[This Transcript is Unedited]
Department of Health and Human Services
National committee on Vital and Health Statistics (NCVHS)
Next Generation Vital Statistics:
A Hearing on Current Status, Issue, and Future Possibilities
September 12, 2017
Hubert H. Humphrey Building
200 Independence Ave., SW
TABLE OF CONTENTS
- Reconvene & Welcome – Dave Ross and Bruce Cohen, Hearing Co-Chairs
- Panel 4: Challenges and Opportunities: Federal Level – Delton Atkinson, Eric Hughes, Caitlin Cross-Barnet, Lekisha Daniel-Robinson, Ursula Bauer, William Riley, Karen Glenn, Mark Bye, Benjamin Bolender, Corryne Carter
- Committee Q&A
- Panel 5: Challenges and Opportunities: State and Local Levels – Elizabeth W. (Lou) Saadi, Terra Abrams, Mary Ann Sens, Gregory G. Davis, Mary Beth Kurilo, Amy Zimmerman, Michael Hogarth, Janet Hamilton
- Committee Q&A
- Audience Input
- Panel 6: Future Directions – John R. Lumpkin, Kalvin Yu, Chesley M. Richards
- Committee Q&A
- Discussion and Pubic Comment
- Moving Forward: An Integrated Vision-Discussion – R. Gibson Parrish
- Wrap-Up – Dave Ross and Bruce Cohen
P R O C E E D I N G S (8:33 a.m.)
DR. ROSS: Good morning. Welcome to the second day of the National Committee on Vital and Health Statistics hearing on Next Generation Vital Statistics. To get us started, I first need to do our statutory duties. I am Dave Ross, co-chair of the planning group, along with Dr. Bruce Cohen. I am a member of the National Committee, and also a member of the Population Health Subcommittee, and I have no conflicts.
DR. COHEN: Bruce Cohen, member of the Full Committee, co-chair Population Health, no conflicts.
DR. RIPPEN: Helga Rippen, member of the Full Committee, I am also a member of the Subcommittee for Privacy and Population Health, no conflicts.
DR. STEAD: Bill Stead, Vanderbilt University, Chair of the National Committee, no conflicts.
MS. KLOSS: Linda Kloss, member of the Full Committee, co-chair of the Privacy, Confidentiality, and Security Subcommittee, member of the Standards Subcommittee, no conflicts.
DR. MAYS: Vickie Mays, University of California Los Angeles, member of the Full Committee, member of Privacy and Pop, and no conflicts.
MS. GOSS: Good morning, Alix Goss with Imprado, a consultancy of Dynavet Solutions, member of the Full Committee, co-chair of the Standards Subcommittee, and no conflicts.
MR. COUSSOULE: Nick Coussoule, BlueCross Blue Shield of Tennessee, member of the Full Committee, co-chair of the Standards Subcommittee, member of the Privacy and Confidentiality Subcommittee and no conflicts.
MS. LOVE: Denise Love, National Association of Health Data Organizations, member of the Standards Subcommittee and Population Health Subcommittee, no conflicts.
DR. PHILLIPS: Bob Phillips, American Board of Family Medicine, member of the Full Committee and co-chair of the Population Health Subcommittee, no conflicts.
DR. ROSS: Let’s check to see if there are any members on the phone. Okay, we have several. Go ahead.
MS. STRICKLAND: Deb Strickland, member of the Full Committee and the Standards Subcommittee, no conflicts.
DR. CORNELIUS: Llewellyn Cornelius, member of the Full Committee, Population Health Subcommittee, from University of Georgia, no conflicts.
DR. ROSS: Lee, how is the weather down there?
DR. CORNELIUS: It wasn’t as bad as we expected, but there are a lot of trees down and power outages around the metro area.
DR. ROSS: Okay, so look forward to coming back soon.
DR. CORNELIUS: Exactly.
DR. ROSS: Good morning, everyone. This morning, we will tackle the next day in this very interesting adventure on e-vitals. I have the help of my co-chair, Dr. Bruce Cohen, and also Kate, the hook, Brett, who will help our speakers be aware of time. Before we begin our first panel, I just thought I’d make a — I can’t resist a few comments. Yesterday, I came in knowing that I knew not that much about vital records, and I came out knowing that I knew very little.
We learned an awful lot, it was incredibly enlightening. We learned that vitals data assumes by many, without any kind of recognition of the challenges that are posed in being able to gather it and use it. This is a very challenging area. We also learned that there are many, many purposes to which these data can be put that create very large societal benefit as well as business benefit; I think that was something else that was important.
At its foundation, vital records really are the beginning of a U.S. public health system. Without them, we don’t know what the denominator is. This is among the most important foundational data public health has. And yet, this federated system that we live in that’s an artifact of our Constitution is just a fact of life. We have to figure out how to work with that.
So state prerogatives matter. Tara Das yesterday said, really admonished us to challenge the public health vitals community to seeking excellence and not really to be held back by the states that want to move more slowly or can’t move fast. So I think today is devoted to examining what that future could look like, or the future state, the future possibilities, and what those future opportunities are.
For me, yesterday, there were a few key things that were important. There were many, many, and we’re going to have to put together a comprehensive report, but I think the idea of decoupling the legal aspects from the use of data, something that’s going to have to be explored. Of course, many opportunities of linking vitals data with other datasets, business benefits, potentials for informing pension plans to reduce payments to dead people, is actually to everybody’s benefit to allow pension plans to survive.
And I think the idea, the overarching theme of aligning and possibly restructuring financial incentives to make it more possible for the states to make their vitals enterprise more timely, more accurate. There is a lot to be done, a lot to be discussed. So today, we move towards this digital future.
Delton challenged us yesterday, or teased us, I should say, with what he’s going to say this morning, so we’ll go ahead and start with our first panel, Challenges and Opportunities at the Federal Level, first with Delton Atkinson.
DR. ATKINSON: Good morning. I am happy to be back, to have an opportunity to speak with you again, and hopefully at the end of this, you’ll still want to speak with me. And let me say that yesterday I talked about my experience in the vitals world from the federal perspective, state perspective, nonprofit perspective, private sector perspective, and I think some of my thoughts and suggestions will come from all of those perspectives in terms of how we need to go forward.
Some will be things for which people have said you can’t do but we have shown that you can make a difference, and I’m going to talk about seven things that I think are going to be critical as we move forward. Number one is the timeliness of records to NCHS, and especially on the mortality side. We got started with some efforts to improve the timeliness of the transmission of records to NCHS.
When we started, we got 7 percent of the records that were transmitted to us within 10 days of the date of the event. That was in 2010. We are right now at the level of 55 percent of the records being transmitted to us within 10 days. We wanted 80 percent of those records transmitted to us within that particular timeframe.
There were a lot of naysayers that say you cannot do that, and we have shown that with the help of some strategic partners that you can make a difference in the system in terms of how it operates, and we like to really appreciate the support of our Office of Public Health Preparedness and Response as well as the Office of Public Health Statistics and Services and NCHS Director’s Office, but also another one, and that’s the Patient-Centered Outcomes Research Initiative, PCORI, that ASPE oversees have been very instrumental in helping us to really try to make a difference.
And while we’re at 55 percent, I don’t believe that we need to stay there. I think that there can be some more improvements. If you look at the range from state performance, we go from 0 percent to 95 percent of a state’s records being transmitted to us within 10 days, and everything else in between.
So we really need to make a difference in that. We need to continue that particular activity, and one of the results of this activity has been the fact that we have now really begun to talk about mortality in the terms of surveillance. You can’t do surveillance if you’re not getting the records more timely than what they are. In this particular effort, there needs to be more targeted interventions with some of the worst or underperforming states. We talked about this range of performance of states, and somehow in this federated system, we need to narrow that gap. Because again, we talk about only being as good as the worst states. Well, we have some states that are not performing, and we need to do some things about it.
And we need to also look at improving our abilities and capabilities to working with the medical examiners and coroners. I have two of my staff members now through some special funds that we’ve gotten from public health preparedness that’s given us an opportunity to look at the interoperability of case management systems and medical examiners and coroners’ offices with electronic death registration systems. We want to really improve on that whole area.
Second, and this issue has come up quite a bit in terms of vitals, and that’s the whole issue around the quality of the vital statistics data. And again, I look at this as we look at the whole issue of timeliness with people sometimes it can’t be done, we can make a difference. We have just not been committed to really making a difference in the 41 years that I’ve worked in the health statistics field. We talk about it, we study it, we haven’t been committed to it.
On the birth side of the picture, we’re beginning to look at and put in place some e-learning training for clinical and nonclinical hospital staffs, redesign of our facility guidebooks, we are creating in the NAPHSIS cooperative agreement that started the first of September a state and federal workgroup that is going to evaluate some of the modifications and additions that we need to do in birth reporting, and again, we really need to get aggressive in the whole area of the interoperability of birth information from medical records to electronic birth systems.
If 99 percent of your births are occurring within the hospital, then I think it’s very achievable goal to look at the electronic transmission of the medical information from the electronic medical records to the electronic birth certificates. And by the way, we have staff within NCHS who are working with my division and we are working with HL7 to put in place national approved standards for birth and death transmission.
Second area obviously relates to the cause of death. We had some discussions yesterday about some of the problems and issues with the quality of the cause of death, and that is an area that truly needs to be worked on. We do have to figure out ways to work better with our physician community.
That community will have to improve its ability to be able to report the causes of death and here also we need to facilitate the interoperability of electronic health records with electronic registration systems, and I’m going to talk about this a little bit more. But the one thing we must do, we must prevent the urge to develop a drop-down list.
That is always one of the solutions that people talk about is let’s just make a list of the most common causes of death, put them in an electronic system, and let the physician just choose one. With that particular strategy, how do we pick up emerging diseases and problems? That’s the thing that people don’t think about sometimes. And also in this particular area, we need to obviously develop better benchmarks and measures on quality.
I’ve talked a little bit about the interoperability. I think this has to be one of the major recommendations coming out of this group. It has to be a much stronger emphasis on the interoperability on vital registration systems and health and medical systems.
We at NCHS, and working with our state counterparts, must be able to put in place these national standards. It’s been a slow process, but as we look at HL7 standards, as we look towards the new FHIR standards, we must move our systems such that they’re interoperable with other systems that are operating. Not only the electronic medical records, but again, as I said before, medical examiners and coroner systems, as well as funeral home systems. We have to get away from the standpoint of entering data two, three, four times into different systems in order to get the information that we want. Sometimes that requires us to get beyond our pigeonhole that we seem to be at.
We had a very interesting project with the state of California and the University of California at Davis working with Dr. Hogarth.
There, we looked at capturing the information from the electronic medical records, that information being transmitted to the state, using draft HL7 standards, that information at the state then transferred to us. We do the coding of the cause of the death, turn it around, send that information back to the state, those codes, and those codes then going back to the hospital for which they came from.
So beginning to link from the local level to the state level to the federal level, these various systems that we want to talk about. The last one I’ve got here is the linkage of the medical examiner coroner case management systems with electronic death registration systems. These are going to have to be improved in the pass.
The fourth one relates to the leveraging of technology. Here we need to take advantage of technology to be able to do the various things with it. I do not believe that any records should be coming to NCHS that does not pass, cannot pass the NCHS edits, period. In this day and age, every system that’s operating out there minimally should have the NCHS edit specifications and no records should be coming to us that can’t pass that. That, folks, is a performance issue of those systems, and we should not have that in place.
We should begin to look at the implementation of VIEWS. Views is called the Validation Interactive Edits Web Servers. That’s where, when a person is entering the cause of death, the system will pick up that information, send it to a web service, and send back some information about the quality of what you just entered.
Some of those things that these vague causes of death, which we see all the time on death certificates, and other kinds of things, that should be caught at that point. Until we get to that point, trying to change it, trying to catch up on it, on the back end of the process, is never going to solve the answers here. I know that VIEWS is operating in about 12 states, and I know that some states, they’re concerned about the physician not participating.
Again, that’s where we need to work with the physician community and talk about the importance of you delivering quality information, and if the system responds back that is not good information, the system says that what you entered is misspelled or what you entered is a vague cancer without telling us the site, that those kinds of things we can’t continue to try to clean up on the back end.
The third area under there is the modernization of our medical mortality coding system. That’s a system totally within NCHS. It’s how we code the causes of death. Tremendous changes need to be made in that and we need to use natural language and machine learning techniques to mine all of the literal text information on the death certificate, including information about what the specific drugs were that was associated with a drug abuse death, and so forth.
Let me just skip a couple of things here. We also talked a lot about the value of vital statistics, and here I think is very important for us to really think about, and I think it came up yesterday, there’s got to be ongoing financial participation of others beyond simply NCHS and SSA. If you remember the amount of money that I said we gave to states, if you divide it by 57 jurisdictions, that’s about $400,000 per state. That will not even buy you the electronic systems which we are demanding that states implement. There’s no way that they can meet that in terms of that.
We really need to have our federal partners in their grants and cooperative agreements include vital statistics as a priority. I’ll give you an example. Over the last year, we’ve negotiated with the Office of Public Health Preparedness and Response, and in their cooperative agreement with states, their state counterparts, they put in there a priority that says their funds can be used to help improve state vital statistics.
We need to look at other kinds of strategies and other places where we are demanding that states do some things to improve the systems. I mentioned before narrowing the gaps between jurisdictions. This is a critical need. Somehow we need to figure out how to modernize some of the vital statistics laws. Some of those are completely out of date.
There needs to be the adoption of standards, and we are working with the public health accreditation board and NAPHSIS and the states to hopefully put into place an accreditation program so that we begin to put in place minimal standards that programs should be achieving so that we can improve the performance of programs.
I just want to touch bases about NDI. Again, the real goal with NDI that we wanted to try to get to is having a quality system but inexpensive. By that, I mean that we really want to look at ways to reduce the cost to researchers while maintaining the state payments for the use of the records. That may seem like a contradictory kind of goal, especially if something that I said was 100 percent receipts-oriented, but that’s the goal that we need to try to work on.
If you look at performance, here you see the chart that says that this is the number of months after the close of the data year for which the final NDI file is ready, we’ve gone from something like 22 and 30 months after the close of the calendar year, down to about 10 or 11, and if you throw in the provisional file, NDI file, we do that one month after the close of. So we’ve made tremendous progress in terms of the data. But it gets to that question, is it really valued or do we devalue this whole process?
With respect to NDI, we approached PCORI last year who provided some funding for us to take a step back and look at NDI. We’ve hopefully come up with a strategic plan with the target of April 2018. You’ll hear a little bit about some of the economic models, but taking a look at some of the economic and non-economic barriers to NDI, but when you look at this, and I’m wrapping up, I think that there are three alternatives.
Now, there is some permutation of each one of these. We can continue NDI as it is, meaning that we continue to charge, we try to improve with the addition of trying to improve the process and the research applications and hopefully with state approvals to begin to make some changes, do some easing of restrictions. How much, not quite sure.
Second is to begin to look at and talk about new economic and operational models for NDI and I’ll let Eric talk a lot about that, but beginning to look at new ways in which this is financed. If it’s valued, look at new ways for which we finance that.
The third option, if nothing can be done and if everybody is complaining and so forth, end NDI as a national dataset at the federal level.
If all of this that we do and we still can’t make folks happy with it, you end it, you give the dataset back to the states and back to DoD and say, it’s not worth it. That, folks, is my thoughts about it.
Let me say, let me take 30 seconds to say one other thing.
DR. ROSS: We are going to let you pull rank today.
MR. ATKINSON: I don’t know how many of you have seen the new report from the evidence-based policy commission. I looked at that report and I was quite disappointed. For vitals to be such a foundation of public health, of the data that we need for research, for policymaking, program management, and so forth, for it not to mention vital statistics at all, I think shows that issue that we are talking about.
I can’t think of another dataset that is more evidence-based decision making, is needed for that than any other dataset around, other than probably the census, but it wasn’t mentioned at all. That to me is a disappointment.
DR. ROSS: Thank you, Delton, for your remarks. Yeah, it’s sort of like, we tend to forget about foundations of buildings. We look at the glitter on the outside and don’t pay attention to the fact of what’s holding it up. I think you’re right. We assume vital records in this country and that’s what this hearing is really about, is to try to bring back some visibility to what’s needed, what the opportunities are, to where we can really improve. So thank you for your remarks.
I am typically terrible about keeping on time. So now I’m going to turn it back to where Kate is going to keep us. We wanted to hear Delton out because it’s really very important, but now we’re going to get back towards schedule.
So Eric, please?
DR. HUGHES: Thank you and thank you for the opportunity to speak with you today. I know your goal in the committee is to give important advice to health and human services on uses of data to improve our health and security and I want to focus on those areas.
I want to tell you about the urgent need to modernize vital records and request your support for a plan to use death records to address the health issues that are facing the nation. For the last year, I’ve had the privilege to lead a project for Delton on modernizing the National Vital Statistics System focused on the collection and use of mortality data or death records and focused on public health applications. We’ve heard about business applications as well.
The project is sponsored by NCHS under the CMS Alliance to Modernize Healthcare, or CAMH. CAMH is Health and Human Services’ federally funded research and development center. It’s operated by Mitre to provide independent, objective advice to Health and Human Services.
So this morning, I’ll give you some background on what we’ve learned about the challenges and the opportunities facing the vital records system and then describe some recommendations to use vital records as a foundational element in improving healthcare and public health in this country.
Vital records are essential to our understanding of and response to important health crises including the opioid epidemic, preparing for the next flu pandemic, developing new cancer treatments, and preparing for emergencies like the hurricanes we’re experiencing.
Mortality data in particular provide us the most comprehensive information available on health outcomes, covering all deaths in the country. Although timeliness and quality of mortality data, as Delton spoke to, have improved significantly, continued improvements in the data are urgently needed to address the complex crises that we face.
We have to also provide this data in a way that’s sustainable for the jurisdictions and for the government. Without your support, I’m concerned that we may continue to underutilize vital records data, and through attrition may lose access to systems like NDI.
So before I discuss our recommendations, I want to give you a little bit of background on a research process so you can put those recommendations into context. For the last year, our project has examined death registration systems and processes used by the 57 jurisdictions to collect death records. We’ve studied the NDI and other systems that make death records available to others for health purposes. We conducted an in-depth environmental scan interviewing over 50 stakeholders in mortality data.
As you can see on the slide, the process for capturing and using mortality data is complex and we’ve heard about many different parts of it, but for me, it’s very illustrative to see it all on one place. In the jurisdictions, data are captured and validated by funeral directors, physicians, medical examiners, coroners, and registrars. We heard from many of those yesterday.
These data are reported to NCHS. Cause of death is coded using the standard scheme that Delton spoke of and then it’s returned to jurisdictions for their use. Death records are used to generate those mortality statistics that we rely on as a country to understand leading causes of death, and they form the basis for NDI, which is the nation’s resource for health research.
So, as detailed in my written testimony, great strides have been made in improving the timeliness and quality of mortality data, but we need the following to happen to modernize vital records to best serve the health needs of the nation.
First, we need wider application of mortality data to important health issues. Second, we need to modernize electronic death registration to help states increase the coverage, timeliness, and accuracy of mortality data. Third, we need a more sustainable funding model for mortality data, and we recommend a public-private partnership to facilitate innovation and partnership with the states.
On the first point, we need wider application of mortality data to important health issues. Although, NDI has been used for 35 years for health research and e-fact of death is becoming available for counter fraud and the administrative uses that we spoke of yesterday, public health officials and healthcare providers aren’t often using mortality data to see health outcomes of the populations they serve.
National mortality data are not often used in these systems in areas like healthcare quality measurement, improvement, post-market surveillance of drugs in medical devices, tracking and improving patient safety. These are foundational parts of the business of healthcare. This means we’re missing out on opportunities to inform our health improvement efforts.
Second, we need to modernize electronic death registration systems to help states collect more timely and accurate data, to keep pace with the needs of public health response.
Mortality data, as Delton mentioned, are still not timely enough to meet the needs of our rapidly evolving problems, health problems, and we still fall well short of the goal of 80 percent of deaths available within 10 days of the event, particularly for causes like drug overdoses. For example, the recent interim report by the Commission on Combating Drug Addiction in the Opioid Crisis cited 2015 mortality data. That’s because 2016 vital mortality data were not available at that time.
Although some states have advanced electronic death registration systems, six states lack even basic electronic registration system capability. Several more states capture less than three quarters of their deaths electronically, hampering our efforts to improve timeliness.
We also need to improve the accuracy of mortality data. One recent study found significant errors in half the deaths reviewed with 9 percent requiring correction. As you heard, NCHS has established goals to improve in these areas, but the funding that’s been used is of limited duration and may not pay for all the modernization that we need for public health.
Jurisdictions need next generation electronic death registration systems to collect more timely, accurate data. These systems must support electronic exchange of data with other systems, including electronic health record systems, case management systems as Delton mentioned, and also systems that are used by funeral homes. Medical examiners and coroners in particular play a key role in registering overdose deaths, and we need to partner with them and other stakeholders to assure that the technology reduces the burden on them to help us collect this important data.
We’ve developed a prototype next gen EDRS and made it available to all. The written testimony includes the details for how to get it. We’re using the prototype to assist states in the process of acquiring their first EDRS and are piloting next gen EDRS services to demonstrate their value in addressing public health issues. We need your support jurisdictions continue to take advantage of these services.
Finally, we need a more sustainable funding model with a public-private partnership to drive support and guide innovation efforts. Researchers, as you’ve heard, need less expensive access to NDI. The cost to use NDI is high compared to other research data sources, limiting its application in important research areas, and also putting its funding to jurisdictions at risk.
As part of a new business model, we should explore wider uses of mortality data in addressing public health issues while protecting privacy and confidentiality. Our conversations with over 50 stakeholders identified the opportunity to form a partnership focused on enhancing readiness and capacity to improve and address public health issues. The partnership can tap into new sources of funding to help states modernize systems.
In conclusion, I urge this committee to support the modernization of vital records, and in particular, mortality data to help improve the health and security of the nation. Thank you.
DR. ROSS: Thank you very much. Right on time. Appreciate it. As with yesterday, we will go through our panel of speakers, and then we’ll open for the committee to have discussion.
Next we have Caitlin Cross-Barnet from CMS.
DR. CROSS-BARNET: Lekisha and I are both from CMS, but she’s going to start and then I’ll go on.
MS. DANIEL-ROBINSON: Thank you for having us and for allowing us to change your agenda just a tad bit with that order. As you heard yesterday, from my CMS colleagues Karen Matsuoka and Cathy Carter, vital records really plays an important role to the work that we do at CMS. So today I’m just going to talk a little bit more about some of our experiences as a part of the CMCS Maternal and Infant Health Initiative with states and vital records, particularly noting the challenges, which I think in itself are opportunities and have already been sort of reflected between conversations yesterday as well as this morning.
Our Maternal and Infant Health Initiative explores program and policy opportunities that improve outcomes and have the potential to reduce the cost of care for women and infants in Medicaid and CHIP. Through our initiative, we are working very closely with states on a variety of activities, and as you might expect, data is an important underpinning of that, and birth certificate information in particular.
The vital statistics information is an important source or potentially an important source of info for monitoring maternal and infant health outcomes, understanding maternal demographics and risk factors, and calculating Medicaid core quality measures that Karen talked about yesterday that become more precise with the addition of vital statistics information.
But as others have mentioned, there are some challenges and I would say perhaps one of the fundamental challenges is the standardization issue. So, given that the data is locally controlled, there is a lot of variability in it, whether it’s variability at the state level or within states at various facilities and sites. Reliability and completeness we’ve also noted to be of challenge. The clinical data in the birth record does not always reflect the medical record information or data is missing as such.
So the other issue, data access. I think this one was a big issue that we noticed with our states. Privacy and data access rules seem to be a bit antiquated, which does not allow for the exchange of information between state agencies, so we’re not even talking about intrastate agencies, we’re talking about Medicaid and the cases that we’re working with with public health not being able to exchange the data, which obviously limits the ability to make informed decisions, informed monitoring, that sort of thing.
As already has been discussed, data availability. So the data lag is an issue. You can’t do some of the real-time data monitoring with data that lags one to two years.
So just to talk a little bit about some of the things that we’ve done and some of the other things that we’ve seen that speaks to the potential of the system. One, given the fact that we saw some states having some challenges with reporting on the measures that require the vital statistics data, we did some training on data linkages. So your committee has already talked a lot about that. That’s where we notice some of the challenges, in particular, some of the communication issues between agencies within the state.
But with that data linkage effort, states were then able to report on the core measures that are part of our core set, so the C-section and the low birthweight measures. But I think there is still a challenge in terms of sustainability given capacity issues. So whether it’s analytic capacity, changeover in staffing at the state, or just the needs, the great needs that are at a state level for analysis.
So while calculating certain birth measures or maternal- and infant-related measures are really important, they’re just a number of things that the staff may have to do. So they’re very much stretched. That’s certainly a big issue.
Another project that we worked on with the Medicaid medical directors really I think speaks to the potential of the system. So a couple of years ago, we worked on a project that the Medicaid medical directors wanted to have a better understanding of early elective deliveries in the Medicaid population. The project collected aggregate data from birth certificates and where possible linked it to Medicaid claims data. The rates were calculated using methodologies that have already been established within the broader industry, and the states were able to report on that, but really, I think that project pointed to a couple of different things.
One, first, I would say the understanding of the differences in data collection between states. So the process was not very standardized because some states may include certain components in a variable that others may not. I mean, just something as simple as the Medicaid payer may or may not be standard between states.
So that’s one challenge, but I think that project also showed the ability to use this information from a public health perspective and to inform policy design and evaluate programs.
Another opportunity, and I will be very brief with this, that we saw was the ability once states were actually able to link datasets to then validate those data against state and health plan financials and really setting the stage for a single source of utilization data for surveillance, quality measurement, as well as improvement and analysis.
DR. CROSS-BARNET: Thank you, Lakisha. I am Caitlin Cross-Barnet from the CMS Innovation Center, and I’m just going to give a few examples of some of the work that we’ve been doing that really requires vital records to be able to do a comprehensive evaluation.
So the primary initiative we’ve been working on is Strong Start for Mothers and Newborns with the goal of lowering incidence of preterm birth, low birthweight, and maternity and newborn care costs among Medicaid and CHIP beneficiaries. We offer care enhancements through three evidence-based initiatives. We have 27 awardees with sites in more than 30 states and we’ve served about 50,000 women.
To use the vital records, we wanted to determine the outcomes and costs amongst Strong Start participants versus similar women who weren’t enrolled, and some advantages of using the vital records and Medicaid data would be that there isn’t an awardee burden for data collection which gives them more time for program services and makes people more likely to want to participate in our programs and be evaluated, that there is some consistency across states. It is the same form even if it’s not always filled out the same way. And that there’s high reliability particularly on gestational age and birthweight, which were our two main outcomes of interest.
But issues include that the identified data can be difficult to obtain. It’s often time consuming and costly. The processes are different for every state. There are different data use agreements. There are different IRB processes. For some states, you actually have to have individual signed consent from women to get individual data. In any case, my contractor’s been working on getting data for 20 states for about three years now and we have it for about half of them.
So medical chart review comparisons, because we did collect medical chart review for our participants, reveal varying data quality with some variables showing high concordance and others moderate or very low. And Medicaid participants aren’t always identified accurately on the birth certificates. Then the data quality in many fields is just kind of random or often not filled out. So we’ll know if there’s a positive, but there are so many false negatives it’s really hard to know what’s going on.
Other potential evaluation uses in our center, we have an initiative beginning for the Communities of Health for Children and Youth. We want to be able to look at a baseline of children’s outcomes when they were born so that we can develop really good comparisons for looking at outcomes for children as they go through the program. We are also looking at a bundled maternity payment model to address neonatal abstinence syndrome where we would want to be able to look at outcomes for infants and be able to link those to our claims data.
So the priorities for the innovation center are the ease of obtaining beneficiary-identified vital records, so vital records that actually have the Medicaid box checked accurately, the reliability of the demographic data, the reliability of the health and outcomes data, and the accurate identification of the birth attendant because for so much of the time now, if a midwifery practice is under an obstetric practice, the obstetrician gets listed as the birth attendant.
DR. ROSS: Sorry we had to cut you short.
DR. CROSS-BARNET: No, I was done.
DR. ROSS: Well, thank you. Thank you very much. We are going to move on to Dr. Ursula Bauer who is head of chronic diseases at CDC.
DR. BAUER: Thanks so much. As I said yesterday, those of us who sort of work on the vital statistics side, including myself, don’t really think about the whole vital registration function that this system was really set up to serve. So we built this surveillance system, not from scratch like we do with the national notifiable diseases system or like the behavioral risk factor surveillance system, which also have a huge state component, but we built it on the back of a different system that was established for a different purpose.
As I listen to the discussions yesterday, I really wondered how streamlined and solvent, even, the vital registration system would be if it were focused exclusively on documenting the fact of the event and weren’t saddled so much with the enormous data collection that the vital statistics system requires as a public health surveillance system.
Then per some of the discussion that we’ve already had this morning, given that we’re in the 21st century, we’re moving toward a century removed from when this system was set up and we have all of these new processes and tools and resources like electronic health records and so on, were we to build a new kind of freestanding system, might that better meet the needs of our public health surveillance. So just some thoughts to think about.
Given that we have the system that we have, these are the challenges for that system. Timeliness, quality, and accuracy. We’ve heard those already this morning. We’ve made huge improvements in each of these areas over time, but certainly have a lot more to do to optimize the system for surveillance.
Quality really refers to completeness of the story and we heard a little bit about that already. Did the person who died of the heart attack also have diabetes? Were they in treatment for lung cancer at the time? Was the motor vehicle crash victim in the car or on the sidewalk? We need so much more information to really flesh out the story. Sometimes we have that and often we do not.
I mentioned the issue of race misclassification yesterday. We do have a fix on the vital statistics side. We don’t have to do that extensive training on the collection registration side. That is something I think we should fix. Certainly, if up to 40 percent of whites were misclassified as American Indians, we would have solved the problem a long time ago.
On the birth record side, we just want more information. We heard how difficult it can be to get that, both at the point of collection and also often the data use agreements really limit what we’re allowed to use and how we’re allowed to use it. For both birth and death records, there are so many stakeholders, and so finding solutions implementing them, the jurisdiction by jurisdiction strategy, is very cumbersome.
The system largely works well. I mean, yes, we have 2015 data when it’s 2017; that’s not optimal. We are certainly not in real-time identifying emerging issues, but largely we have a pretty robust surveillance system and more resources would make it work better.
We were asked to address this question of the consequences of inaction. Not really sure kind of what those might be. One imagines we can make ourselves irrelevant on the vital statistics side if we’re unable to keep pace with new developments in technology, with innovations in the way data are used. We heard a little bit yesterday on the private sector side about going around the vital statistics system to find data elsewhere. So that’s potentially a consequence of inaction.
And without investing the appropriate resources, the system can become overwhelmed. We had the foundation metaphor already, perhaps building a house that’s too big for the foundation and that could collapse. Thank you.
DR. ROSS: Very good. Thank you very much. Our next speaker is Dr. Bill Riley from NIH.
DR. RILEY: Thank you, David. Yesterday I spent some time talking about the critical value of our vital health statistics program for NIH support and research. I’m going to spend some today on opportunities to strengthen and improve that vital system and improvements that we could make, but also put that in the context of larger efforts to create a more integrated national surveillance system and some of the support that’s going on in various places that seems a bit surprising when you look at it relative to what we put into the vital statistics area.
I will repeat what others have said about the importance of timeliness for this data, that that’s particularly critical. It typically takes over two years or longer from study completion to the date of a publication. There are many factors other than the lag in death data that account for that, but that is certainly one of the things when people are including death data in development of their publication.
We currently, as others have said, we curate on an annual basis and we pull that data together on an annual basis. The requests from researchers coming back to us is usually at the end of the study and one large request to gather all the data at the end. If we had a more up-to-date sort of case by case basis of queries into the system, being able to pull those out, not only would be able to get data in a more readily basis than we currently are able to do, but would also allow us to detect serious adverse events more quickly than we do at this point. I think we miss some of those over the course of studies.
Obviously to get this done just in terms of getting the data in, it’s surprising that EDRS isn’t adopted across all the systems for death, that even though it is mostly for births, and that we don’t have a common system for that so there is some sharing of that across the system, it would make it easier to at least get the data in and that would certainly be helpful in the process.
The other is the cost structure which other people have discussed as well. Let me just say a few things about that. So right now, we have a fee-for-service structure. Nobody knows exactly why we ended up with that, though it makes more sense in a paper-based system where most of our costs were variable and very few were fixed. We now have mostly fixed costs and upfront costs of electronic systems and a fee-for-service model doesn’t work particularly well under that system.
As other people have argued, states are allowing these revenues and other services for support of their vital statistics system, some entirely for these fees. I’ll just remind us all that transaction fees cost money as well. So those transactions also cost money, so the more of those we have to do on a fee by service basis, the more complicated and costly that becomes.
I’ll also note that the federal government already pays for some of these services, but we do it in a very inefficient and indirect way. So at least for the NIH, what we do is we send money to grantees. Those grantees are paying for their employees. They’re also paying for the fees and we’re paying in direct costs that are sometimes 50 to 100 percent over that original fee.
So it’s pretty easy to assume that for every dollar that’s going to Delton’s NDI system to be able to get that data, the NIH is paying two dollars to get that one dollar of service. So there are obviously more efficient ways to do that than we’re currently doing.
Streamlined process for linking NDI and health service research data, we’ve already talked about researchers send a list of personal identifiers, NDI sends back candidate matches for that and probabilistic system, and again, a fairly batch sort of way. There have been lots of people, including the AHA, but a number of other organizations have talked about doing this in a more streamlined manner.
If we could, and again, this is outside of my area of expertise, but there’s certainly secure APIs and there are secure data enclaves. There are secure ways in which we could more readily on a case by case basis be able to go in, put in a personal identifier, in real-time get that data back whether it’s provisional and/or final, just so we know by metatags which one of those they happen to be, then be able to extract that data as needed and only the data for the most likely probabilistic match, not for all of the possible probabilistic matches. That would be useful, I think, as we move forward.
I want to put this in context, though. So as we know from the recent hurricanes, we have a really nice weather service national surveillance system that allows us to do advanced warning for that. It’s integrated, it’s comprehensive, it’s international, it drives data back to users in real-time for access whether you’re a researcher or a consumer or commercial entity.
We have it for plate tectonics and earthquakes. We have it for the background radiation in the cosmos. You can even, we have an international surveillance system from those radio telescopes that even allow us to determine whether there’s intelligent life on other planets sending data to us.
Yet we do not have that same sort of integrated comprehensive system when it comes to health. What we’ve been talking about is on that top right; it’s the integration of birth and death certificates and other vital records, but it’s only a piece of the system.
As you know, we have lots of surveys, also not all that well-linked even though we’re increasing our ability to link those survey systems better than we have done in the past. We have multiple vendors of electronic health record systems that also don’t link to one another.
As you’re probably well aware if you’ve ever moved to one doctor to another and had to go get your CD-ROM of all of your data and carry it to another system that can’t integrate it into there at all, and so as a result your doctor probably pays no attention to that data.
We have lots of very sexy things that we’re doing in genomics and proteomics and metabolomics and all the other -omics of the world, and all of that data is there and there are efforts to try to both standardize and make them a little bit better. Then of course we have all the emerging technologies of data being collected by consumers that can be pulled into that system. All of this of course is disjointed, fragmented, does not connect, and what’s crucial in all of that is that we know people from birth to death. Without knowing that, all the rest of it makes no sense.
Some of the work that we’ve done has been in an effort to try to do some of that, some of the work that NIH has done. I just want to briefly in my no time left to go talk about what this might look like.
I’m going to say some things that I know are all doable, it’s just that we don’t have either political will or the finances or whatever it takes to be able to do this, but clearly doable if we think about it. I think we ought to put it in this context, that we monitor people’s health in the U.S. health population from birth to death. We have unique health ID for every individual. I know that’s been problematic for many, many years. Electronic health records that link to each other and to all these related datasets, increasing integration of protected genetic data with EHRs, core basic survey data that’s obtained on everyone, more in-depth and intensive data on representative subsets.
The ability of participants to donate data, especially nontraditional datasets like their activity trackers and heart monitors and that sort of thing. Invitations into that system for clinical studies and trials, as you know, our studies are not representative. Our epidemiological studies are, but almost all of our clinical trials are convenient samples, and as a result I always wonder what we are inferring, what population we’re inferring to from those studies.
And integrated and accessible sort of data enclave that researchers can access. I mention this not because it’s beyond the scope of this particular committee, but I think if we are thinking about a national surveillance system for health, death and birth have to be a critical part of that system and without it, there’s no way for that to happen. Thank you.
DR. ROSS: Thank you very much. Our next speaker is not Stephen, is it? Please introduce yourself.
MS. GLENN: Good morning, everybody. I am Karen Glenn. I work with Steve Goss. I’ve also got my colleague, Mark Bye, with me. We’re going to do a little tag team here. We do have some slides as well.
While we’re waiting, we approach this as sort of our wish list for what we could get for vital records. Steve gave you a little bit of background yesterday on what we do. We are the Social Security actuaries. Every year, we’ve got to project out the finances of the Social Security program for 75 years. So the vital statistics are really the basis for everything we do. Our 75-year demographic projections depend very heavily on those vital stats.
As an overview, the births and deaths data we get is very reliable. Actually, we are pretty happy with what we get. Obviously, like everybody else, we would like to get things sooner, but overall, it really is meeting our needs. Marriage and divorce data is another story, and Mark is going to go a little more in-depth on that. The other thing we really need is more data on immigration. I know that isn’t your focus, but if we have a couple minutes at the end, we may go over that real briefly.
Births and deaths, like I said, it’s been very reliable, even back a few years from there, we’ve got some good data. Underreporting has been an issue in the past, but it’s better now.
One thing we focus on is really the entire population that gets Social Security benefits, which is a little bit different than just the United States. We look at the states, the territories, D.C., federal employees and their dependents who are overseas, Armed Forces overseas, Social Security beneficiaries living abroad, and some other Americans living abroad. So it is a little broader area than I think most of you typically focus on.
We would love to get data on the territories from NCHS. We get that, but we would like the real data for births and deaths from NCHS, and we would love to get population estimates from Census on that. Any birth and death data we can get outside the mainland United States — not the mainland, the United States — would be really useful to us.
Now Mark is going to talk a little bit about marriages and deaths.
MR. BYE: One of the main things we need, and this is the existing marriage grids — age of husband, crossed with age of wife — so obviously we can’t get that directly from vital statistics, at least we can’t get it because we need it at each point in time. So, vital statistics is marriages and divorces. So in theory we could build it up, but there’s immigration stuff. So we need it each year.
So we use the ACS. Is everyone familiar with the American Community Survey? Not vital statistics, but it’s been very helpful for us. So we’ve been using that. That gives us — but the problem with it is, the ACS only gives us people living together. So there are some couples not living together, maybe they’re separated, maybe someone’s in a home, or whatever, but that’s one of our issues. And if we could get that with vital statistics, that would be great. I’m not expecting to the existing married population.
And the other thing here is same-sex marriages. This is really, this is a huge need for us, which you’ll see in a couple more slides. It was federally recognized in, I guess, 2013, so we need grids for each age of husband, age of husband, age of wife, age of wife.
So the next slide we can do, so we also need the same thing with new marriages. Our old data, from NCHS in the 1980s and a little bit into the 1990s, to a lesser degree, actually did have age of husband, it was age group of husband, and then we got it crossed with age group of wife. But now we only get the total.
The other thing is if anyone has any ideas for us on other data sources, please let us know. We may have overlooked something. I’m always willing to look at some other data. So ideally we would get this each year, age of husband, age of wife, and for same-sex marriages, similar grids. So we’d rather have vital statistics, but instead we’re using the ACS to do this once again. The ACS actually asks you if you’ve been married within the last 12 months, so that’s how we’re getting these grids each year now.
But once again, it would be really nice if the ACS actually asked, if you’re not living with your spouse, what their age and what their sex is. They actually were going to get rid of the were you married in the last 12 months question on the ACS, and I almost cried. But they are going to keep it, for now.
Divorces data is even more sparse. We did have some age group of husband crossed with age group of wife data from the 1980s, but now we only get totals, and those are very rough estimates, I guess. I don’t know exactly how NCHS does it, but I know they don’t get every state.
Since we were using 1980s data for a long time, one of the things I did a couple of years ago was try to get state-level data, and 19 states actually gave me data, for free by the way; some states wanted us to pay but we don’t have that much money in our budget. So I did update that, and we also have the National Survey of Family Growth we use for some things. I didn’t wind up using it for marriage and divorce just because it only covers 15-44; much smaller sample size, too. Obviously the ACS is a great sample size, 1 to 2 percent now, I think, of the population each year. Although I’ve never had to fill it out and I don’t know anyone who has, either. But I love the data.
So once again — and obviously if you’re divorced you’re probably not living with your spouse, your ex-spouse, so it would be great if the ACS could ask what’s the age and sex of your ex-spouse. Also, we really love marriage duration data. So if we got the marriage date and divorce data, that would be wonderful, too. Ideally we would have NCHS give us all this, but we understand the issues with certain states collecting divorce data, particularly. But that is our real wish, is to have all that data, in real time, instead of using a survey.
And then, once again, with same-sex marriage, we really, really need this data. So I just have that broken down by age of spouse 1, sex of spouse 1, et cetera, same with the other spouse. And duration. So same-sex divorces, there’s even less data on this, of course. I think Massachusetts was the first state to allow same-sex marriage in 2004, but we really want to start looking more into same-sex divorces, so if we could somehow get that data with the survey, or hopefully vital statistics, that would be wonderful.
I’ve read some things, in other countries, where I think female-female divorces were at a higher rate for whatever reasons, so I’d be curious to see if we find that true in the United States as well. We do actually get total divorces in Puerto Rico from the ACS. It’s called the Puerto Rico Community Survey, but it’s the same questions, I suppose. I suppose exactly the same questions.
MS. GLENN: I know we have about 30 seconds left. So instead of going over our immigration needs, which we find fascinating, but maybe the rest of you don’t, let me just clarify why we really need this marriage and divorce data. Marriage and divorce really matters for the structure of our program. People get survivor and spouse benefits from the program, depending on the duration of the marriage, so that’s really critical for us in our projections.
So thank you all.
DR. ROSS: Thank you very much. Our next speaker is Benjamin Bolender, from the Census Bureau.
DR. BOLENDER: Hi, thanks for having me. I run the area that makes the population estimates and projections, the official ones that are used as controls for a variety of things, and so Kate doesn’t get me I’m going to start with the punchline and then work my way back.
If you don’t get anything else from what I’m talking about, three points. One, the vital statistics have importance, high importance, outside of health-related fields, health studies, and I’ll go through that. So it’s not just public health stuff.
Two, a lot of the data that we use to make the population estimates are from disparate sources, and they’re hard to put together, and anything that we can do to make that easier makes the estimates better.
And three, I would say the vital statistics system has a strong incentive to modernize a lot of the things that colleagues from NCHS have talked about. I won’t have a lot to say on specific solutions, because I think a lot of people have already said it, so I’ll just kind of skip over that. But that’s basically what I’m going to talk about.
Our population estimates are the basis for federal funding allocations, hundreds of billions of dollars. They’re used as denominators for health statistics, so for death rates, cancer rates, et cetera. They’re used as controls for the American Community Survey. They’re used as controls for the current population survey for unemployment rates. They’re used as controls for most major government surveys that deal with population. They’re also used by the public sector, the private sector, local government leaders. You want to figure out if there’s enough people in this town to support a particular store, like those are the kinds of uses that people use our estimates for, so it’s very, very broad.
Not only do we publish our estimates on the website every year, but we also deliver over 150 extracts to both internal and external customers to be used for controls. So it’s important, and the vital statistics are our main source of our population estimates. They’re one of the core components. We also use data from Internal Revenue Service for domestic migration. We get data from our federal-state cooperative folks. We use data from the American Community Survey. We get data from Medicare. But the vital statistics are really a core component.
Now when we try to put those all together, specifically for the vital statistics — now I need to make a slight digression because I’m a data guy, but we get three types of data, or we make three types of data in terms of our vital statistics. The first we get from the National Center for Health Statistics. Again, that’s probably our most consistent and highest quality data source.
The difficulties are that they are two years old by the time we’re using them for a vintage. So when we make — we call them vintages, it’s a time series of estimates — when we make our 2017 population estimates, we’re using NCHS data from two years prior. So the final data by race detailed by geography is two years old at that point. It’s 2015. We also use the preliminary totals for the year prior to the vintage year, but we’re still required to make estimates out to 2017, so we need to come up with some more information.
NCHS data, like I said, is probably the highest quality input that we have. It’s clean, it’s — I know with a lot of different uses you may be looking at where did the birth happen, where did the death happen, but we really need to know where did they live. So NCHS cleans it up for cross-state events, that type of thing, which is great.
In the years that we don’t have NCHS data, we also get data from our federal-state cooperative for population estimates members. So those folks are typically state demographers from every state, or someone put in place by the governor to act as a representative with us. We have meetings with them, we have a lot of communication. But the data that comes from them is mostly — a lot of times we can get data for the year prior to the vintage year, which is great. And we use those for county-level distributions.
Now, the problem with that is their totals are sometimes different than what NCHS is going to give us, so we actually use them just mostly for distributions. But that does allow to get a little bit more recent. We don’t have to project quite as far.
The third source of data is us. We do a couple of things to the data. One is we have to do those short-range projections to get to the vintage year. And we even need to go out a little bit farther because we need to provide controls for the current population survey for the current month.
The other thing that we do, the data that we get from NCHS is only in the 1977 race detail, from the OMB standards. So the OMB changed the race standards in 1997, and I believe the first year that all states were reporting in that standard was 2016, and somebody here can correct me if I’m wrong. But that means for the last 20 years, give or take, we’ve been converting data from the four race categories to the 31 race categories.
Yeah. See, that’s the response.
We worked with NCHS back in 2000, we came up with a fairly good way to do that, but there’s always some questions involved. And then, like I said, we use a system of rates and population estimates to create projections to go out to the current period.
So in terms of challenges, I think you can kind of see how complicated this can get very quickly. So anything that we can do to get the data earlier, more quality, more flexible system, we would be an advocate for. Like I said, we get data from NCHS that’s two years old. We get data from Defense Manpower Data Center month by month. So there’s quite a wide range in when things can be reported. We get data from IRS up to the current month.
So in terms of our inputs, the vital really does lag behind the other ones. That doesn’t mean everything’s right up to date. I mean, the American Community Survey data we use is a year old. But anything that we can do to sort of move that line forward would be helpful. We’d have to do less work with the county distributions, we’d have to do less projecting, and it would just make the estimates more consistent.
And the last thing I would say is there’s a real — the fact that the states didn’t switch over entirely to the new OMB race standards I think shows that there’s a flexibility problem in the system. And it’s not necessarily because of the states didn’t want to, or they were resisting; they just don’t have the funding, they don’t have the time. We heard yesterday, there’s 57 reporting districts, so there’s a lot of reasons why places couldn’t do that. But anything that we could do to make that easier would be beneficial, especially because OMB has been examining the race standards again, and if they were to make changes, we might be in the same situation we’ve been in for the last 20 years.
DR. ROSS: Thank you very much. Our final speaker on this panel is Corryne Carter, from the Department of State.
MS. CARTER: Thank you, and good morning. This has truly been an enlightening experience for me. While our business need at Passport Services is quite different from some of the other panel members, I’m encouraged because our challenges are the same, and the functionality that we are needing out of this new process and new system are the same.
Before I get into the topics of the day I just want to provide a brief overview for those who weren’t here yesterday about what I discussed yesterday. The Bureau of Consular Affairs Office of Passport Services issues U.S. passports to U.S. citizens and nationals. In fiscal year 2016, we issued over 18 million U.S. passports, and we anticipate issuing close to or over 20 million U.S. passports at the end of fiscal year 2017.
Vital records play an important part in our administrative review process, because they are relied upon to determine identity, citizenship, and entitlement to a U.S. passport, and that we value and rely on vital statistics offices and our partner organizations to resolve discrepancies and to verify and confirm authenticity of vital records.
In my previous life as a passport specialist, I was responsible for reviewing and making judgment calls on roughly 150 applications a day. I was responsible for preventing fraud, terrorism, child abduction, identity theft, and the list goes on. In an 8-hour day, minus lunches, breaks, completing refresher training, and other administrative tasks, that left me on average under three minutes per application to make the appropriate decision. With such a high demand for U.S. passports and a corps of roughly 1,200 specialized professionals performing this work every day, security and efficiency are the highest priorities in our process, and are always at the forefront of what we do.
Naturally, challenges arise between the two. Some include challenges around document security and counterfeiting, especially when older vital records documents remain in circulation. Data integrity and accuracy in the documents we review, changing state vital statistics offices, and our ability to maintain partnerships and communication, and access and timeliness of record for us and our customers.
Generally, in our line of work customers don’t need vital records until they need a service. So timeliness is really critical in our review process. And lastly, access is an important thing for us and also a challenge we face, in that a lot of our customers are not geographically located in the jurisdiction that their vital event is registered in.
As Passport Services continues to modernize and innovate and provide more convenient and digital service options to our customers, and become more secure and efficient, we support any improvement and are actively seeking processes and systems that will meet our current and future business needs, such as direct access and need-to-know access to vital records for verification purposes, electronic delivery or acceptable electronic certificate options for our customers and for us, increased adoption of the recommended standards for birth records, and as we become more digitally capable, introducing standards for electronic vital records options.
The natural phaseout of the circulation older and less secure certificates in formats no longer in production, faster registration or reporting processes for vital records to federal users, and the adoption of those processes. Our partners such as NAPHSIS are already doing exciting work in these areas. Not listed, but definitely most important to us is increased and hopefully full participation from vital statistics offices, and whatever system or process we come up with or comes out of this, it is only as valuable as those who participate in it.
DR. ROSS: Thank you very much. I appreciate all of our speakers staying to the clock. That leaves us now with a few minutes, so we have an opportunity for committee members to ask questions. Let me start by asking if our colleagues on our phone, if either of you have any questions you’d like to ask.
MS. STRICKLAND: I do not. Thank you.
DR. ROSS: All right, hearing none, we’ll start questions, Bruce.
DR. COHEN: Thank you all. I could ask everybody seven questions, but I’ll constrain myself for now. One observation around the — Social Security colleagues — around nuptiality statistics, the way I learned them, in public health school 50 years ago. I know the jurisdictions, most jurisdictions, still collect them in pretty significant detail. I don’t know whether you’ve tried to connect directly with the jurisdictions to get the data that you need.
MR. BYE: Yes, I did that specifically for divorce. It wouldn’t give me age, single year of age, they gave me age groups, and it depended on each state, the level of detail they gave me. But I was able to get something. We could try to do that each year, but that’s going to 50 states and outside areas, it would be very time consuming.
DR. COHEN: Perhaps you could work with NAPHSIS the way other parts of Social Security Administration’s work with NAPHSIS to address the economies of scale.
MR. BYE: That’s a good idea.
DR. COHEN: I had one other —
MS. WEBSTER: I literally just, two minutes ago, scheduled a meeting with some of your colleagues at SSA, not in the actual actuary area, but the other area. One of the other 40 areas. To talk about a national marriage data set and a pilot program with some of our states to do that. Which is super timely, because our board meeting, our NAPHSIS board meeting in October, is going to include a half a day on brainstorming about what that would look like, how we would go about doing it, what system requirements we would need, probably not that level of detail, but just kind of high level.
So this has been on our radar for years, and I think there’s probably, it feels like there’s some momentum here and possibly some funding, too, and some corporate partners who are willing to help out. So this is going to be an exciting new era for us.
MR. BYE: That would be great, I’d love to get involved in that. At least I’ll listen.
DR. COHEN: I had one more question for our CMS colleagues, for both Lakesha and Caitlin. I think it’s really fantastic using the birth data for evaluation and examination of the state Medicaid programs. I think it’s one of the potential most important uses of these data. I was wondering how CMS reimburses states for the multiple uses of the birth data.
DR. CROSS-BARNET: For Strong Start we’ve been purchasing the data from them, or reimbursing them for their time, or whatever the process is. It’s actually not clear whether or not we would be able to reuse it for any other purpose right now.
MS. DANIEL-ROBINSON: And so I think that’s just one application of the use of the data. In the work that I do, it’s really about supporting the states in their particular areas of interest, and so we have not supported the purchase of the data or any of those kinds of things. We provide technical assistance about how to use it and things like that, but not in terms of the reimbursement piece. And I’m not sure if there is — I don’t know what the role might be in terms of that reimbursement.
DR. CROSS-BARNET: And CMS is not doing it directly, right? It’s through a contractor. And so a part of the contract was to conduct this evaluation, we gave them a budget, the evaluation strategy was to link this data, so this contractor has gone out and done, under my supervision, but has been the one to do this work. It’s not like CMS is cutting the checks to the states.
DR. COHEN: Again, the issue around — of sharing the federal cost burden to support vital statistics is an issue that has come up in a variety of situations, and I think this is an opportunity to think about CMS and how they can support that.
MR. ATKINSON: I think this is probably one of those areas where, if we can somehow get beyond our individual programs, there is a tremendous need for the linkage of birth records, especially — probably also deaths — but especially birth records, at the state level, in a couple of different programs. Medicaid is one, HRSA is another, and so forth. If there could somehow be the synergy where those entities are helping to put in place the capacity within the state vital statistics programs to do the linkages and then do the analysis — whether that state vital statistics program uses a contractor or their own folks.
And I’ll give you an example. When I was at the state, we had a contract with the state Medicaid program, and that funded a couple of positions within the center, for which we then did the linkages and all kinds of analysis that was available, and it was not only for the Medicaid program, but for our maternal and child health program. So there’s an opportunity here that, working together, I think that we might be able to put some things in place that will help in building that capacity at the state level.
DR. ROSS: Thanks. That is very helpful. All right, we’ll go Vickie, Linda, and then Bob.
DR. MAYS: Thank you very much. I agree with Bruce, we could ask a ton of questions, but I’m going to kind of start. What’s the reason, Delton, that you all stopped collecting or compiling, or whatever the accurate way to say, the data on marriages and divorces? Because I’m very sympathetic to what you’re asking for, and I think it’s critical. I’ve done work in this area. It’s like, can we fix it, I guess, is the question.
MR. ATKINSON: Charlie was around then.
DR. ROTHWELL: When Eve got the apple —
This is actually a good example of the decentralized fiscal system, some of its benefits and some its negatives. When we had the vital statistics system, when it first began, we created birth registration areas, and death registration areas, and that was in certain cities, like New York City, and a variety of states. And as we built those up in the early 1900s, until we got to about 1935, we finally had complete coverage. We did the same thing for marriages and divorces later. And we started having states, encouraging states to collect marriage and divorce data.
As usual, agencies go through funding problems. And at the time, NCHS was looking at what it could do to cut. We were looking at our surveys, which were doing very well, and they were collecting covariate information on marriages and divorces. And in fact, we are a health statistics agency, even though we have great demographers within vitals. And so the fight went on, and we decided that we really weren’t a demography organization, officially, and so we decide that marriage and divorce should — would get the cut.
I would also like to say that we had not full coverage of either event, and we had never really funded it appropriately to build up the states to be even close to what the birth and the death registration was. So, if we were looking at it from a national perspective, we would really need to invest more money in marriages and divorces, and we knew we weren’t going to get that funds; we had to lose it.
So that was a long explanation of why it was that we walked away from marriage and divorce, and we do collect marriages and divorces from states, but just counts. And of course we can’t verify what those counts are. But we greatly appreciate the states providing that information to us.
DR. MAYS: Let me just ask, then, do you think that NCHS would be the home if this, if we were able to recommend or reinstitute? Or is there another agency that you think should do this?
DR. ROTHWELL: I think Census could do it. I think we could do it. There would be a variety of agencies that could do it. The critical issue will be, especially on divorce end, and I will get into all the problems with divorce registration and how it’s handled or not handled, is at the state end, you’re going to have to build up capacity at the state end to do that, so it’s going to be a fairly significant investment up front in order to get that going wherever — so it’s the state end of things that are going to be the most critical thing, not the federal end.
MS. GASTON: Just to echo what Charles was saying, we also cut marriages and divorces. We don’t report on them anymore, and one of the additional challenges we had is with same sex marriages. We stopped getting — we got spouse one, spouse two. And so, it made it really challenging to actually report on the data. So that’s perhaps something to think about when you’re considering your next steps.
DR. MAYS: I just wanted to ask one more question, and then I’ll let go of the other, to Ursula. One of the things you talked about in terms of wanting more information about the underlying causes of death. What I haven’t heard discussed too much here has been the issue of the use of narratives, like in the National Violent Death Reporting System, they actually use narratives. It really helps a lot. I know it would be expensive, but do you have some sense of what you’d like to see in order to get that extra data? Would doing narratives on all deaths — I mean what — do you have an idea?
DR. BAUER: Yes. We are with our national program of cancer registries and with some support from PCORI, looking at how to analyze language, freeform language, in a cancer report or in a medical record. We’re trying to develop that out, and that would potentially be an application. There is, of course, training one can do in terms of the death certificate if we can explain, what are the critical things that we need to know as we try to prevent these events.
You have the immediate cause of death, the underlying cause of death, you have contributing causes of death. If you list lung cancer there, and it’s a heart attack, and you’re getting chemotherapy, we can put a picture together a little bit better than if we just have MI.
MS. KLOSS: Thank you. Like Vickie and others, I have many questions. I want to just ask for clarification from what we heard yesterday from Valerie. Does the United States have anything like the coroner and medical examiner database? Is that something we have and if not —
DR. COHEN: Coroners —
MR. FUDENBERG: No.
MS. KLOSS: Okay. Is that anything that’s been discussed? It’s just —
MR. FUDENBERG: No, but it should be.
MS. KLOSS: Okay. It seemed like something that was valuable to Canada and why wouldn’t it be certainly to the United States, so I wanted to know if that was something we should put on our list.
MS. WARNER: I am Margaret Warner from the National Center for Health Statistics and I work for Delton, and it is something that we’ve talked about quite a bit, and you mentioned NVDRS, which is the National Violent Death Reporting System, which gets information from medical examiner and coroners, but actually goes out and actually goes to every single office to get that information for NVDRS.
There’s also the SUID, the Sudden Unexpected Infant Death Case Registry, which does a similar process, there’s a Sudden Death in the Young Registry that NIH funds that does the same process. So there are many people seeking data from medical examiners and coroners because they certify deaths that occur outside the hospital, outside of medical settings, which are actually quite a few deaths and very preventable deaths, potentially, and of young people, often, so drug overdoses is a way, a reason that we’re actually looking more carefully at medical examiner coroner databases.
The other thing that I wanted to say is that somebody mentioned Canada yesterday. Australia had a great system for quality of information on death certificates, and they actually have quite an elaborate coroner and medical examiner database, even more elaborate than what Canada has. We had an interchange a couple of years ago about how they collect data from medical examiners and coroners, so I know a bit about their system. Australia’s system is even more advanced, and it integrates with the vital statistics system to get higher quality data on death certificates. So, it’s an option.
MR. ATKINSON: And that is why I think that any report coming from this committee has to talk a little bit about the medical examiner coroner systems and the need to strengthen that, because that is going to be a tremendous aid to vitals.
DR. PHILLIPS: This has been like a master class for me. This is incredible. Delton, I’m struck by the fact that you started talking about how PCORI and emergency preparedness have really made some strategic investments that have brought down NDI reporting times in particular.
And we moved to CMS which has made strategic investments in contractors who have helped with making data connections in states, but it’s not a direct investment in the vital records process, and then, Bill, I’m sorry, I don’t want to pick on you, but NIH spends a lot of money on institutions, but not a dollar gets out to the coroners and medical examiners who we rely on to collect these data, so it’s almost like a tragedy of the commons.
I mean, we have this incredible, common, valuable system that has no — none of the stakeholders have — at least most of the stakeholders don’t have a direct investment in the process, and it sounds like that may be one of the biggest hurdles we have to tackle. I was struck last week, I was at the Board of Scientific Counselors, thank you, for NCHS, where Nancy Potok, the new chief statistician, said that the federal system is unsustainable. She was talking very specifically about our survey systems, but it really applies, I think, to the entire process that we’re talking about. So I wonder if there’s a way to bring stakeholders into the conversation about how do they support and get the value they need from this system, so that there’s a negotiation for building up the vital statistics process we have.
DR. ROSS: I second your comment. It is sort of the challenge of the commons. We’ve heard yesterday and today this enormous benefit that so many different people accrue, but yet over time, we’ve created this very curious way of gathering data, creating different surveys, and I’m sure if I were in the states’ shoes, I would feel like there’s piling on, here.
One group after the other is asking us to do things, or if I’m a coroner or a medical examiner, hey, can’t you give us this, can’t you give us that? And who pays? This is one of the things I think we’ll have to put together.
Ursula, you had a comment?
DR. BAUER: Just a quick comment here. Putting my former State Health Department hat on, here, which is that states are one of the users, and they are a huge user and have a keen interest too, and public health at least is primarily a state function, so they are one of the payers in the system, and we just shouldn’t lose sight of that as we are trying to make sure that every other user is contributing as well.
DR. ROSS: Nick, last question.
MR. COUSSOULE: Last question, that’s a hint, I think. I will try to be brief. I guess, I’m interested in — we’ve heard lots of comments about the data accuracy and completeness struggling, and we’ve also heard lots of comments about we need more data, more and more data, and those two are naturally conflicting.
I’m a theoretical guy, I believe you can do both at the same time, but I’m trying to understand the dynamic between, do we approach data quality first, or do we approach data volume and completeness first, and how do we determine the value differential there? Does that question make sense? And it’s probably directed at just about anybody sitting here, frankly.
DR. HUGHES: I think the way NCHS is organized, as Delton explained, is to look at timeliness first and now start to look at quality. Not that there haven’t been efforts on quality, but from a jurisdiction’s perspective, we need a complete death certificate to be able to handle all of the registration things that need to happen. So, in that sense, completeness is number one.
I think what we’re recommending in terms of quality, it has to be problem driven. So the needs for tackling the opioid epidemic, in terms of quality, are very different from the needs of emergency preparedness in terms of quality. And until we have that focus, it’s hard to explain to a medical examiner or coroner or to a physician, this is why what you’re doing is important, and so that’s my view on that.
MR. ATKINSON: That is a question that we have wrestled with quite a bit. When we started the issues, the focus around timeliness, and there was that debate, should it be timeliness, should it be quality, should it be completeness, and we sort of ended up just saying, well, you know, if we don’t ever do something about timeliness, it really doesn’t make a difference about the quality. If I’ve got the highest quality system and it’s producing data that’s five and six years old, nobody’s going to be interested in it. So that was more of a strategic decision, but it was not one to say we’re only going to do quality, but we just started there and that’s where we got funding to help us move in that area.
DR. COHEN: Bill wants the last word.
DR. RILEY: I don’t want the last word, but I’ll also say the same thing is true. I think from the research user perspective that timeliness is probably more valuable than quality. The issue is a metatag to tell us what the level of quality of that particular data point, but as long as we have that, we know is this provisional, is it not quite complete yet, is it finally curated? Those kind of things are very useful, but more critical is, do we know the person is dead or alive, and we can move from there after that.
DR. ROSS: This has been, yet again, another very interesting panel, thank you very much. We are going to start on time, which would be 10:30, so I ask you to make your break quick. Thank you.
DR. ROSS: I was wrong. I said we would start promptly at 10:30. It was aspirational, right. Time is a fluid thing with me.
So we’ve just had a panel on the federal perspectives. Now we’re going to move to state and local perspectives. It’s another group of very informed panelists. I’d like to start with Dr. Lou Saadi, from Kansas Department of Health Environment.
DR. SAADI: Thank you very much for this opportunity to speak from the local perspective. You have heard a lot of the challenges that we have, and I don’t want to belabor a number of these. I do just want to hit a few topics that have become apparent, and I’d like to also make sure we have plenty of time for our local officials that have come, particularly in the pathology area, where we’ve certainly got to hear their perspectives on what we do here.
Dave, I wanted to just share, you’ve given a really good analogy about vital statistics being a foundation. I make it a little more like electricity, and from a state who went dark in August of 2010, it’s emblazoned in my brain, about how our whole agency had a big failure in our servers. We went dark for the whole agency for 30 days.
So we depend on these electronic systems. What if we couldn’t get them? And it was very apparent to us on a daily basis of how important the records were, and our leadership got to see how important the records were and the information that was made available. I mean, people were having grants due and were digging the papers out of recycle bins. So it was really a tough time for us, and that was school enrollment time, and everybody needed their records for their jobs.
But the point being is that we’re dependent on these electronic sources and our resources within our agencies, and we have to recoup some costs, because what turned out to be the case for us was improper backup, improper attention from our state ITs, on how we properly manage the resources. And we’re extremely challenged in our state health agencies. We have a lot that is put on us, and when you have things like a governor’s fitness program that takes priority over assuring that you’ve got proper backup, there are problems there. So we definitely have some areas to work on.
So I want to make just a couple of comments. Many of us have talked about, well, you need to update your laws. And I just want to give you — well, it sounds easy, but it isn’t. And I just want to give you a flavor for what some of us have had to endure in trying to do that.
Some missed opportunities, for example, like Real ID. I probably used that for five years, hoping that we could use Real ID as a way to boost our security in our vital records offices, to improve the way we capture information, issue the information. There were just so many benefits that it would have provided some significant underpinning for us in the direct civil registration process. And whatever’s going to help civil registration is going to help the vital statistics systems, there’s no doubt.
But that faded on us. It’s an example of how we were not allowed to — we had to stop talking about it. Because it didn’t come to fruition.
Adoption of the model law. That’s 2011 that that came forward, and even though we have been promoting it at NAPHSIS, it’s one of our core products that we offer jurisdictions and we support them in any way that we can to try to get jurisdictions to promote them, promulgate them, we’ve only had a handful of us that have been able to do it. I’m proposing it; it’ll be my third time.
But things like same sex marriage. We know you all want the statistics, but if you’re in a red state, it’s not going anywhere if you propose that. And we have to have it. It’s the law of the land. But how do we get it on our birth certificates?
Now, just to give you a flavor of the challenge that gave us, it wasn’t all about marriage. It all became about parentage as well. It filtered through the entire records issuance system and the statistical process. You had to add sex to the marriage and divorce systems. We didn’t have that. You had to add age. You had to add a number of things. So there are difficulties with that.
We also have a lot of concern for unintended consequences. You always have your partners, but the situation is those of us who are managing these systems and managing the civil registration, it’s up to us, it’s up to the jurisdiction to promote those laws on civil registration and vital statistics collection.
It can’t be the funeral directors’ association, who are great partners. Our EDRS sailed through with only one no vote in 120 house members. Who gets that? I mean, nobody gets it. But, when the industry helps us we can get it. But some of these things have to be promoted as state policy. And that’s difficult. So you always get, we don’t want to open that up. So then it gets denied from out of your agency.
So those are examples. We also have to think about our partners that manage the parentage act. We need them to help us promote that information, and get that updating their statutes. I think you’ve heard, we definitely and desperately need support from the federal level in supporting us getting our resources. It is mentioned, and I get it, that states are responsible for using this information. And we do use it for all of our programs across public health, but that money, only by relationship building do we get some of those dollars. And you can’t tell me there’s not a line item in many of these grants for data. But it doesn’t find its way to us. So, again, it’s about relationship building, and many of us have those but many of us don’t.
We also need for NCHS to have the capacity to be more flexible in their ability to address some of the changing needs within their IT systems. I manage a couple of surveillance systems, in infectious disease and syndromic surveillance, and the capacity, the technical capacity, they have and the flexibility is phenomenal. The amount of information coming out of those systems is really beneficial to our programs.
We’re configuring our HL7 messages or our processes to a checkbox. Form. We need to find ways to support NCHS to look at receiving a feed from an EHR based on trigger codes, and those of you involved with surveillance will know what that means for us. But then the problem is, for NCHS, is that it is a federal statistics system. A federal statistics member. And they have to be able to give you all data from many years ago, on a trend, the idea of things changing is difficult to map to. So they are boxed in in being able to look at the ways that the data can be sent to them in a more timely manner and utilizing some of these capacities that the country is building.
We definitely missed out on a lot of the HIT initiatives and the EHR initiatives, and we certainly want to catch up on those activities.
The last thing I’d like to share is, the opioid crisis is one of many. Many of those of us that have been in this area have dealt with the condition of the day. But it certainly has raised for us attention and some significant deficits in our system, and you’re going to hear about some of those, and some unique ways, and the ways that our professional partners capture this information.
But it certainly has highlighted areas that we need to focus. But think of other ways. But whatever we can do in these crisis times will support the system. But we frankly need to work hard at finding how we can get federal support. I need to be able to tell my leadership, this is what is needed by the feds, even though we have our members that some of them pooh-pooh, and they’ll not be able to comply. But that doesn’t mean, like Tara mentioned, the rest of us can’t proceed and innovate, because they’ll come along eventually.
And then the idea about, okay, it took us so many years to adopt the new regulations, I mean, excuse me, the new race-ethnicity provisions. You know, changes get made, but it costs us money to change our systems. And I don’t want to make it sound like money is all the answer, there are some policy changes that could help go a long way as well. And just the oomph of the federal support will be very helpful for us.
DR. ROSS: Thank you very much. Our next speaker is Terra Abrams from D.C. Department of Health.
MS. ABRAMS: Good morning. I am Terra Abrams, District of Columbia Vital Records Registrar.
The D.C. Vital Records Division has been working to modernize laws, operations, and technology. As we have discussed on yesterday and this morning, the ideas and solutions are vital to improvement of the health outcomes of our citizens. D.C. has faced challenges but we have had many improvements, many of which were supported by partners like NCHS and NAPHSIS.
First, let me give you a quick snapshot of D.C. vital records. My staff of 15 has extensive experience and expertise in the vital records field. Together we annually serve approximately 50,000 walkup customers in our fulfillment center and 17,000 online, mail, and phone customers. Our customers request certification, modification, and registration services for birth, death, and domestic partnership events.
D.C. vital records has had electronic death registration since 2004, and electronic birth registration since 2009. Our roadmap to improvement has included preserving and digitizing our paper records, which date back to the late 1800s, upgrading our electronic registration systems, training our data providers, sharing data reporting timeliness and quality performance metrics with our data providers, and proposing to modernize the District’s Vital Records Act of 1981.
As we move down this path, we are looking forward to completing significant technology improvements to our EDRS and electronic fetal death registration system, strengthening our technical assistance and quality assurance activities in the field, and increasing our focus on birth and death record data quality. These initiatives at the state level are often supported by funding and collaboration from external partners.
As Delton will tell you, there is a clear correlation between this type of support for vitals and performance improvement at the state level. In D.C., we have been driving data provider performance improvement through training, system improvement, and measurement.
It was mentioned yesterday that state VROs use funding to address deficiencies that may be unique to their jurisdiction. We often discuss barriers to success with our data providers, and universally they cite aged technology, lack of training, and redundant processes as issues. We have taken this feedback seriously in D.C. as we realized that our success depends on the quality of service we provide to medical facilities, funeral services providers, and constituents. We have also expanded the conversation with data providers to include facility and, sometimes, individual-level performance reporting.
For our medical facility administrators and funeral services management, this has been extremely helpful in keenly addressing data quality and timeliness issues.
The previous slide illustrated our improvements in death reporting timeliness. We expect death record reporting performance to continue to trend in the right direction as we upgrade our EDRS this fall. On this slide, I have provided a snapshot of our birth record timeliness improvements. In April of this year, we upgraded our EBRS, taking care to address many of the concerns reported by our data providers. I often tell my staff that the best way to understand what the issues are with vital data reporting is to talk to the data providers. Surprise, surprise.
D.C. is working to continue these improvements. As our attention to reporting timeliness has realized great efficiencies, we are now shifting some of our focus to data quality. And so, as I said, you see here at this end, calendar year to date, 2017, our improvement in birth record reporting within five days, and we truly believe that is connected to our improvements in our system.
This meeting has had a common theme. There’s a need to educate data consumers about the value of vitals data, and assessing and addressing the risk to inaction. I won’t restate the obvious. We are all clear that vital records are one of the best sources of data to inform our public health programs and policies. We are also all clear that our vital records offices face resource challenges when attempting to address dated vital records laws, aging or obsolete technology, aggressive counterfeiters and fraudsters, and data providers who may never have been trained on how to complete a birth or a death record accurately.
Our challenge is now to spread that message. Education for our data providers and consumers will be key to advancing our shared objectives. Additionally, federal support by way of mandates or policies, and cooperative agreements, will aid states in appropriately prioritizing modernization and enforcement of vital records laws. Increased funding will support system and resource improvements to increase the feasibility of meeting these objectives. The risk to inaction is clear, as are the timeliness and quality deficiencies within our data.
What’s next in D.C.? Well, we have quite a bit going on. We are leveraging technology in order to refocus staff toward back of the house data provider support and data management. In our office, most of our resources have been focused on fulfillment and we anticipate that our upcoming technology improvements will help us to balance the focus.
Additionally, we spent two years drafting a modernized vital records act for D.C. which brings our statute to close alignment with the Model State Vital Statistics Act. The bill is scheduled for D.C. City Council hearing in late October and we are super excited.
We are continuing to leverage our partnerships and networks to share best practices and apply innovative concepts within our vital statistics system. This audience and the opportunity to share our challenges and needs is exactly the type of collaboration and support that is needed to move the needle.
I truly appreciate your time and commitment to the work that we do. Thank you.
DR. ROSS: Thank you very much. Our next speaker is Dr. Mary Ann Sens from University of North Dakota. Mary Ann, I’m from Wyoming.
DR. SENS: Close, very close. One of those states with more senators than representatives.
DR. ROSS: That’s right. One of those states with many more cows than people.
DR. SENS: Absolutely. Okay, I thank you very much for inviting us. I am the executive vice president of the National Association of Medical Examiners Forensic Pathologists. I also have a tristate, 49-county rural and frontier medical examiner system that’s accredited. So I can share some of the rural experiences.
We’re going to tag team, and I’ll introduce Dr. Davis.
DR. DAVIS: You want me to take it from here? All right, I am Greg Davis. I’m the chief coroner medical examiner for Jefferson County, Alabama. That’s the county that Birmingham resides in. I also have an academic appointment. I’m professor and division director of the forensic division in the Department of Pathology at the University of Alabama at Birmingham.
So we are going to answer the questions that you asked of us and we’ll start with mentioning some challenges and barriers that exist. We’ve already heard that death certification is an imperfect system. I don’t want to belabor that, but there are several aspects of it that haven’t been mentioned explicitly yet.
Terra actually just referred to one of them. That is that many clinicians have never been formally trained or have little formal training in death certification, and yet they’re completing 80 percent of death certificates. Only about 20 percent of death certificates are completed by coroners and medical examiners.
Now that’s an important 20 percent, but 80 percent are completed by people with little or no formal training.
But death certification is best thought of as the practice of medicine. So good death certification, we’ve been talking about how cause of death are determined and written down, but we haven’t actually talked at all about the process of making that determination as though somehow it were a simple matter.
Occasionally it’s simple, but often it’s complicated, and it’s done according to medical paradigm where you take medical history or narrative and you combine that with the physical findings that we see at autopsy and the laboratory results such as toxicology analysis, then pathologists put all that information together.
Pathologists are trained not to just make some sort of wild determination. That’s not the pathology approach. Pathologists are taught, dare I say, it’s drilled into us, that you make a diagnosis that you can defend and show to other people with compelling evidence, evidence-based diagnoses. That’s how it is that pathologists are putting the cause of death on death certificates. So you can be confident that that is correct, but if somebody is just putting down what their Ouija board told them to put down, it’s not going to necessarily be as useful information.
The forensic pathologists who are practicing in this country you can expect to be boarded by the American Board of Pathology. So they will be certified in their practice, but beyond that, there’s no requirement for certification or medical training, so many death investigation offices in the nation do not have that.
Also, accreditation of offices, Mary Ann’s is accredited, mine is too, but that’s voluntary. It’s not required that we do such a thing.
We spoke yesterday about the patchwork nature of our federated system. This simply shows that network for the states representing the 57 registrar jurisdictions, but you should know that there are actually more than 3,000 death investigation jurisdictions. That would be the various counties that are performing these investigations. Then within my own county, one county, I deal with some two dozen or so different funeral homes. So there is a bewildering number of different agencies and computer systems that are pouring data into the death certificate, way more than the 57.
DR. COHEN: Greg, excuse me, in that map you showed, I can’t discern what the colors stand for.
DR. DAVIS: Oh, actually, the importance that I wanted to show is that some are coroner states, some are medical examiner states, some are mixed, coroner, medical examiner. The point is that it’s a patchwork.
DR. SENS: I will further talk about some of the challenges and barriers. I’ve got a lot listed here. There are more in the notes and they fall into really probably three major areas.
The first one is garbage diagnosis and garbage questions. Don’t ask me on a death certificate, did tobacco contribute to the death? Ask me if he was a smoker or an ex-smoker. The tobacco, the pack years, the type of tobacco he used, and that’s data. The rest is opinion, not that my opinion isn’t important, but data exceeds that.
The second major conglomeration is the data, the form that it’s in. As an example, time of injury. For one of the states I work with, I enter a time, I assume it’s a time field. For the other state, I have a pick list, series of times, and then an unknown. Well, most pick lists, I don’t know much about IT, but most pick lists go in as text, especially when there’s text in the pick list. In my own system, when I enter time, that goes into a time field or it goes into a fully text field. Right there, that simple question that we all know as time, has three incompatible transfers from the IT.
The last group of complaints from this area is really the lack of user-friendliness. Obviously, user groups did not get going when these systems were set up. As a user, I would like to know how many pending cases did I pend that I have to go back and change? I can’t find that out readily. As an administrator, I want to know which of my docs are slackers and not getting them out. I can’t tell that. The systems may work well at the state level, they may interface at the federal, but they’re not friendly to the ultimate users.
So what are some potential solutions? We’ve talked about many of these. We’ve talked that federal agencies should obtain data from the vital registrars and that’s probably, right now they’re getting it for free and we’re looking at models for that, but I think you need to, as other speakers have pointed out, we’re the ones with the data. We’ve got the meat, as Wendy’s commercials would say.
We’ve got tremendous amounts of data. We’ve got much more than is on that death certification that we want to share, that’s good for public health. We need to be able to give that to you.
DR. DAVIS: So one recommendation that we make as a potential solution has already been stated. We’d reiterate it, simply that, and this is how I think of it, public health should move beyond the electronic death certificate that currently exists. In my mind, it’s simply a glorified paper statistic. It’s really not doing anything other than paper did except that you’re getting your data entered into a field and you don’t have to have someone manually enter it yourself. We have to manually enter it for you, but it’s really no different than a paper certificate.
A system that, Charlie Rothwell mentioned it yesterday, if we’re on the cusp of developing a system, the technology exists where we can have data from the various databases being put into the other databases, share the data, and that’s part of what it is that our offices have the responsibility of doing. We’re happy to do it.
When the time comes to pilot such a system, and I hope it’s sooner rather than later, you’ll find that medical examiner coroner offices, at least some of them, I can’t say everyone, but you’ll find many who are willing, even eager, to get involved in piloting such a project because we see its usefulness and know that that’s the proper future for us. That will require, as has been said, some standardization of terms and avoiding duplicate fields and whatnot, but it can be done.
Another thing that was mentioned yesterday was the possibility of having death certificates or all the information entered into this electronic death certificate and some given to the families for their legal need for a fact of death statement, but other more sensitive health information such as, for example, that it’s a suicide, families are embarrassed to have to show that. We get a lot of pressure to change that because they don’t want anyone to know it, but what happened is what happened. That’s what we’re going to say. But families would be helped if they didn’t have to share this and you kept that information separate from the certificate they need to unfreeze accounts and to move on with their lives.
Do you have something you want to add?
DR. SENS: Yeah, a lot of that could also be used for pending toxicology. It could be used, you could say, drug overdose and you don’t mind showing that to your minister, your bank, you don’t have to say heroin on it. You would actually have more information on a death certificate if it went that way. If we went to an expanded death certificate, it’s similar to a fetal death certificate. We could ask about linking health records into that and get good information.
Finally, for a selected group of deaths, and this would take and complement some of the federal systems that are out there, the sudden death in the young, the national violent death reporting system, but it’d be very robust and for, say, drug deaths, or for suicides in teens or suicides in recently discharged military, or any number of things that we identify as public health professionals that we want to look at, or that commercial entities, drug companies, want us to look at and pay appropriately. We could actually expand that information gathering in a confidential manner, use it for our own planning purposes, or deliver it to the people willing to make that data.
Every clinical trial should probably have a mandated autopsy and that information entered in and that’s part of the funding.
So for challenges and opportunities, we can go business as usual, there will be a delay in response, a loss of data, if we don’t have these coroner and medical examiner information systems that are transparent, that are readily available, if it’s not backed up by accreditation, we will continue to delay the public health response and you will lose data. There is 90 percent of the data that’s in our office you cannot have right now. We would love to give it to you. It’s lost.
DR. DAVIS: It won’t fit.
DR. SENS: It won’t fit.
DR. DAVIS: Then I want to add one other statement. So medical, legal death investigation systems provide you a foundation of information that’s necessary to all the enterprises we’ve been hearing discussed, but there’s another foundation that hasn’t been studied explicitly and we thought that it was worth mentioning. That is public health actually fights two battles. There’s the obvious battle that everyone thinks of, the battle of trying to stop disease, stop an epidemic so that we promote better public health.
But there’s a second battle and we’ve gotten to see it with the opioid epidemic. That is the battle that public health has with people and what they think does or does not constitute disease and what does or does not constitute the cause of the disease.
If we had believed the data that were coming forward about the number of opioid deaths 15 years ago when it began, we would be in a much better shape now than we are because we waited as a people until it was much, much later. That’s a difficult problem to overcome that’s inherent in human nature, but I think it’s important to mention that that is a battle that public health fights all the time.
DR. SENS: All the time. The opiate epidemic and certainly in our office, I think John would agree, we knew about this and started saying this in the mid-2000s. I know in my jurisdiction, I went before the county council in 2009. They didn’t believe me. We, NAME, wanted to go to the AMA, very unpopular, and we were told the biggest problem with opiate addiction is you guys write it on the death certificates.
DR. DAVIS: In closing, I’ll say NAME, which she mentioned, is the National Association of Medical Examiners, and finally, both Dr. Sens and I thank you for inviting us here to share our experience.
DR. ROSS: Thank you very much. This is really very, very helpful. It’s kind of a staggering comment.
Our next speaker is Mary Beth Kurilo from the American Immunization Registry Association, AIRA. Mary Beth, great to have you, please go ahead.
MS. KURILO: Thanks so much and thanks very much for inviting me here today. So as mentioned, I work with the American Immunization Registry Association and we’re a member organization for 64 IIS jurisdictions. Yesterday you heard from Rebecca Coyle about the ways IIS benefit and use vital records data. So today I’m going to talk more directly about some of the challenges within the immunization vital records relationship as well as some of the potential solutions for consideration.
So I just want to start by echoing Rebecca’s comments from yesterday about the strength of the relationship between IIS and vital records. It’s longstanding, very solid, and in a vast majority of jurisdictions, they’re exchanging data. There is of course always room for improvement, but I think it’s important to underscore that that strong foundation exists. I also want to mention that much of the information I’m going to share today was actually sourced through an informal survey of AIRA members across the IIS community. So thanks to everyone at the state level who provided input and information toward this effort.
So, challenges. Perhaps the most challenging area for this exchange is in the realm of policy and that was mentioned yesterday. Not only do statutes and laws vary across jurisdictions, but the policy tools used to implement the statute or law vary as well. These tools could include data use agreements, memoranda of understanding, interagency agreements, or other policy vehicles that define the terms and responsibilities of data exchange.
One example of this variation is that in some jurisdictions, following an adoption, immunization information is required to be deleted or removed from the IIS as a safeguard to privacy. However, this also destroys that valuable consolidated record that has been built, often over years. A uniform policy approach might allow both confidentiality and clinical needs to be met by securely preserving the consolidated record under the client’s newly adopted name.
So an additional challenge is a variety of specific data received when comparing across jurisdictions. Eighty-six percent of IIS reported receiving birth demographic data while only two thirds received death data and only 50 percent received birth dose and adoption data. So completeness across the IIS network of systems will vary substantially.
Similarly, completeness within a jurisdiction can vary if states are relying on individuals to opt in to data sharing. It’s also important to recognize that as of 2016, 14 IIS weren’t received any vital statistics data in their jurisdictions.
The timing of data submission varies broadly as well. Increasingly, electronic health records, or EHRs, in hospitals and health systems are set up for real-time or near-real-time exchange with IIS, which means that an EHR report of a newborn or of that newborn’s shots may make it to the IIS before the vital record report.
To the extent that we can consider vital records to be the gold standard report, we would want those data to reach the IIS first to populate that demographic record with the most accurate data.
So, of course, funding is a significant challenge across public health, and there are a few areas where it directly impacts the vital statistics IIS interfaces. There is no dedicated funding source that funds the exchange of data between these two public health programs. It is still prioritized across systems, but a direct funding stream that supports a base level of exchange does not exist.
Vendors on both sides of the exchange fence may be operating with limited visibility. As a result, even when vendors are paid to make changes to accommodate vital records and IIS data exchange, requirements may not be articulated or documented clearly. Finally, routine development work may delay or break working interfaces on either side of that data exchange, and resources will then be needed to research the problem and to get the interface reinstated.
So it was interesting to note that when the data came back from our informal survey, several IIS called out the need for funding, not for their own programs, but for their vital statistics programs in their jurisdiction.
Yesterday several of the speakers spoke of the need to address data quality within vital records and these quality issues are compounded by a lack of standards to support the data exchange. We don’t currently have uniform standards for the capture and submission of vital records data for IIS. This affects the data quality coming in.
Some examples where this can negatively impact data quality include, in some jurisdictions, Hepatitis B dose is reported as a yes/no variable, and lacks any of the information needed, manufacturer, lot number, to actually record an administered dose in the IIS. Placeholder information, like baby Smith or twin one and twin two, can create false merges if it’s received. So that’s often rejected upon import.
Finally, many data exchange processes themselves are heavily manual or outdated, using flat file or SFTP exchanges when they could be leveraging more current data exchange standards that support data quality.
A final challenge before I move into solutions is about the sensitivity about vital statistics data and the essential need to be thoughtful and careful in our approaches to data exchange and data use. In the IIS world, we recognize the need for rigorous confidentiality and privacy, and it’s critical we protect the sensitivity of adoption records and that we track the records of deceased clients to avoid any inappropriate outreach following a death. So close collaboration between these two public health programs is essential going forward.
In terms of proposed solutions, the one that would move IIS and vitals forward the most swiftly involves the development of standards, particularly regarding the message format and transport method of vitals to IIS. Ideally, a common transport method, like SOAP web services, could help with prompt reporting and data exchange and improved timeliness. Likewise, a common messaging format, such as HL7, could improve quality and consistency. I was thrilled to hear Delton mention that this morning. We’re big HL7 devotees across the IIS network.
We’d also love to see capture of common data elements across all jurisdictions, including birth dose, hep B, and HBIG. Finally, a suggestion has been made to consider how to better incorporate data feeds from EHRs to feed into both IIS and vitals birth data in an effort to minimize double data entry and increase quality.
The second potential solution involves automating the process of data exchange between vital statistics and IIS. We believe that investing in automated interfaces with limit human error, decrease resources needed to sustain interactions, increase quality, and certainly increase timeliness. We also think there are opportunities to explore synergy in the exchange of data across jurisdictional lines. Both vital statistics and IIS have a need for interstate exchange and so we want to continue to explore opportunities to leverage the same pathways and policy approaches.
Opportunities exist to improve accurate data capture, transfer import and use through training and knowledge transfer across vital statistics and IIS programs. This extends out to birth centers and hospitals where the data are being gathered since this is really where good data quality begins.
For cross training and knowledge transfer to be sustained, leadership must be engaged and committed on all sides. Finally, we recognize that this isn’t easy to come by, but increases in dedicated funding towards vital statistics programs but also towards IIS could potentially support this work on both sides. In particular, funding that is made available and dedicated to pursue standards development may incentivize some jurisdictions to tackle the policy, technical, and operational issues that need attention.
So what happens if we do nothing and we just continue with the status quo? On the plus side, most jurisdictions have a good working relationship with vital statistics partners and those agreements that are being individually negotiated would likely continue. However, we continue to live with our current state of variation and ad hoc processes. These manual processes will eat up more resources than they need to and will adversely affect data quality, timeliness, completeness, and accuracy.
Those IIS not currently receiving vital statistics data will likely continue to not receive data unless a new champion arises to take up the cause. Ultimately, this will mean that as a community, we’ll fall short of our full potential for leveraging vital records data within IIS.
So in closing, I just want to reinforce three key themes that we hear repeatedly from our IIS community. The relationship between the IIS community and vital statistics is strong and foundational across public health program. However, opportunities exist to increase standardization, automation, and modernization. Working together to leverage investment across our nationwide networks would save resources, improve data quality, and bring benefits to the populations we all serve.
So thanks very much for the opportunity to provide testimony and I’d be happy to answer any questions at the end. Thank you.
DR. ROSS: Thanks, Mary Beth. It is an interesting comment, we’ve heard this theme repeatedly. I hope our vital records colleagues feel supported. You have so many people saying you need more money, a lot of advocates on your behalf.
Our next speaker is Amy Zimmerman from Rhode Island Executive Office of Health and Human Services. Amy?
MS. ZIMMERMAN: Thank you. Good morning and thank you for having me come before you today. What I want to do is really focus on some things that have not been said because a lot of my slides will cover things that have been said. So I’m going to go through some of the slides pretty quickly so I’m not redundant and where there may be a few nuggets that have not been touched upon.
First slide, and if you get these, you can look at them later, just gives a little bit of overview of Rhode Island. I will say from the beginning that we are one of those few states that do not have an electronic death registration system. While we have been a leader in a number of IT initiatives, this is one area where we are not clearly leading and are well behind.
Challenges and barriers I’ve sort of broken down into three areas, infrastructure and interoperability, understanding the value of the data and potential uses, and I think for those users, and I’ll talk about this in a minute, there is a great appreciation, but I think there are a lot of individuals including legislatures and others that do not appreciate how wide and how far the data is used, and I think that’s an area that can really use some thought in recommendations about how to get the message out to non-users, how important and where the data is used. Then of course funding and I’ll talk a little bit about that.
So challenges around the birth data information, again, a lot of these have been mentioned, a lot about the EHR not interacting with our electronic birth certificate data, so data double entry, prenatal records not getting to hospitals in time. Again, in our state, only a few of the electronic interfaces of where the birth data goes is done through files transfers. So that’s not sort of real-time data either.
Enrolling newborns in Medicaid is a big issue. I don’t know whether it is for other states, but it’s a timeliness issue, I think. In fact, in our state, because of enumeration in birth, because babies aren’t always named — let me back up. In our state, mothers that are on Medicaid, their newborns are automatically on Medicaid. So it’s really important to very timely get that information to get those kids into Medicaid. I haven’t drilled down, I wasn’t able to drill down to all of the issues that are happening there, but there are a number of issues that complicate that.
In fact, what I was told by one of our Medicaid colleagues was that the hospitals, for proof of birth, actually just get a list from the hospital as a sort of agreement while they’re waiting for more formal birth data. So I’m just sort of pointing that out as an area.
Also, what was raised to me by our vital registrar was the need for certificate of parentage, particularly for same-sex marriages. In our state, it is law that you can’t get on the birth certificate unless it’s a full adoption or an affidavit of paternity. So same-sex marriages, there are some challenges there around parentage.
We have 39 cities and towns in Rhode Island despite how small we are and there’s one system for them to issue birth certificates and it is web-based.
So for death data, I think basically the fact that we don’t have electronic death certificate is the biggest challenge for us. No single unified death registry, so not the National Death Index, but the fact of states changing, I know there’s work through NAPHSIS and stuff even other to do that. Again, ability for timely surveillance. In our state, by the way, fetal deaths, all fetal deaths, including under 20 ages of gestation, so there’s a lot of underreporting and lack of desire to report. So that’s not really complete data.
Again, the greatest challenge here I think is funding for our system. Our state is now facing, it was just released, like a $237 million deficit. So there are major budget cuts coming in the next budget session. So any hope I think for our department of health to be able to move to an electronic death certificate with state funding is probably not going to be a high priority. There’s no tie-in to the ME system either.
With marriage data, I think we’ve talked about this, not electronic, and again, no linkage to divorce. So this again has some fraud waste and abuse implications where benefits are tied to marriage and there’s no cross-check on that.
So potential solutions, again, we’ve talked a lot about the interoperability between EHRs and electronic birth registration systems and with prenatal. One idea that I threw up here was, and I’m sorry for my typo there or lack of spacing, but we’re really working in Rhode Island through our health information exchange and through other things about patient-generated data, and I don’t know whether there’s a way for the mother’s worksheet to become more automated, either through a patient portal, either through an HIE, and then integrate it and becoming more timely because that’s double data entry that has to occur by the medical record staff at the hospital.
So I just wanted to call that out, that there may be some thought and opportunities there. Also, again, interfacing with health information exchanges, where consent models permit. In our state, we are an opt-in so that wouldn’t work, but prepopulating a health information exchange with birth data for those either statewide or local or jurisdictional systems where that would work would really jumpstart and help health information exchanges, and the same for indicating that there are some deaths there.
So moving along to, again, I think I mentioned really I think the value in uses of vital records data quantifying the return of investment on fraud waste and abuse and then taking that to legislators and others is really one way to communicate the value around that. Then supporting where states have data warehouses or statewide integrated systems, really making sure vitals get integrated there and seeing if there is some funding there.
We have in our state, through our Rhode Island Innovation Policy Lab, done some stuff with birth registration and Medicaid data and identified that there were significant differences in health outcomes for mothers who were enrolled in Medicaid prior to birth as opposed to not.
Importantly, I just want to talk a little bit about funding solutions. I know there were a lot of questions about Medicaid before. We have not pursued this, but this is an area that I think I’ll be working with my colleagues to try to pursue, and that is around enhanced match 90/10 funding through Medicaid, 75/25 for ongoing. There are different tracks to get enhanced match. It is for the IT system, not for the data and analysis. For development, it’s 90/10, maintenance and operations is 75/25, but it could be tied into either the eligibility and enrollment systems with Medicaid and insurance exchanges or the MMIS claims.
All of these are done through what’s called implementation advanced planning documents, IAPDs, or potentially through fraud waste and abuse pre- and post-payment systems for Medicaid, and/or meaningful use; could vital and death registration systems be considered a specialized registry and count towards public health? If so, then that might enable a Medicaid or a state to be able to get 90/10 funding to connect the public health components. So those are just some ideas to continue to pursue.
Lastly, two other things. One, a lot of states have all payer claims databases and they charge for their data files. I’m not sure we do this with vital records. It didn’t really come to me. I mean, I was thinking about this yesterday. So there are models. In our state, we have an all payer claims database and we will be charging not for other state agencies necessarily but for researchers and external entities. So that’s something to look at as a model.
Also, identifying other sources of funding that include health information technology components. For example, the SIM, state innovation model, cooperative agreements for states. You’re required to have an HIT plan. There are only two years left. This is an example. We didn’t include vital in this. We did some other things in our HIT plan, but again, just to get people thinking that there are other opportunities here and ways to do it as it relates in this case to population health or assessing the outcome of value-based payments.
Then my last thought, which isn’t up here, was around whether model laws, and this was raised yesterday, could include something around what I would call in my state a restricted receipt account. There were comments about how states can charge but the money doesn’t come back to vital records. Well, if in the model law, there were some language or something that sort of said the money goes back to support the program, because for us oftentimes that’s done legislatively and then the money can’t go to the general fund. It has to be used to support the program.
I’ll stop there.
DR. ROSS: Thank you. Thank you very much. Our next speaker is Dr. Michael Hogarth from UCSD. Michael, please?
DR. HOGARTH: Thank you and thank you for this opportunity to discuss challenges in vital health statistics. I have an academic appointment in internal medicine as well as in pathology and I have an interesting perspective which I’ll share with you in my discussion. I’ll focus my comments primarily on the challenge of access to vital statistics data by healthcare providers and by medical researchers.
Ultimately, access to this data has many benefits for the patients we serve. As has been reported, there is no existing process whereby clinical providers can know which of their patients is deceased. It’s not that this information doesn’t exist in electronic form. It actually does. The National Death Index has this information, but laws and regulations prevent healthcare providers from using it to systematically and regularly match EHR records and death records to determine who is deceased.
It is available for clinical research projects, but not for quality improvement. This is shortsighted as we as a learning healthcare system need to learn from every patient. Considering that healthcare delivery affects all patients, it would behoove us to allow this added use.
It was also ironic, given the healthcare providers partnered with public health to source key medical information in this dataset, but when it comes to access to vital record data to help improve healthcare delivery or to uncover the next generation of cures, we’re no longer a partner. We’re given non-preferred status to the very information we helped to source for public health.
I wear several hats that give me an interesting vantage point on this problem. I’ve practiced internal medicine for 25 years and have certified causes on many death certificates. I have knowledge of vital records process that most of my colleagues do not because my informatics groups developed, and today operates, the California EDRS system, which electronically registers roughly one in ten deaths in this country.
We also work with Maryland to develop an EDRS system for them based on the California system, but at one tenth the cost. We make sure both states own their respective systems. We’re currently building a new birth registration system for California and we’ve also built what we call the Vital Records Business Intelligence System, which is a reporting and dashboard system for California.
I’ve also spent over 20 years in medical or health informatics and in the past year, my team funded and collaboratively with NCHS and California Department of Public Health developed an integration with the California EDRS and a commercial, broadly-used EHR system which claims to manage over 60 percent of all the records in this country. This integration is in use today at UC Davis.
I also participate in clinical research. I’m the clinical research information officer for UC San Diego today. Recently, I’ve been involved in the California Precision Medicine Consortium and last month began accruing for the groundbreaking NIH one million participant All of Us program, which was mentioned yesterday. Mortality information is key to the success of this initiative.
As a physician and researcher, I’m disappointed. The EHR use can do amazing things, like speech-to-text conversion, clinical decision support, electronic prescribing, and medical record exchange with hundreds of hospitals across the United States. It gives me a myriad of metrics, but the one important statistic it cannot provide reliably is mortality.
The problem becomes front and center when we inadvertently attempt to communicate with individuals who are deceased. Imagine how we feel when our office sends a medical appointment reminder for a patient for whom we’ve cared for many years, only to have a distressed spouse call wondering why we don’t know the passing of their loved one. Imagine how that spouse feels when we send a medical bill to their deceased loved one and we never send a card or a letter expressed condolences either.
As a physician, when my patients pass away, our relationship does not end, it only changes. It changes from me helping them to them helping me. Knowing about their death is critical in helping me be a better doctor. Medical practice is not static. It is a process of continuous improvement and mortality rate is fundamental in enabling providers like myself to continuously improve how we deliver care.
Having accurate vital status of my patients is key to having a correct denominator for many quality measures. Although 30-day readmission is an important quality measure, 30-day post-procedure mortality is even more important. I’m unable to reliably calculate the statistic, however. The problem impacts every quality metric that presumes only alive individuals are in the denominator.
It also affects biomedical research. It has an impact in several ways. Last year, a colleague contacted a randomly selected cohort of a thousand patients between 18 and 90 years of age about participating in a study about weight management. One in 15 of the invitations ended up going to a deceased individual. My colleague heard from their family members. It was not only wasteful of resources, it also created anger and frustration among their loved ones. The currently allowed research use of the NDI would not avoid this problem.
How much vital records status error do we actually have in our EHR systems today? I have access to a federated data system with de-identified EHR records of 15 million patients from the five UC academic medical centers.
I performed a query earlier this year for all non-deceased individuals who are 80 or older. The answer, 1.85 million people. The 2010 census reported only 600,000 such individuals alive in California. As has been already stated, in 2011, the SSA removed protected state death records from the DMF, making it far less useful, as it excludes 40 percent of deaths in the United States after that year.
The recently introduced NAPHSIS Fact of Death service is a step in the right direction. However, it is not complete and the model of submitting data to the service will be problematic for health systems. We would be sending millions of records of non-deceased individuals to an external entity. Consider with 90 percent adoption to EHRs today, it is likely that there’s at least one EHR record for every person in the United States. That’s 326 million records. That is not practical nor good security practice to submit that volume of records to an entity for multi-matching of such a small number.
The optimal solution is to securely send the much smaller file of decadents out to healthcare systems. The healthcare systems or EHRs have record-making functionality in a piece of infrastructure they all have. It is called the enterprise master patient index system.
I do not pretend to have the answer to how best to broadly provision systematic EDRS EHR matching. However I wonder if the NAPHSIS fact of death service could be connected to the National Health Information Exchange Network known as e-Health Exchange, supported by the Sequoia Project, and must be configured to send individual death notifications to all those provider EHR systems connected to this network. One might also be so bold as to envision that one day public health would receive and in turn the structured and computable consolidated CDFL for the decedent.
That functionality is programmed into every certified EHR system used in the United States today, that added information would be of significant epidemiological value to public health and outcomes researchers who use the vital statistics data. I recognize the national EDRS and EHR matching service and the integration of a fact of death service into HIE networks are nontrivial proposals, but my experience in implementing an electronic vital record system in California has taught me nothing is impossible.
We just need to set the goal, be patient, and be persistent. It can happen. In the words of the British contemporary philosopher Bernard Williams, there was never a night or a problem that could defeat sunrise or hope.
DR. ROSS: Thank you very much. If you can do it in California, you can do it anywhere. Thank you, Dr. Hogarth. That’s great.
We have one last speaker on this panel, and that is our colleague Dr. Janet Hamilton, who is calling in from Florida. Because of the storm, she was unable to be here.
Janet, are you there?
DR. HAMILTON: I am.
DR. ROSS: Great. Janet is calling as state epidemiologist in Florida but also as the representative from the Council of State and Territorial Epidemiologists, CSTE. So, go ahead, Janet.
DR. HAMILTON: Okay, great. Thank you so much for the opportunity to speak with you all today and for making special arrangements to accommodate CSTE’s testimony in light of Hurricane Irma.
The Council of State and Territorial Epidemiologists or CSTE, greatly values our partnership with vital statistics and welcomes the opportunity to provide testimony today. I am Janet Hamilton, an epidemiologist with the Florida Department of Health and currently serving as CSTE president. CSTE is a national organization that represents over 1,800 public health epidemiologists working in state, local, and territorial health departments. We work on the front lines of public health, investigating and controlling communicable and noncommunicable diseases nationwide.
So how do we use these data and why are they important? The cornerstone of public health practice, epidemiology is the science of determining the patterns, causes, and effects of health and disease conditions in defined populations. Epidemiologists play a pivotal role in shaping policy decisions and evidence-based clinical and public health practice by turning population-based health data, including vital stats data, into structured information that helps explain the context of a problem or situation and thereby inform wise decision-making.
Epidemiologists track diseases not only to quickly detect and control outbreaks and emerging threats, but also to monitor trends over time, which helps us determine the impact of policy and practice interventions. Epidemiologists count. We count people and risk factors associated with outcomes.
Birth and death certificates are unique in our work. Everybody gets one. At its core, vitals data provides the denominator or ultimate foundation for our work as epidemiologists. Death certificates are the granddaddy of surveillance systems, going back to the late 18th and 19th centuries in some locations. My great-grandfather pictured here was a farmer and a toolmaker during those heady times. He lived to be 99, just three months shy of 100. He attributed this longevity to daily chaws of Red Man.
Of course, we know, due in large part to vital records data and the work of epidemiologists to understand causal relationships, that this in fact cannot be true.
As evidenced by this hearing, there are multiple separate ways to use and look at data about births and deaths. From an administrative perspective, the fact of birth is useful for proof of identity. The fact of death is useful information for people concerned about pensions, Social Security, life insurance, and permission to bury.
As epidemiologists, the circumstances at birth and the cause of death are important, and we use this data every day to look across the records to draw conclusions to improve health status, health equity, quality of health, and health outcomes. It is the work of public health epidemiologists and their use of births and death certificates that helped us understand the dangers of tobacco use during pregnancy and as a carcinogenic, black lung among coal miners, silicosis, asbestosis, and to implement seat belt laws.
Vital records data provides foundational data and is useful all on its own, but public health epidemiologists often link the vitals data with other sources and registries to draw additional population-based inferences for those that cannot be made with vital stats data alone. A lot of examples on this slide are familiar to you. This list is far from complete, but it should provide some insight as to the breadth of our work and interaction with death data.
I’ll highlight just one piece on this slide, the issues around disease control, which may not have been brought out by others. Death data is used, for example, to identify next of kin and contacts to TB cases and to identify then potential missed sites of TB exposure. For example, funeral homes and worksites.
As in the previous list, here are just a few examples of our collective work as it pertains to the utilization of birth records. We are confronted now with the Zika epidemic and how to use our birth defects surveillance to combat that. Linkages of birth records to mothers’ prenatal care and behaviors to combat neonatal abstinent syndrome. Again, to highlight issues around disease control, for example, matching birth records to identify HIV-exposed infants for investigation and follow-up, and matches to birth records for hospital exposures among newborns to tuberculosis, as well as using birth records to identify next of kin and contacts to TB cases.
So how does all of this work? How does the transmission of data work? Ultimately, what I would say is that data is protected securely. Across public health, data security and protection is paramount. Across all jurisdictions, staff members receive specialized training in order to ensure data is protected.
For as many different uses of the data, there are equally as many different types of ways of sharing data. Data use agreements, memorandums of understandings, I think you’ve heard about this from others. And there is variability within a state as across states. To state it simply, it’s complicated. Different types of datasets are also provided, not necessarily transmission of data.
Again, we will be looking now at challenges and opportunities. The National Notifiable Disease Surveillance System is another federated multifaceted public health disease surveillance system that all states and territories participate in that gives public health officials powerful capabilities to monitor the occurrence and spread of diseases. CSTE maintains the responsibility for defining and recommending which diseases and conditions are reportable within states and locations and which of these diseases and conditions will be voluntarily reported to CDC. In collaboration with CDC, CSTE works to determine changes to the list of national notifiable conditions through the passing of CSTE position statements.
Each position statement has a criteria for case ascertainment and reporting. See example tables on this slide. Every year, in every state, the review of ICD-10 codes is used to identify unreported cases of reportable diseases. In many states, this review has also been associated with diseases of importance, like the identification of hantavirus. In some diseases, like influenza-associated pediatric mortality, it is almost exclusively identified via review of death certificate data and, sadly, often goes unreported.
When identification of cases is done through death certificate data because of the inherent data delays in many states, the opportunity then to collect a specimen and have it tested in specialized public health laboratories to determine the strain or type of influenza has often passed. This cripples one of the goals of this surveillance effort to identify potentially drifted strains which could lead to more severe presentations or even the identification of a novel strain.
In Florida, we have established a particularly innovative and useful approach to receive death certificate data every 24 hours into our syndromic surveillance system. We use these tools in the system to perform pretext queries on the literal causes of death across all records to identify pediatric influenza deaths rapidly and other potentially non-ICD-10 coded data. Review of the literals in this way has also proved useful for estimating fentanyl-related deaths.
Death certificate review is used to match known cases of disease to track outcomes associated with long-term sequelae as well, like taking the information on West Nile cases and looking to identify the long-term sequelae of a known neuro-invasive West Nile case, identify outcomes for causes of death in known TB and HIV/AIDS cases. We are currently exploring this same approach for congenital Zika cases.
DR. ROSS: Janet, you are getting actually past your time, but if you could think towards wrapping up relatively soon, all right?
DR. HAMILTON: Okay. Let me just move on then to slide 14, post-storm or event surveillance.
So I cannot talk to you today without making mention of the use of death certificates post-storm or event. The information provided on death certificates including the cause of death is collected and used to guide the response effort and later produce verified and accurate state and national estimates of disaster related mortality. These data support the development, policymaking, and funding that lead to the adoption of evidence-based prevention strategies for future disasters, and has been used to set evacuation policies and timing, curfews, requirements for carbon monoxide detectors, in rental housing.
Let’s go ahead then and move on. We have talked about combating opioid abuse, and move to the last slide, which is recommendations, slide 17.
So our recommendations are threefold. The first, to expand automation beyond electronic filing to decision support and bidirectional communication to support querying back to gather additional information, to improve linkages, as we have heard, with electronic medical records, medical examiners’ data, and toxicologic data. These data sources in combination with public health reporting could fill important gaps in information, especially as healthcare is expected to shift to more decentralized care settings with variable data reporting infrastructures.
Imagine how those techniques and data sources could fill in important information and provide the opportunity to look at a single person longitudinally and episodically across multiple settings of care.
Our third recommendation is to invest in data quality. We want to use these data to improve health across the continuum and longevity of life. We want our children’s lives to be healthier than our own. The ability to make data-driven decisions and inferences, to set policy, is only as good as the data in the system. Invest in training and education about the intrinsic value of the data, how the data are used, and how to complete the form.
Pieces listed above in automation such as decision support and bidirectional communication will also assist in meeting this goal.
And the third item is let the data flow. We are all one public health committed to the protection of the data that we possess. Develop processes and procedures to share the data in timely, automated fashions, and standard based fashions. Develop model language to share identifiable as well as de-identifiable data.
None of this, of course, can happen without funding to support infrastructure. State and local public health informatics infrastructure needs modernization and greater human capacity for our nation to fully benefit from better and more timely vital statistics data and public/private partnerships, like the Digital Bridge, will be critical to working out implementation details in mutually beneficial ways.
CSTE appreciates the opportunity to participate in this conversation, our partnership, and looks forward to working with partners across the public health spectrum at the local, state, and federal levels.
Thank you so much.
DR. ROSS: Thank you, Janet, and thanks for participating with us, knowing you have just come through this bad storm. I hope you and the family and everyone has survived it well. I know you are going to have your hands busy for the next few weeks and months.
With this, I would like to now open for committee questions and/or comments. Do we have any questions? All right. We will start with Bruce.
DR. COHEN: I guess this is for you, Lou, and for NAPHSIS in general, but an observation and comment. The issue of permission based access to determine fact of death on a real-time basis has come up in numerous contexts, whether commercial use, Bill talked about NIH, Michael just talked extensively about the need to be able to do this.
Can you tell me what’s going on? I am asking you, because I guess the most efficient and effective way at this point seems to be getting NAPHSIS to get the jurisdictions to work together to contribute their mortality data on a real-time basis to a central hub that would allow use for individuals who have permission to use it. Could you tell me what’s going on in that area?
DR. SAADI: Okay, and Shawna can comment to fill in where I’m not clear on a few things. But we have asked all jurisdictions to participate, share the process of what we are hoping to achieve, to provide access again with the death master file being delineated from the state information. We have established based on the same traditions of our birth EVVE, and you heard Passport Services talk about how they use birth EVVE, the same provision, for death registration and death information, and I can’t remember; we have, Shawna, 20 jurisdictions participating? Is it up to 22 now? Fact of death.
MS. WEBSTER: I think it is up to 22.
DR. SAADI: Twenty-two. So it comes down to like we’ve mentioned through this conversation, what jurisdictions think that they can do within their statutes, and like Shawna mentioned earlier, we can share information among government agencies in our statutes, most everybody is allowed to do that. Where we come through difficulty is approaching sharing information for nongovernment entities.
Now, the way e-fact of death and the way we have been marketing it to our jurisdictions is send your data to that — your fact of death information, just fact of death, not cause of death, which again there are some states with difficulty, but send your fact of death information to that. Then you can decide who to share the information with. Now that provides difficulty, like you heard from pension benefits institute, they are a third-party administrator, and so jurisdictions are still a little bit concerned about how that information is going to be used for those kinds of entities, because there’s not a direct relationship between them and the policyholder.
So we have a long way to go in getting our jurisdictions to look at their laws. For Kansas, our attorneys looked at the tangible interest process, and they felt that pension benefits, annuities, those kinds of groups would have permission to access our information in an identifiable manner.
However, when we get into research and further follow-back activities sometimes that are involved with research, we have difficulty. So we are encouraging and have been encouraging for the last three years for jurisdiction to — each jurisdiction to update their statutes, and many of us can’t get — like I’ve mentioned before the difficulties in getting our statute provisions outside of our agencies and into the conversation.
I hope that gives you the detail. We are growing. It’s coming. It’s just a matter of getting complete participation and the processes in place.
MS. KLOSS: Thank you. My question is for Dr. Hogarth. I was interested in hearing a little bit more about the metrics, the performance metrics dashboards that you have developed and what kind of measures are being tracked.
DR. HOGARTH: So for the vital records process, these are operational metrics for jurisdictions. So turnaround time. What’s interesting is when you implement an electronic system in a previously paper process, one side benefit is what we call electronic breadcrumbs. You get to see the process actually happen, which before you didn’t know much about until the certificate showed up at the central office, for example.
So we track all — the status of the certificate from its creation all the way through to the end. We have implemented electronic signature systems, or at a station using fax and voice, which was a key adoption advantage in California, which now does 100 percent fully digital certificates. So we track the whole process. We don’t drop anything to paper, other than a certified copy, and all that information is now available to the jurisdictions and the state office through this vital records business intelligence system. It might be an overreach to call it business intelligence, but it was the buzzword of the day.
MS. KLOSS: But you are able then to monitor how all the jurisdictions are tracking and —
DR. HOGARTH: Yes. That has been a key actually in when you present jurisdictions with data, anonymously, and how they stack up. It’s amazing how peer pressure works in improving the process.
MS. KLOSS: I was thinking about that in light of Delton’s earlier comment about some states still using paper and it seems to me, if you just kind of threw up that list and showed everybody, that there would be some self-direction.
DR. HOGARTH: I do want to mention that going from paper to electronic is often a difficult process, as we all know, and we have seen that with electronic health records. Those that provision the data spend more time doing that with the electronic system than they would in a paper system, but there are so many advantages to the electronic system in terms of structured data and the reusability of it down the road that from an enterprise perspective, it’s a big advantage if you can actually line up the value with those that are paying for that.
MS. KLOSS: Building in this performance improvement feature is something that really needs to be part of any modernization.
DR. ROSS: Vickie.
DR. MAYS: There was a comment, I think, Amy, you made it. Every clinical trial should have a mandated autopsy and the funding that goes with it. I’m really trying to get a sense of why in terms of — you’re asking for, like, a full autopsy and then the funding, the funding would come from whom?
DR. ZIMMERMAN: That comment wasn’t made by me.
DR. SENS: That was me. I made that comment. The autopsy is the ultimate — and death — is the ultimate outcome measure, and there is a lot of very disease-specific information that can be obtained. Many of the trials involve some modification of gene therapy, of pharmacology, of new drugs, receptor drugs. We can get information from that from the autopsy.
DR. MAYS: But why clinical trials only? They are not like other kinds —
DR. SENS: I am a strong advocate of the autopsy for everything. I am a strong advocate of performing an autopsy. That said, you can get specific information — I think there’s two levels of information with an autopsy. You certainly get information about why the person died, and that is very helpful, that drives quality improvement. It provides some resolution for the family. Surprisingly, I have never had a family regret doing an autopsy. I have had many families regret not doing one after a few weeks of thinking.
The other and more important and totally unutilized version or help of the autopsy, and this is where a medical death investigation hardware/software could be so valuable is looking at how many people died of myocardial infarctions and what was the background for them, and you look at not an individual but as a group of individuals and it’s aimed for quality improvement.
The other untapped potential of an autopsy is in the quality improvement that you refer to. Our office will routinely, when we get a — say, if a person showed up at an ER and was sent home and then they died, we will contact that health system, we contact the peer review of that healthcare system and say, listen, these are our findings. Do you want to look at it? And they usually say yes, and then I don’t know what they do. I have been on peer reviews, but it’s a nice internal process.
And they’re not out to get the people that could be — but it’s to improve the whole system.
DR. MAYS: So this is beyond then just clinical trials. I mean, it’s almost like you’re an advocate for this in general.
DR. SENS: I’m an advocate for this in general, but for clinical trials, there’s a tremendous investment already being made, and there’s a tremendous decision process that are being made. I mean, this is how cancer treatment may be given or not given, and if it wasn’t working in a patient, we may need to know more about that, the pharmacogenomics of that patient or the other unknown conditions of that patient that may have made a difference. So I think it would be very helpful for — now obviously, that ultimately is the patients and their families, it’s a consented process. But that should be a part of the recruitment to a trial and that should be offered to the family and to the healthcare system that is conducting the trial.
Any additional comments?
DR. DAVIS: No, well said.
DR. HOGARTH: One quick comment, if I could make it. I have seen research in the past that is — I’m a big advocate for autopsies, and that has shown that even though we have great imaging and we believe ourselves to be great diagnosticians, that the autopsy has repeatedly proven that around 10 percent of the time, the cause of death is not what we thought it was. I think that’s really important, particularly for clinical trial use.
DR. SENS: Especially, that has not changed since the 1800s. That’s been that way, you go back to 1920, big huge series, it’s 10 percent we don’t know. There’s one study out there where it was a German hospital continued to have a high autopsy rate, and as they acquired imaging and better diagnostics, their rates may have started to go down. That was the only hospital in the world that did that, and that was probably because they were at a baseline with their autopsy, and then the others enhanced it. But 10 percent of time, the cause of death is totally different than what we expected.
DR. DAVIS: The most recent such study was published this month in the Archives of Pathology and Laboratory Medicine and said the same thing.
DR. ROSS: Let me check in on the NCVHS members who are on the phone. Any questions from the phone?
MS. STRICKLAND: No, thank you.
DR. ROSS: Seeing no additional questions, so comments from the audience. So now we are opening up for audience input.
NCHS staff: I have a question based on — oh, I’m Louise Savia(phonetic). I work as a coder and a contractor with NCHS. So I help code infant mortality and HL7 stuff. So this is a question based on Amy Zimmerman and Michael Hogarth’s presentations. If we are trying to improve the timeliness and accuracy of the data, has the committee considered leveraging crowd-sourced information from researchers to improve the accuracy of the data as a way to save time, money, and specifically target at school. If you would like, I can explain a real-world example of this.
Okay, so Waze is what you used perhaps to get here to D.C. on time without traffic. It’s an app that is user generated. A user can enter when they see a police officer, where there is traffic, and what’s going on, and who knows the roads better than you who are stuck on 495?
So if we were to use a crowd-sourced model broadly and specifically focus on the researchers who are receiving the data and even the families who are dealing with those who are deceased, they understand where accuracy needs to be changed. So across the United States, what are researchers focusing on, and is maybe do we find in a broad study of this that there’s only one field of data that needs to be improved? This could lighten the load for federal employees by giving them scope, and it could save money because you could even incentivize the cost for getting these records if they gave accurate feedback on the data, they could get some money off the data that they are receiving.
DR. ROSS: This is out of the box thinking.
DR. ZIMMERMAN: I think two things. I think there’s certainly a role for what I would call patient-generated data, but it is in the context of patient-generated data, meaning it’s patient-reported. So to the extent that there are legal or state or certification or civil requirements, I think if it’s something that’s already and only can be gotten from the patient, you have to blend that with the legal framework and context, but I do believe that there may be efficiencies, and I’m not sure if this is why I got posed the question, because I do think that there is opportunity perhaps to streamline things where patients can enter information electronically versus on paper, and then having to get someone either in the hospital or someplace else to enter the data.
With patient portals and HIEs having patient portals off of them, I think there is more and more applicability where this has not been considered and could be considered in the realm of these types of documents. So I don’t know that I really answered your question, sort of roundabout.
DR. HOGARTH: Just a quick comment. I would say a fair amount of the data is already sourced from family already, comes from the funeral director, but it’s coming from family. The other is research has shown in healthcare that data that’s patient-reported is actually as accurate or sometimes more accurate than what the healthcare professional records after talking to the patient.
DR. DAVIS: But I would say that your model assumes that all the information that’s entered will be accurate, and I would look to Wikipedia and the examples of how sometimes people with ulterior motives go in and change an entry for their own purpose. So that would be a danger.
DR. HUGHES: Just a quick comment on an in-the-box version of — this is Eric Hughes. For cancer deaths, we are already linking all cancer deaths to — excuse me, deaths to cancer registry data across the country, and perfect opportunity to do this in sort of an in-the-box way to compare the cancer diagnosis that the registry has with the cause of death that the death record has. But feed that back to the death registry, death registration system.
DR. PARRISH: Gib Parrish with a question for Michael Hogarth. I am interested in your electronic breadcrumbs, and in particular wonder if you have in California looked at the certifier, the medical certifiers, of cause of death and, in particular, you mentioned that you actually have certified a lot of cause of death but looked at the distribution of, by physician, the number of certifications they do and whether you have in fact a very large number of certifications done by very few people and then a long tail with very few certifications by many people?
DR. HOGARTH: You just defined the histogram for death certification by providers. We have 120,000 licensed physicians who can do a death certificate in California. The vast majority do one to two a year. They are not trained usually, maybe one lecture in medical school, which is an odd time, because they don’t do these until they’re in residence, and what I think is important is we integrated cause of death entry into the EHR, and the big opportunity there is the same thing that happened with physician order entry.
It wasn’t that the clerk didn’t have to enter the order anymore. It was the opportunity to interact with the person actually doing the order or entering the data at the point it’s being entered. That is the best time to get your highest quality, and I think CDC has pioneered the VIEWS system, which I hope will continue and actually be a web service available to systems like the ones that I’m designing, because now I have an opportunity to actually perhaps score or provide some feedback to the certifier about the cause of death.
In my hospital, we have the ability to actually send the causes of death to the local registrar for validation before it’s submitted for real and everything is locked, and 60 percent of the time, it’s rejected by local registrar. You get things like they died of cancer, pneumonia, not enough specificity. So there’s clearly lack of training. I think training is important, but just in time training, feedback right as they do the certificate, the causes, I think would be key to improving the quality and timeliness.
DR. PARRISH: Thank you.
DR. ROSS: Michael, is that just in time training in your view something that could be embedded into the EHR products? Because I think about my experience in hospitals and presenting — that tool for better or worse is the way that now with physician order entry you’re interacting. Is that the time to have it happen?
DR. HOGARTH: So a little nuance. We actually provide access to the actual California electronic death registration system in a different form factor within the EHR, using an interaction between the two systems. So they are actually interacting with the EDRS system in our integration, which gives us a lot of flexibility because we now in the EDRS system are interacting with the data provisioner. It’s not the EHR vendor.
The EHR vendor facilitated this by just simply implementing a standard by IHE, Integrated Healthcare Enterprise, and HL7, called VRDR, and this particular vendor, which is broadly implemented, has done that, Epic. We successfully used that very quickly to do this maneuver, and so they are enacting with our system as most of the EHR vendor — I just don’t think you could get enough velocity to get the vendors to constantly adapt to the edits and things that you want to do and to be really dynamic and flexible with what you want to do with the interactions. There’s a lot of interactions that can happen.
Last thing, we created an NVDRS data entry form in our EDRS for coroners, same with California Highway Patrol traffic fatalities. So I just want to make sure we don’t think about vital record systems as just vital records. They can collect additional information around that event that’s really helpful for epidemiologic purposes.
DR. ATKINSON: Just a couple of points just to follow up on it, I think Utah is doing something very similar where they have integrated that cause of death information into the electronic medical record, and it’s a seamless transition to the user.
Second thing, you mentioned VIEWS. VIEWS is available. It is no cost. It can be used. If anybody tells you it can’t be, please call me.
The third thing, this whole issue around physicians and the information and completing the cause of death, we have just a real project with Georgia Tech University where we are looking at and trying to put together a decision support tool, and what that tool will do is pull the relevant information from an electronic medical record and try to provide the physician with all the relevant information they need in completing the cause of death.
That is still under development. We hope to be trying to pilot that in the next year or so, but looking at sort of these decision support tools to help in this particular process, because we do have that problem of very few positions were filling out ones and twos and threes, or a resident at a hospital who know nothing about that particular patient, that individual, happened to kind of show up to that particular hospital.
So again, looking at technology and whatever ways that we can to see if we can improve the way in which we complete the cause of death.
DR. STEAD: I am sorry. I am getting real cognitive dissonance here. I get the idea, if I have a person and they’re walking down the street and they get hit by a car and they’re pronounced dead at the scene that we probably know the cause of death, although we may not. They may have had a heart attack. That’s why they hit the car, et cetera.
So I’m really — I think we are into a slippery slope here. Increasingly our population has multiple comorbid diseases, and so I think the idea of the fact that we might be able to link an intelligent abstract from an EHR to a fact of death and possibly a circumstance of death, I think we might could do that, but I think that the amount of information that we may have, it’s a real judgment call and in most cases it’s probably multifactorial, in many cases.
So I would be cautious of what I’m — it comes back to the loose coupling idea. I’m getting uncomfortable with what I’m hearing, just as your generalist.
DR. ROSS: Is this an opportunity though, a research opportunity, to explore some of this?
DR. STEAD: Absolutely. There are any number of research opportunities if in fact we could link the real medical record to the degree we have it to the fact and circumstance of death. I think you can then use all sorts of intelligent algorithms to probably get at what we want, but I think we may need to be careful of somehow of distinguishing when there is a clear cause of death, and when we in fact have a fact of death and a set of relevant information, and I think unless we — I’m just getting cognitive dissonance as I listen to this conversation.
DR. ROSS: We are at the end of our time. I am compelled to make one sort of — oh, Tara. Before we break.
MS. DAS: I just wanted to make a couple of just overarching comments based on the past two days. One is that you are hearing a lot about data quality. You hear it not only from the data users but from those of us in the local vital records jurisdictions about challenges with data quality.
I don’t think that that’s necessarily a bad thing, though. I think it’s a very good thing that we are all very aware of the data quality issues, and that’s because our birth and death data are rooted in local realities and everyday lives as we talked more so about yesterday, and I think because those of us at the state and local levels are so close to the data collection and have excellent relationships with our data providers that you hear so much about data quality as an issue, but I also want us to keep that in perspective and understand that what big data collection system does not have data quality issues?
So I don’t think that that’s necessarily a bad thing. I think it’s a very good thing that we are all so keenly aware of data quality being an issue and that we are so honest and up front about it. So I did want to keep that into perspective.
The other thing that I wanted to mention is that given that we talked about a vast spectrum of activities that we undertake in our world of vital records and vital statistics, I don’t think that necessarily the solution to all that is to reduce the amount of activities that occur in our world or perhaps to decouple all public health purposes of birth and death reporting. I for one am very — I think it’s very — and makes sense to me that there are both legal and public health purposes to the work we do. As I mentioned, many of our local vital records offices were rooted in health departments. We see a core public health mission to what we do.
But at the same time, I do think that the public health purposes may have become too expansive, particularly on the birth reporting side. I think one of the things that we talked about yesterday is a very good point, which is reconsidering how many data items we collect on the birth certificate, particularly in the confidential medical report.
There are many, many items there, and those of us again at the state level, one reason that we have so many data items on the birth certificate is because many of our data partners have requested for those items to be put on the birth certificate, like results from hearing and screening registries, blood spot screenings, because they see it as a very simple thing for them to just tack on what their needs are onto our birth certificate because we are reporting that for each and every newborn.
So I think that that is a request from the states to our data partners at the state and federal level is to think about that, think about the demands that you ask of us and increasing them without providing any kind of support, or really also consideration as to whether they are appropriate to ask of us.
So those were some points that I wanted to make, and in addition, just to reinforce what I had said yesterday, and I think that came out wonderfully with experiences that we heard from my colleagues in Washington, D.C. and California today, that while we may have challenges as in the United States at the national federal level in terms of timeliness, electronic modernization, and variation in laws across all the states, we do have very strong states and state leaders who have surpassed that baseline and are above that floor.
And I think that we also need to look internally within our vital records community and leverage our state leaders and see how we can help the laggards in our community in reaching that floor, because I do want to recognize that many of us have surpassed that floor, that baseline. So the things that we are hearing from our federal partners as to the challenges with the data, that that is across all states, but that doesn’t mean that it applies to each and every state.
So as one of the states, I wanted to make sure to say that, and that also as Dave was saying yesterday, I think our developing a community of practice would be very welcome so that we can help one another not only to ensure that we are all at that same baseline with supporting all of you and your needs but also that for those of us who thrive on being on the forefront of the community and doing cutting edge things and innovation, that we also have support in doing that, and if there are ways that we can learn from our colleagues in the private sector and at the federal level — we heard yesterday about some things that were being done with cause of death information that, Helga, I think you were pointing out, maybe there is a way to feed that back to the states so that we can learn from those efforts as well.
I think that was an — I really appreciated that point, because I think that that is how we would develop a thriving community of practice.
DR. ROSS: Tara, before you step away, I want to follow up with a question on — you mentioned regarding the birth certificate, the appropriateness, sort of adjudicating or having a way to decide what really is appropriate request of other programs, adding onto the birth certificate.
So what would be the vehicle for that? I think of like in the case of like a reportable disease, the CSTE as a community has a process for deciding both case definition and then the decision about is it reportable or not. What about this? This isn’t an issue of legal reportability but really it’s almost of reasonableness and fairness. What constitutes a reasonable ask, right? How would you do that?
MS. DAS: I think that is an excellent question, and I don’t have the answer for you. But I think that’s really good, and I think that gets back to our really needing to develop a thriving community of practice where we can all, as a federated system, come together through NAPHSIS and NCHS and kind of come to standards for how we would respond to those kinds of questions, because even things — what would we consider to be reasonable appropriate ask of adding to the birth certificate?
That is all something that we all just sort of come up — we all have our own standards for what we think is reasonable or fair across the states, and at the same time, you know, I received a question from one of my statistics colleagues in Texas about, well, how do you, in terms of reviewing research applications, whether it’s a reasonable or fair use to provide particular birth or death data to a researcher, what are our standards for deciding is that a reasonable use or is that going to violate confidentiality or privacy, and honestly there aren’t really any standards. It becomes — it’s what do we decide at the state level is, again, reasonable or fair, but if we could have more consistency in our standards across, I think that will be good.
And also, just again, just getting back to learning from one another. I mean, just having this hearing was so useful and just hearing from my colleagues in other states as to what is going on, we don’t really have a very robust venue for doing that.
DR. ROTHWELL: My MCH friends would not like to hear this, but let me just throw this out. When we redid the birth certificate years and years ago, there weren’t electronic health records, and the reason why if you put it on a piece of paper it would be two pages, maybe two and a half pages long, is because we wanted to collect everything we could think of that wasn’t available electronically.
I’m wondering if we ought to maybe rethink that, and maybe what we should do as NCHS is reduce drastically what is on the birth certificate from a medical perspective, which would then put pressure where it should be on the electronic health records to collect this information, and then we would, from a sample perspective, go back and collect that which we need from a national perspective and states could as well.
Instead of trying to force the issue of having this really — yes, it’s an automated system, but it was based on a paper view of life. We have to move away from that. I won’t get into death end of things, but I think our California partner is right on, right on the mark.
DR. ROSS: I think you have teed up very well the afternoon’s conversation, because looking to the future, really it’s an issue of do we pull back and do kind of a pretty substantial rethink or continue to build on a world that really evolved out of paper processes that do not in any way reflect where we stand today.
We have another committee —
MS. LOVE: I would just say this afternoon, I could talk a little bit about working with states over decades to get rid of state unique fields. There’s really very little that each state is unique on, and we have knocked out state unique fields on hospital discharge data and we spent a year with claims, and the other states are the best ones to push back on why is Massachusetts collecting these three variables that are crappy and the other states bring them in line? Because it really is not fair to the data suppliers and it’s not useful data.
So I talk about some models and I could say more, but I’ll let Lou have the final word here.
DR. SAADI: If you don’t mind, I just wanted to mention, and I think maybe Delton alluded to this in his presentation, that we are working with NCHS and the PHAB accreditation board, some of the things that I have heard probably like the immunization registry, sharing the data, perhaps that’s a place for us to, as Tara mentioned, set the standards for where we ought to be and the functions that we perform, and hopefully we knowing that that is the core source of the information for an immunization registry, that our agency would consider assistance with that.
But at the same time, maybe there are some core functions that a vital records office could be held in as if it’s deemed what a standard practice activity should be. But that’s a mechanism that we are doing to try to improve what we do and get the dialogue going about what constitutes a fully functioning high quality vital records office. That’s a mechanism we could start looking at.
DR. ROSS: That is great. Having been one who has dealt a lot with trying to get multiple states to see that they are mostly doing the same work and mostly doing it the same way, we have come to say you’re special, but you’re not that different.
I think we should keep that in mind as we move to this afternoon’s discussions about moving into the future, because even though we are in a federated system, it is sometimes very helpful to have states who think they are different. When you peel it back to what are you doing, they discover we’re doing almost exactly or actually exactly the same thing, and that to have created all sorts of wild variation for no purpose is kind of a revealing exercise.
So I’m looking forward to our afternoon. We are, amazing, exactly on time. Thank you all. We are going to reconvene after lunch at 1:30, and start looking to the future.
(Luncheon recess taken at 12:30 p.m.)
A F T E R N O O N S E S S I O N (1:35 p.m.)
DR. ROSS: We are going to reconvene. We have got several committee members who are I guess in the hallway, so they’ll drift in. But it’s now time to bring us into the future directions part of this hearing. We’ve been leading up to this through yesterday and this morning what some of the opportunities are for the future, what the needs are that would provoke changes of various types in the way we approach vital records, the gathering of the data, the use of the data.
Speaking just for myself this has been an incredibly useful dialogue, I’ve learned a lot. As I’ve said this morning I thought I knew a fair amount coming into this hearing, I’ve learned that I don’t know very much. This is a pretty substantial field of effort.
So this afternoon we’re going to account for the evolving technology and change in practices and understand what the opportunities to potentially reinforce and strengthen the National Vital Statistics System might be. To begin with I’d like to ask Dr. John Lumpkin who is Senior Vice President of Robert Wood Johnson Foundation, and also a former state health official of long standing to be our first speaker.
DR. LUMPKIN: Good afternoon. It is a pleasure to be here back at the NCVHS —
DR. ROSS: That is right. Former chair of NCVHS, I forgot to mention that.
DR. LUMPKIN: Where I spent a lot more time on that side of this table, so it’s really a thrill to be here. And to talk about an issue of vital statistics, which has played a critical role not only in identifying causes of death, identifying what’s going on with births, but in innovation related to information in our nation. When we go back to the 1850s and we look at some of the vital records, they were just measures of deaths, and not very accurate.
So the first innovation happened in 1860, when that census included causes of death, and became very important to answer some of the important questions of the time. For instance, in Chicago, the question was if you look at 1870 why were 70 percent of the deaths occurring amongst children under the age of five?
What were the issues that had to be addressed within that community of sanitation, in order to have a tremendous impact? There was a census of 1880 that created the real challenge, because it took almost a decade to complete that census, and it was the thought of how do we improve upon that process of getting data quicker, like over a decade, that led to the rise of automation in data processing.
And the first electronic tallying machine was done to do the 1890 census. Vital records, vital statistics systems were leading the way to innovation. By the 1950s vital records continued to increase. In the State of Illinois where I was State Health Director the first tallying machine in all state government was in the Department of Public Health to do disease statistics. The first computer in 1850 using this kind of memory is an IBM 1401, which had magnetic core memory.
So that is about 1K of memory. Moving forward from that time however we’ve lost the edge on innovation. When we’re talking about microelectronics, and you notice there’s a difference because they’ve actually just come out with a micro SD that has now 400 gigabytes of data, and you compare that to just one byte of data from magnetic core, it shows you how dramatically information systems have changed.
So how do we continue to get that edge? I think at this particular point I would say that the key issues are that the vital records, the vital statistics system needs to be forward facing, embedded in the workflow, meaningful to the data generator, and I’m not talking about the local health department, I’m talking about where the data is actually generated, timely, integrated within the public health enterprise, and granular. So let me talk about how that fits in.
So first is that the vital statistics system must transform from answering yesterday’s questions, why are children dying at a high rate in the City of Chicago, but to begin to answer tomorrow’s questions using tomorrow’s tools. So those kinds of questions include why are we as a nation spending so much on healthcare, and at the same time dropping behind other organized nations, other developed nations, in statistics of life expectancy.
Why are we seeing major changes in the way that we live as a culture, where over the last few decades the total calories have increased, the amount of vegetables have decreased, and we’ve seen a major increase in added fats and oils in our diets. We have changed our very culture. That’s the reason why we at the Robert Wood Johnson Foundation have suggested that what we as a nation need to do is strive together to build a culture of health, enabling all in our diverse society to lead health lives now and for generations to come.
Understanding vital statistics can play a role in that. For instance, if we were to be able to look at granular data, we can take the neighborhood where the Robert Wood Johnson Foundation is built, and a distance of just 10 miles results in a 14-year difference in life expectancy. We can’t usually do this without some sophisticated analyses, because our vital records systems aren’t built to answer these kinds of questions and address these key health issues.
So we will only be able to be successful if we take chances and completely consider new approaches and ways of thinking. Thinking outside of the box, that is the challenge, the vital record system has done that in the past and we need to do that in the future. So part of that is beginning to embed the vital record systems into the workflow.
And let me give an example of how that has happened in another system, or how that’s developing in another system which is notifiable conditions. The data shows that notifiable conditions have a reporting completeness rate of about nine to 99 percent. So 99 percent for monkey pox and Zika, and nine percent for most every other thing. How do we change that system so we actually get the data we need to understand what’s happening with contagious diseases and infections in the country?
Well one of the ways is to change how the conversation is occurring. Through a project called digital bridge we’ve brought together the major electronic health record vendors, healthcare providers, and the public health community to talk about ways to change the flow of data collection. And that data flow collection process has changed. So if you look at the far left, you can see that the first step is that it’s built into the electronic health record.
The System S design is being field tested how, it’s relocation seven by the end of this year, will use the electronic health record that will respond to trigger codes that are built into the electronic health record to generate reports and notifiable conditions.
It will be completely transparent to the healthcare provider. They don’t have to push a button, they don’t have to do anything, and a report is generated. How can we apply this approach to the vital statistics system so that it is in fact generated as part of the workflow, not as an additional step?
The next step is thinking about creating data that is meaningful to the data generator. While it’s true with our system of collecting notifiable conditions, it’s also true when we think about the vital record system. As I was going through some of the history in the data I actually ran across a diagram. I think you have a similar diagram that has been shared to the committee. This one was from a report in 1950. The system hasn’t changed dramatically, the data flows one way.
If you’re a data generator in this system, and I’m not again talking about local health departments, I’m talking about clinicians who are filling out death certificates, what the heck is this system for and why am I filling out this record? I get nothing back. In fact, when we talk to some of the vendors about our system of collecting notifiable conditions, one of them said public health is like a thousand gnats, they constantly pester you but they never help you do anything.
And you heard discussions over the last couple days about the gnats in the vital statistic system, because every state, every jurisdiction wants something different. And is it really necessary? Because until the public health community got together and agreed upon a common set of trigger codes we could not have this conversation around the digital bridge with the vendors, because they will not build a system that has to respond to every jurisdiction with a different set of codes and a different set of responses.
So the Vital Health Record System also needs to be — And so that led to the vision that we have with that is that we will work together to improve the health of our nation by creating a bidirectional exchange of information between public health and healthcare. And think about the ways that the data can begin to flow. Because when you understand why data is being collected and how it impacts what you do you’re much more likely to be actively involved in the collection of that data.
Our vital statistic systems need to be timely, and we’ve heard plenty of conversation about that today and some yesterday, so I’m not going to go over that issue. But I always was struck by the fact that as a State Health Director I’m trying to address the issue of what happened two years ago with infant mortality, and does it really make a difference what we’ve done? Well I’ll tell you in two years. That doesn’t work anymore. Our system needs to be timely.
The system also needs to be integrated. It needs to be integrated within the public health data info space. What do I mean by that? Well let’s go back to an example that was raised yesterday. What happens when a child is born? A fairly common event in this country. In fact it’s probably happening just this very moment.
In the state of Illinois when a child is born what happens is there’s a birth record, then there’s Social Security enumeration at birth, and then there’s Medicaid enrollment for the roughly 45-50 percent of children who are born in this country on Medicaid. Then they may be involved in an immunization rate registry, because they should be getting their first dose of their Hep V vaccine. And then if a child has any birth defect they’re registered in an adverse pregnancy outcome registry, and then of course every child gets a metabolic screening.
For children who are born in rural communities someone from the local health department needs to be notified, so they go out and check the well water to make sure that it’s safe for that child to drink. And you heard also that family case management is an important aspect, particularly for vulnerable children to be healthy throughout their life and home visiting. All of those are separate information systems that pull at the healthcare provider to fill out.
And the reason is because we think about it this way, this is the thought about the vital statistics system. The data on that record, by the way that’s Elvis Presley’s birth certificate, one of the few I could find online, the data on there, the information could easily be drawn from the Electronic Health Record, and could populate all these other data systems.
But the vital statistics and the public health systems are all different silos, even when they’re within the same agency. And it’s because we’re thinking about how the paper will flow. And we don’t live in a paper environment anymore, we have to think outside of the box. And so the challenge is to think about creating this public health info space that healthcare and funeral directors interface with, and it actually gives them something back that’s meaningful in what they’re attempting to accomplish, because the goals of these systems are in fact the same.
And finally, the vital record system needs to be granular. Communities want to know what the difference is in life expectancy, but it even goes a little bit further, one of the projects that we’re engaged with, and we’ve done this with the National Center for Health Statistics, we’re doing a similar one for death data, is to begin to pull the data down to the census track. And we’re looking at various statistical approaches to make that data meaningful because really if we think about building a culture of health we’re going to build those culture of health in a community.
And we have to move from thinking about it at a national level and even at a state level, because health happens in a community where people live, learn, work, and play. And as we move away from thinking about this broader sense to a more granular sense to creating the conditions in which communities can be successful in building a culture of health, they want to know the data. They’re the ones the data is being generated from. So why shouldn’t they know as they begin to address what they believe are critical health problems.
So in the 500 cities project that I mentioned, for instance in Nashville Tennessee you could begin to take the data down and know that in some areas that if you look by census tract and you see the areas that are the darkest, they have the highest rate of smoking, are also the ones with the highest rates of blood pressure, and therefore it’s no surprise they’re also the highest rates of heart disease.
So a community that says they want to address heart disease begins to see data that says okay let’s look at the risk factors in ways that help them understand not just what the government says are the challenges in the community, what they can see and what they can feel as they begin to improve the health of their community.
So to make all this happen there are some key things that need to occur. We need to have governance within government. Think about areas where for instance there have been some efforts in Philadelphia and other locations, and I want to encourage the committee to go or at least talk to those communities about how they’ve changed the governance structure for health information at the local level to begin to achieve this goal.
There needs to be governance within the private sector. This is not an endeavor just for government, it is government working with vendors, with healthcare providers, for funeral directors, and all those who are involved in this data enterprise to think about a seamless system that answers the key questions for each community. And sure you can take the data and parse it out and spit out a birth record, or spit out a death record, but it shouldn’t be designed around those records themselves.
Then we need to create the legal framework to allow that to occur, that facilitates it in a secure fashion, that’s respectful to privacy and allows the data to flow where it needs to go, so that there is a commitment to get the right data to the right place at the right time to improve health.
Albert Dunn, in 1946, wrote a paper. He was Chief of the National Office of Vital Statistics, a forerunner of the NCHS, where he said that each person in the world creates a book of life. This book starts with birth, ends with death. Its pages are made up of the records of the principle events of life.
Record linkage is the name given to the process of assembling those pages. The goal of the vital records, the vital statistics system should be to not only create for each individual that book of life, but to allow those books to be combined in a community, to give direction as that community seeks its efforts to enhance the health of the community and build a culture of health.
DR. ROSS: Thank you very much. Our next presenter is Dr. Kalvin Yu of the Kaiser Permanente of Southern California.
DR. YU: As that is being set up, full disclosure, some Uber driver named Pajak has my notes for this talk, so I’m just going to have to wing it. That was a perfect setup, Dr. Lumpkin thank you. So it’s true, the private sector, we do want data. And I’ve put together a slide deck that shows some use cases of how we at Kaiser Permanente are trying to influence community health. So I think I need to do some level setting here. Particularly on the east coast I think Kaiser Permanente is a black box.
So we are a combination of the Kaiser Foundation Hospital Plan, which owns the hospitals, they’re nonprofit, they have deep roots in the communities often, especially in the California regions, seven decades worth, senior centers, community centers, faith based organizations. And so we have an interdependence in that regard. And then on the other side you have the Permanente medical group, and we take care of the patients and we make the clinical decisions.
And so together we are Kaiser Permanente. And so then our mission statement of improving our patients’ health and also the communities in which they live is not such a weird thing to say, because that latter portion usually throws people off. But we do strive to improve community health, and so I put together some use cases of collaborations with local, state, and national public health agencies, in particular projects that highlight data deficits where we think this group can help influence or at least broker some sort of dialogue.
So the first example is the local level. In this project we collaborated with the LA County Department of Public Health with a two week prediction forecast of cold and flu activity two weeks in advance, and we mapped out hotspots by zip code. We almost used census tracks, because they don’t change as much as zip codes.
But basically, areas of higher risk of morbidity and mortality due to flu and bacterial pneumonia. And so the LA County Department of Public Health shared this with all hospitals in LA County several times, particularly at the height of flu season, to give some situational awareness of the areas and geographies and medical centers that would be at risk for higher morbidity and mortality from flu.
And so I’m bringing this up because in our retrospective analysis, and you can see the grey shaded area is our two week forecast where it’s actually going out two weeks in advance, and the dark blue line is actually what happened in terms of ILI encounters in real life, and so we weren’t perfect, it was about 89 percent accurate. We do think that if we had access to mortality data that was related to flu or hospital encounters from cold and flu that we could kick it up to 95 percent.
And so the reason why we think this is important is the intent of this was to share it with all of the counties in southern California to provide an awareness of what your staffing needs would be two weeks in advance for cold and flu season. And so I want to make sure that you understand that this is more than just a cold and flu situation. Every hospital throughout the nation during cold and flu months have ER throughput issues, a bigger burden in surge capacity issues for more patients that need to be seen.
Our population is growing and increasingly getting older, and therefore higher risk for morbidity and mortality for flu. And so what this essentially is doing is providing some awareness on how to staff up and staff down in the hospital setting, which I think is a really big pressure point for us delivering healthcare, and I think any public agency dealing with health issues really needs to have this on the radar.
Moving a little bit further upstream, and also at the level of the state level, this is a collaboration with the California Department of Public Health, we actually heat mapped low flu vaccine penetration and we saw a correlation with low socioeconomic income. And so to your point of finding out where the pockets of need are and equitable care and access to something as easy as flu vaccines, we helped divert thousands of CDPH flu vaccine to free clinics in these deficit zip codes. And so that was a success in that we were able to do that.
But really to evaluate this program and to let it live further and expand we needed to show a ROI which was basically di we influence mortality/morbidity due to flu, and so that would mean that we need more timely access as you were alluding to, to deaths and hospitalizations due to flu. I think of mortality as much more easily collated since that’s universal, it’s hard to broker with six different insurance plans, hey share your ICD coding for inpatients. So that seems possible, but the timeliness and the format are issues.
And so moving a little bit further, looking at opioid use has been mentioned in almost every talk. And obesity. So we are trying to demarcate the social determinants of health further upstream, obesity as you know leads to cardiovascular complications, myocardial infarctions, diabetes, and opioid use for overdoses.
We are trying to look at younger patients with these problems, and that brings into our radar neonatal deaths, premature births. So we have programs designed to intervene in the community for education and resources, but again we would need access to these types of specific deaths in order to make our business case and really move the dial on community health issues.
And so there’s a common challenge here, in terms of community health we can paint part of the picture but not all of it. And there’s data integrity issues, there’s antiquated processes. So typically, with public health agencies it will be the case that projects like these that I’ve detailed for you die on an email thread, because oh we can’t handle that type of data format, or we only accept CSV data type on Excel, which is alien to our new graduates who do data analytics.
And so I do think it’s possible to get on the same page because we’re actually doing it with the CC, their AU module which is an inpatient antibiotic overuse, sort of outlier, demarcate, has the potential, they’re working with more than 350 hospitals that are downloading data, de-identified aggregate, and really if this works in a big way we can decrease multi-drug-resistant bacteria cultivation that infects two million people a year and causes significant morbidity and mortality, not to mention cost of healthcare. And so this really is a bidirectional learning project where the users know beforehand what format to send the CDC and what to expect.
And so we can quickly operationalize this information, and more importantly there’s a forum in which all of us hospital users have now gotten to know each other and we can trade sort of secrets of data collation which I think is key. And so that brings us back to NCVHS. I think your influence not just on NCHS but as an organization is not to be underestimated.
We would support and really wish for when we eventually submit data to the state and it gets to somehow a federal level, when we purchase it back if it could be in an electronic format and not paper that would really help us. Data integrity is an issue, I’ll give you one example from my specialty of infectious diseases. Most electronic medical records, no matter what vendor you use, there are three to five places where a nurse can put in where a urine Foley catheter is.
So when you report catheter associated UTIs, urinal tract infections, when we started doing this years ago our first data pool showed multiple people with three Foleys. There’s one urethra per person, so we caught this before we submitted it because we have a whole data analytics department.
I worry about small community hospitals where their data analytics team is like one person in the basement, they don’t know the difference between a triple lumen, a Foley catheter, a midline catheter, not all those were reportable for infected devices.
And so I really wonder how confident you are when you’re collating data, when you’re calculating baselines for the nation on how accurate that is. And so those are some of the concerns, and these are some of the ways that we’re using data in healthcare. And so hopefully I’ve engendered some topics for discussion.
DR. ROSS: Thank you very much. That is great. Lots to talk about data quality. Our final presenter is Dr. Chesley Richards who is Deputy Director over Surveillance at CDC.
DR. RICHARDS: Thank you, Dave. I don’t have any slides, because I knew that by the time we got to this point my mind would have changed about what to emphasize. The second thing I’ll say about my job at the agency, I’m the Deputy Director for Public Health Scientific Services, and under my supervision is NCHS, which means I have the impossible job of trying to supervise Charlie Rothwell. But it has allowed me to get very close to this issue and really sort of try to understand some of the challenges but opportunities as well.
In the role that I have at the agency, Dr. Frieden, our former director, charged me three years ago to develop a surveillance strategy for the agency, partly because of many of the things that we talked here specific to registration, but broadly across our agency we have 120 data systems, most of which are siloed and not interconnected, most of which can’t leverage the emerging technologies in ways that are effective and within their resources. And we’re not able to make the data as available to all the various partners who want the data as we should.
I chose at the time four initiatives to get started in areas where there was the ability to touch virtually every program in the agency and in our public health partners in the states. Notifiable diseases, syndromic surveillance, electronic lab reporting, and mortality reporting. And the star of all of those initiatives has been mortality reporting.
If you look at why we’re even talking about this today, it’s in part because of Dalton and Charlie’s leadership to get the timeliness of mortality reports to a point where we would actually care. And I think this meeting is evidence that we do care, and you have a very important role as a committee to chart for us some clear signals and a path forward. So I want to thank you for the opportunity to talk about this.
The current system I think we have to acknowledge functions to achieve our basic needs at a federal, state, and local level. It may not be perfect, but it’s functioning. I think for the future if we want to improve timeliness and quality at a national, local, and state level, we really have to think both about policy, and we have to think about resources.
There are three areas that I want to highlight, and John has already set us up with a beautiful presentation on some of this, but there are three areas that I want to think about highlighting for innovation. The first would be the record, and I agree with John’s point that the record is not the record we’ve been using, it’s the record of the future. It’s really comprised of the data. The data is really what underlies that record.
The collection and submission of vital record data through electronic registration systems linked to EHRs and supported by clinical decision support I think is a very important effort that we’ve already made that has to continue and has to grow. It needs to advance to be able to link locally to other data sources so that we’re not capturing all the data we think we need in that record.
The second area I want to talk about is sharing of the data, that data that resides within the record. And the transmission of the vital record data from the vital registration programs to relevant partners and to national partners needs to be supported by electronic platforms.
It’s not necessarily one platform, but it needs to be supported in an environment where we can use all the tools at our disposal to really make that data move in the ways that it should. I think that’s critical for us to be able to facilitate the record sharing that needs to occur at the state level.
And then at the national level, I think the creation of national datasets and the ability to make that data available to all the appropriate business and research users or partners is really going to hinge on how effectively we can use emerging technologies like block chain for security, natural language processing to really be able to get from the data the things that we’re not able to get now, and cloud computing.
But I don’t think any of these technologies we talk about can yield results again unless we can align policy, resources, and a workforce that’s right-sized with the right skills to be able to do the work. So to me the critical piece out of all of this is really the business model, what’s the business model for vital registration, vital statistics. It needs to be robust. And it needs to be primed not just to operate where we are currently, but to be able to move into the future and take advantage of all the various opportunities we have.
And again, I want to reiterate, your committee through this very good hearing, has an ability to send some strong signals that will help us at the federal level to be able to make the types of resource decisions and strategic decisions to support state and locals in their efforts. So thank you for the opportunity to testify.
DR. ROSS: Thank you very much Chesley. I think we’ve got the opportunity for a good discussion. So I want to begin by asking if committee members will have any specific question to start with.
(Invites Charlie Rothwell and Delton Atkinson to the committee table)
DR. ROSS: Bruce, I’ve got a bunch of questions, go ahead.
DR. COHEN: Trying to figure out how to move forward in this area is an incredible challenge. I appreciated, Dr. Richards, your comment about developing a business model, although it’s really difficult for me to conceive a business model that addresses state legislatures’ reluctance to include information in these systems all the way up to the needs for cloud computing to create interactive exchanges. So if anyone up here can help guide this committee about approaches for these business models, I’d love to hear your suggestions and insights.
DR. YU: This might not be exactly what you’re referring to as far as the business model, but from today and yesterday I heard a lot of concern about data integrity as with myself, and what we found helpful was actually a rubric that the CDC developed for their AU module, which is they actually approached the major vendors that make electronic medical records to support the first few meetings we had in this antibiotic overuse workshop.
And that evolved into a voluntary opt-in by hospital systems to participate, and that has evolved into a quarterly meeting where it really has formed a network of people who are sharing how to collate data in a certain way, and sharing sort of pratfalls like I told you about the three Foleys. And there are multiple examples where we who are sending data, ultimately where the sources are cleaning up our data so that you all don’t have to worry about it as much. And so I think that’s something that could be self-sustainable by industry because we all want good data.
The electronic medical records vendors want good data because it speaks to their functionality and them as a business model of selling a good product. We of course in healthcare want it because we want to take good care of our patients, and you want it because you want accurate data. And so if we’re all on the same page, these coalitions if you will I think it doesn’t solve your problem but it’s one part.
DR. LUMPKIN: I think it is critical to think about this not as what the vital statistics movement does or even what state health agencies do, but it’s the coalition. What do healthcare providers think about? They think about the fact that they have bundled payments, that they are facing changed in the payment system, that they’re being responsible for care after the patient leaves the hospital for the first 30 days, and they’re beginning to understand that what happens in the community is an impact upon their bottom line. And they want data. if the system that the public health and the vital statistics system is a black hole that they send data to they’re not going to care one wit about it.
But if we build the kind of partnership that we’ve done with Digital Bridge, where the vendors are in the room and the providers are in the room, and the message for the providers is very simple, we’re going to reduce your burden of reporting. The message for the vendors is very clear, we’re getting together so that you can actually produce a product that has the end results.
And you find that when you have those conversations with the legislators, it’s not you walking into the room to have them, it’s the medical society or it’s a specialty society, or it’s an organization representing the vendors of electronic health records. And then the message is different and you’re more likely to be successful.
DR. ROSS: Charlie or Delton, do you want to say anything?
DR. ROTHWELL: If I knew what the business model was, believe me we wouldn’t have had to have this meeting. I think there are a lot of threads that have gone through today, and I’m just trying to think of some of them, one of which is the reticence of states, and I was there, to open up its public health law to make any changes, unless it’s absolutely necessary. Because once you do then your legislature has a chance to do things that you never thought they might do.
I think there’s a great opportunity, if we let our private sector folks do that lobbying for us, and not have the state health department do that, but for them to go in and say this is what, if you change your law for us to be able to access this information, this is what it’s going to save us, this is what it’s going to save from a financial perspective our customers, which is your citizens.
And I think you have a much better shot at making very focused changes in the state public health laws. If we can’t do that it’s not an issue of technology, we have requirements in place right now on the NDI which are state based. Those requirements were written up by people like me more than 30 years ago. And we did it to protect ourselves so that we could at least get the money we were giving up when researchers came to us directly for services.
And at the time we had just started to do business with NCHS to send electronic files, data tapes to NCHS, and we thought well why not kill two birds with one stone if you will, and maybe we can convince NCHS to take over the responsibility if we keep control of the data. And thus, that has been the relationship that NCHS has had since that time.
I’d like to point out that NIH funded this activity very heavily for many of the first years, and then Manny Feinleib, who then became the NCHS director required the center to either fish or cut bait, that is either get self-supporting or walk away from it.
And so we became self-supporting by putting the cause of death information on the NDI, and that’s what really got the NDI going. Most of the money from the NDI that comes to us is going back to the state and should be being used, although we don’t require it, should be used to support the very systems that are providing us the data. And that is the crux of what we’re talking about here folks, outside of the issues of not sharing.
From a systems perspective, we need to make sure that this one system, whether it is a combined system with electronic health records in the future, or some combination of electronic birth certificate systems and electronic death certificates and electronic health records, however that works out, we need to make sure that the vital registration process is also supported to do the health statistics activities of it.
And that has been the problem, is that we as a federal, we have not paid our dues. One of my first activities back in the 70s was to come up with what we thought the federal share should be to support vital statistic activity, and we never got agreement, and we still aren’t even close to where we thought things should be. So we’re really talking about a funding issue as from a systems perspective.
But just as important is the issue of changing the laws that allow us to really take advantage of the information that we have. And I completely agree with John, we need to provide this information back to the local. And the strength of the vital statistics is that it is local. And yet we’ve turned our back on them. You’ve heard it from the providers who are giving us the information, we’re saying you can’t have it back? Come on now.
We need to be able to change those laws so that it’s more rationale. I would like to say that it’s uncomfortable for me to sit here and hear all the negative things about what we should be doing, what we could be doing and all this. But I think it’s necessary. But I would like to say this. The states have made remarkable improvement, and they’ve made remarkable improvement with a damn little bit of money from NCHS and others.
I mean we’ve given them — What we’ve done is cajoled, tried to motivate, whatever, and they’ve pulled off a miracle. They’ve really pulled off a miracle on where we are right now compared to where we were four years ago. I would suggest that states want to make these changes. They wouldn’t have done it. It’s not that you have a broken system, you have people who want to do it, they just need help in order to go to the next step.
So it’s a matter of funding, it’s a matter of changing the laws that are there for food reason, and I think this is doable, I really do. I really see in the future, and some of you have heard me say this, I really think the first electronic health record will be your birth record. And this is no different than Hal Dunn said years and years ago, and then we’re going to be collecting information from the electronic health records.
And guess what? The last electronic health record is going to be your death record. And we’re going to have at that point a passive registration system for the United States on health. And we need to make sure that we don’t do anything to keep us from going that route. And so the systems that we put in place, the activities that we try to support should be moving us in that direction.
And it’s also not about vital statistics. I think the medical examiners and coroners are some of the most overlooked, overworked staff that we have in public health. And I put them in public health. Because most of our public health events are really issues of suicide, infant mortality, drug overdose, you name it. Those issues are what they’re looking at, and their systems have to be supported as well, and they have to be embedded, or we have to be embedded with them.
I don’t care how it happens, but our system has to be their system, and their system has to be ours from a core perspective. We don’t need all their information, but we need a good portion of it in order to make sense, and we need to go that way, and we have not really partnered with them until recently, and we need to do a lot more work in that area, and so does I would submit all of CDC.
DR. ROSS: Thank you. There was a reason we needed him sitting up here. So we have some committee questions. Linda?
MS. KLOSS: Thank you. Because we have approached this as next generation, and we want to be forward looking, I would appreciate your thoughts on two subjects. That to tackle the business model of the day suggests that we’re going to design it around the way things currently are, or improve that flow of dollars, instead of jumping ahead and looking at what that business model would look like if we were fully leveraging the electronic health record and having the vitals be bookends to that lifetime record.
And I wondered if there is any projection from your work with the reporting project to think about what the impact of that will be on cost, and even what the university of California example we heard earlier. What does that do to the costs of operating this system? Because I think we need to think of business model perhaps a little differently if we’re thinking next generation.
And the same is true I think with the governance issue. That turns out to be the trickiest part of this probably. Again, to think about spending the next some years to encourage states to tweak the current laws without really making it clear what that future goal is or how it needs to operate in the future might not be the best use of tie.
So I really take seriously this idea of next generation and trying to leverage this committee to do what your committee did some years ago in painting that future of health statistics, and now we really need a business model for that future, not just kind of tweaking the flow of dollars today or the governance today or even the details of the law, because that might get simpler actually.
DR. LUMPKIN: I totally agree. I think that some of the lessons that we’ve learned is that the actual coding the vendors were willing to do, and some of the costs that we haven’t figured out but we will as part of the implementation sides will be the cost to the public health system, because now that they have the capability to receive these records or the records will be potentially sent do they have the systems in place to be able to manage that.
I think that what we will see is that the biggest change may not be in volume because of the electronic lab reporting has made a lot of progress, but the richness of the data because that just generally is what’s on the lab slip. And so we begin to look at it in much more rich notifiable condition.
And the NCVHS is well positioned to begin to answer some of these questions. What’s the future of the electronic health record? Well there’s been an initiative that out foundation has supported called Open Notes where individuals have access at the pilot sites to their medical record.
And even though it seemed that many of the clinicians were opposed to it when they started it, once they actually started open notes, giving the patient open access to all of their medical records, the physicians felt it was really something very positive.
Think about that in relationship to vital records, where we heard yesterday that data reported by the patients is probably the most accurate, about where I was born and what my race is, and now you give the individual an opportunity to have access to that part and make sure it’s accurate.
So then when you pull the data out for a vital event you’re pulling it from a very rich database. So the cost of acquiring that data becomes negligible because in the process of care, in the process of routine preventive services all those fields are being populated.
MS. KLOSS: And the costs shift of the use of the data.
DR. LUMPKIN: And when you think about the model, and I think about things in a couple of ways based upon my experience. First of all, we developed an integrated maternal child health system in the State of Illinois we put into place in 1993, it was paperless. And there was a nurse in Madison County who was one of the first one to use the system, and a WIC client came in and she referred them to a well-child visit, and when the patient left she said oh, my work is done.
And we had very detailed records on productivity, on referrals, there was a decision support. That’s where we need to go with this system, that when they’re finished taking care of the patient or the funeral director is finished working with the family, if they don’t have to do something else, their work is done.
The second experience I think is really critical is the power of data to make decisions. And when communities have that data in meaningful ways, they want to do things. And again, the cost equation changes. In a way it’s something along the lines of not what will government pay me to do, but what can I do from my community.
And we see these in communities across the country, particularly look at some of the ones that have been identified as Culture of Health prizewinners by our foundation, where communities get together using the local data to make decisions about their directions.
DR. ROTHWELL: Linda, I think you are correct. But I would suggest that unfortunately electronic health records aren’t there, and I will just say that we have moved away from doing abstraction in hospitals and physicians offices to go to electronic health records, and we haven’t published what we should have been publishing for years.
And the reason is that there’s no similarity within even many times vendors in how they have put the data out or put their systems out to their hospital and physicians’ offices. I think it’s fantastic that we have electronic health records, and they are going to coalesce for a variety of reasons.
But right now I think we have to have a short range view of how we can really improve things and really meet most of the requirements you have here and then decide at some point in time if this confederated approach that we now have isn’t working and is being held back by state law, then we need to look at something else, for example if in fact the electronic health workers could be standardized in such a way, then we would do business with somebody else, and not with the states directly in order to be able to get this information and then provide back to the states. But right now, if we were to go in this approach we would have to turn off the spigot on vitals.
And let’s look at what the money is. We’re providing about $21-$22 million directly to all of the states and territories for vitals on the VSCP. We’re also providing additional money through a variety of mechanisms with partners, PCORI being one, and others, but primarily, and the NDI, you put that all together, and my guess is it’s less than $38 million. And why are we worried, what’s going on here?
This is like, say maybe we don’t need the sewer system that we have to protect us, and this is the sanitation systems that we have in place. Unfortunately we need to move this up. We have a metro system here in DC that’s not working. We all of a sudden decide that we’re going to make it safe, it’s going to cost money to do that. This is a small bit of money, it needs to come from somewhere, it’s not a huge amount, in order to keep us on the trajectory, and then towards what the goal should be in the future, which I’m not sure how it’s going to work.
MS. KLOSS: I could not agree more. I know this isn’t something you flip the switch on tomorrow, but if we can look at a trajectory and make decisions going in that direction. And it’s not unlike in some ways what this country did when a strong public-private collaborative effort came together around the Electronic Health Record and worked through a lot of the early painful decisions, for better or worse, and 20 years later here we are with a lot of saturation. We did that before, and I think it probably needs to take that kind of initiative again.
DR. LUMPKIN: Let me jump on that one, because I think talking about a new business case is really critical. Recognize that in government you have a paper structured system. And by that I mean the information first goes to a local registrar. And the local registrar then sends it to the state, and the state sends it to the federal government, to NCHS. In a digital environment there’s no reason why that information can’t go to all three at once. So that changes the whole paradigm.
But the local registrar gets paid and has a source of income based upon each one of those transactions. And we’ve heard that there is a transaction between NCHS and the states. So it’s not an issue of whether NCHS gives us the data back to the states, if we have the right environment and the right skill sets that becomes a meaningless statement. Because in a timely fashion a state will be able to do their own analysis and make their own decisions.
But how all that money now changes hands is going to have a dramatic impact. And when you talk about going to the state legislators, the local registrars and municipal leagues in each one of those jurisdictions have an incredible sway in the legislature. So as we think about the transitions, we have to have a vision of where we want to go but recognize that there is some really hard work about redesigning that business model.
DR. STEAD: I think what I want to say picks up on John’s idea of the coalition and Digital Bridge. And it also a little bit plays to what I remind people is how Don Lindbergh got us to how we have free internet access to the medical literature. It was basically because he was able to show Congress it was less expensive to provide free access to everybody than to continue to provide the tapes to the vendors.
And so if we’re going to have a new business model it’s going to in essence have to be that disruptive if we’re going to make it work. We’re not going to be able to jiggle things around people’s current vested interest. So the current digital bridge pilot is case reporting based on standard triggers.
It seems to me that we might be able, and that uses the existing permissions, the providers have the permissions to do case reports, they’re in fact required to. So it all fits within the current legal structure. What if we use that same process, at least for births and deaths that take place in hospitals as another Digital Bridge pilot in which we would auto-create the extract of the registration record, whichever one it was, and transmit it either in paper or electronically to the registration office, however they can handle it.
But we retain it in the healthcare system so they’ve got it natively as an anchor in their own quality improvement, and we report it to a national registry that would operate under the rules essentially of an HIE. Then in essence would that allow us to take changing the laws about access out of the critical path? Because those laws could stay in place for access through the registration office to the legal record.
And so a decoupling again, so they would have access to that, and yet this registry could be for quality improvement and public health improvement, other rules which we can at least under either an opt-in or opt-out model allow people to play, could that begin to break the game open without, I think anything that involves changing the laws of the legislature will take place after our time.
That doesn’t mean we shouldn’t continue to work on that, but I’m looking for ways that we could get out of that. Is that feasible as a next step for Digital Bridge? Because it seems to me that could break open the line. Once you’ve got a critical pass of stuff moving that way people will move.
DR. LUMPKIN: I am going to waffle and say it’s feasible as a future step, not necessarily the next step. And that’s just because I think one of the critical parts of the Digital Bridge Partnership and the governance structure is we want to build in on bidirectional flow, and the notifiable conditions really is much more of a concern to the public health community than providers.
And perhaps the next step for the Digital Bridge will be something that is much more of concern for the providers than public health, but where public health can provide a service. So as the governance group begins to explore that, certainly one of the takeaways I have from this session is that we want to at least put the issue of the vital statistics system as something in the queue to think about as a future use case.
DR. STEAD: Wouldn’t feedback from the registry of mortality data be of immense help to the provider community? That was what I got out of the UCSD proposal discussion. I know it would be valuable to Vanderbilt.
DR. RIPPEN: It’s more from a quality perspective I think, endpoints and for researchers. And some physicians, it depends on your cohort. Hopefully most docs are assuming their patients are living.
DR. LUMPKIN: David and I had this conversation at lunch, I think that most of the policy decisions in this country are based upon what is making people sick rather than what people are dying from. And that really is the priories. And we know how important these records and statistics are.
I think it’s a very good point, and one that we can bring back to the digital bridge, but part of it is building the partnership and preserving the partnership, and expanding it. We couldn’t do that without bringing in the funeral directors into the partnership to have those conversations. So it may be a different iteration of the Digital Bridge.
DR. ROTHWELL: If I could interrupt and sort of go back to both of the — maybe an interim approach is really different. In other words, we go to this bridge route, and NCHS does what it normally does, is that we do a survey for vital statistics, for both births and deaths we could collect the information that we probably need from a national perspective and provide basic data per state.
We wouldn’t need to collect nearly probably one tenth of the records that we’re doing now. We could do that through hospitals and medical examiners offices, and we could probably do that for a lot less money. We’d shut down the NDI. In other words, what I’m saying is we’d still be providing vital statistics for the United States, it would be as valid, you just wouldn’t be able to do small area work and that type of thing, while you take on the future.
To me it’s either shut it completely down and look to the future and wait for three or four or five years until we’re able to tell people what their issues are on infant mortality and low birth weight births and a variety of other issues that we do, or we do something in between, a survey perspective while we’re going to the future, or we try to get this system as it now exists working as well as it can while we work to the future. I think to me those are the opportunities that we have.
DR. RIPPEN: I guess I try to focus on going back to what we started talking about, which is biostatistics. So what’s the scope? Because we already had conversations about adding additional things, we’re talking about electronic health records, we’re talking about all these things that all of us from many times in public health really want access to because we want to use as much information as we can to understand a lot of different things. We’ve also heard a lot about all the different important uses of biostatistics for the government.
So then from a scope perspective we have a national need, a federal need, and we’re getting data from the same source through multiple different ways, so there is the question there of future state we’re still going to need specific information that’s identifiable, future state we still need to match different things together at the federal level that we could actually provide to.
So again, there’s a national component, there is a critical obviously state component, which really they’re the ones that need to use it for their state, and then the local. So we have these three different levels, and then we have the scope of what the information is about. And then we have the question of what is it, because we talked about death and birth because we’re focusing on that, but we talked about marriage, divorce, and we know those are health related to some degree.
So again, talking about all of these things that touched federal needs as it relates to policies, as it relates to payments, as it relates to identity and all these other things. So I think we do have to think about future state on the scope as far as what is it, because we can’t solve everything, all electronic health records, the future.
The other thing becomes then what are the implications in an electronic world in healthcare, electronic health records, where information can be linked. So now it’s not a question of the ever expanding but the question of what information do we need and how do we think about future state of linking as opposed to adding, because you don’t have to replicate everything.
And then we got what are the requirements, because we do need to think through what are the requirements of such a system given what we need. And we’ve heard timeliness, ability to link the quality. We have the question of should it be federated, the standards.
And then we have the overall needs, which to me is a practice that no one is really talking about. I mean we have 50 states and territories, so how do you actually allow that in standardization, and then we’ve got the business, how are we going to pay for it. So we’ve got lots of duplication of effort, we’ve heard it, everyone is doing the same thing multiple times. We need to support the data providers, to your point, we have to pay to do it.
And then we’ve got the appropriate use and governance. Because a lot of the laws are related to people feeling uncomfortable about how data is being used. So the big question becomes how do you enable something to be able to be flexible enough to be able to do that. So if we step back and think about scope and we think about these requirements, what should our scope be?
DR. ROTHWELL: As far as marriages and divorces go, I don’t think that is in scope. I think we’re precariously close to having a working system here folks. So let’s figure out how to make it work as best it can under this process while we look at the long-term advantages of going a different route with electronic health records in some way.
The strength, and as far as linkages go I’m not trying to brag, it’s just a fact, NCHS has probably the biggest linkage program not only in the department, probably in federal government my guess is, with what we do as far as linkages to our surveys with vital statistics by the way, as well as linking our surveys with outside databases with CMS and with HUD data for example.
So we are all for linkages, and that’s in fact in the short-run, maybe in the intermediate run that’s where the future lies, is in linkages capability, and that’s the reason why I would hate to go to a survey for vitals, because vitals gives us this universe to really do some interesting linkages to our surveys and other data systems that we don’t have anywhere else. That’s its beauty I think.
DR. RICHARDS: Charlie, I am going to interrupt here. You do have a great linkage program. But we’ve heard here for the last two days, and you and I have been part of conversations where a whole raft of partners say the linkage rates, the fees that they’re paying are too high. And there’s a whole bunch of business out there that won’t even come because they can’t afford it.
And we’ve talked about and you’ve highlighted that those linkage fees are set by agreements with the states, and that most of that revenue goes back to the states, and they’re dependent on that, and they probably maybe would get more if we reduced the linkage fees and had more people paying into the pot. But we don’t know that for sure, and they can’t run a day to day business without having what they’ve got.
So that’s why I brought up business case. Unless we can delve into that and figure out how we make people whole, at least for a system that as you highlight is near functioning, it’s hard to think about how we then put the energy to getting to the next level of the system. And I’m afraid when I’ve been part of these conversations that we make one decision over here, it has three impacts down the road and you end up with a system falling apart when you thought you were actually helping, that’s what I’m worried about.
DR. ROSS Are you talking about then the possibility of a few state innovation model exercises on business model, where maybe they’re given a five or seven year run at a different model, but just in a few places to really try to explore and understand the full complexity and the pros and the cons of the shift in model? Is there any way to do this other than purely theoretically and then roll the dice and do it?
DR. RICHARDS: I will take a stab at this and then Charlie will tell me where I’m wrong on this. The first time I ever came to visit NCHS Charlie invited me to stay at his house overnight. A delightful experience, but I left a t-shirt that I wore there when I went to sleep. And for two years Charlie has been telling me I’ve got your t-shirt, I’ve got your t-shirt, that nasty t-shirt you had on, I’ve still got it.
But I think that, I don’t think today we know what the business model is, and we say that. The second thing, and I’m always fearful in my role is to say a nice group of people who are trying to do the right thing gets a $5 million appropriation for two years which doesn’t address anything, but we’ve suddenly got more money that’s going to get taken away down the road. So I want to avoid that.
I do think if the committee can help chart out a path with recommendations that really start speaking to what the vision down the road is, and then some transition time where maybe what you do is you have enough funding to do innovation projects that really help with data now to chart while we still maintain what we’ve got, I think that something like that feels like the right thing, I don’t know if you have any points Charlie.
DR. ROTHWELL: Yes. First of all, we know we can show how many dollars are going to which state. It varies each year depending on what the requests are and how it balances, but we know what the dollars are going to the state for the NDI, just for the NDI now.
I think what Chesley is bringing up is that if we were to find funding to replace that so that there would be little or no cost for matching, but the same money would go to the states, then they wouldn’t lose anything, the issue would be how long will that money be there. And we’ve received money in the past from federal government through appropriations that have been tied to one year or two years or three years.
The states look at it, I believe, or at least I did when I was there, is the only way I can ensure getting this money is that I have control over what the charges are going to be for it, and that has worked for the states up until now. They’ve gotten the money for it. But I think it has held them back long-term. So I agree that we need to have some way to make sure that the states know that at least they’ll get this money back, that they’re now getting from the NDI. And I think that would serve a great purpose.
DR. RIPPEN: Quick question just to clarify, so with regards to the finance, the business model, so do you believe any of you that it’s better for the NCVHS to be just a client to the states and actually just be involved in getting data for the needs of CDC and HHS and that Social Security, which they already directly serve as the clients also of the state, and that you aren’t the middle man for linking datasets and things like that, or do you think that there is a role for NCHS to be the actual broker.
DR. ROTHWELL: I think multiple people could be brokers on this. We’ve done I think pretty well in the past, but that doesn’t mean that it couldn’t be a third party, somebody else doing what we now call the NDI, so I don’t think it has to be tied to us.
I think we have a pretty good track record but other than that we would definitely want to use the NDI because we use that to link to our surveys, and that’s the only reason why NCHS said yes to this over 30 years ago, is that they saw this as an opportunity for them to link to their surveys, and we’ve done that ever since.
So that would be something that we’re interested in. Other than that, I sort of, if we weren’t getting pillared by the cost of doing the NDI, to the researchers, I think it’s good for us to deal with the researchers in this way, I think it expands us. So I would hate to lose that tie-in, but at this point if I’m a small, I have a small grant, I might not be able to really afford to use the NDI. And that’s where the rub is.
DR. LUMPKIN: If I could just jump in here and make a suggestion and a challenge to the committee, as important as the business model is, that can be quicksand. I think the first thing, and this was my approach when I was State Health Director, is what does the science tell us, and then how close can we get that in the political environment.
And I think it’s really critical for the committee because you don’t have to implement the program to create the vision, and identify where you want to go, and then based upon that issue, the charge of how do you create the business model to support that vision. If you start tinkering with all the pieces inside the box you quickly find out that nothing is going to work other than the herculean efforts that have been going on in NCHS for decades without straight appropriation. So I think create the vision and move towards that.
DR. COHEN: That is where I am coming from too. I’m trying to think about what the sweet spot is for NCVHS, how we can really add to this conversation to move forward. And I see it in a couple of areas. One is the vision and the roadmap forward for creating the next generation vital statistics and how it integrates into the existing data streams that we need to monitor population health. And the other, I don’t know whether it’s low hanging fruit, yesterday I was struck by the diagram that I think Carrie put up describing the complexities of data collection for the birth process.
I’ve yet to see the analogous diagram for how vital statistics are used, and after listening to folks for these two days to better understand the multiplicity of uses of these data and how the data uses interact I think would help me understand how we could move forward developing a business case for financing this and for the duplications that need to be eliminated.
I really don’t think we have a handle on the integration of uses of vital statistics, and I think that would be something that would be worthwhile for the committee pursuing as well as part of this roadmap forward.
DR. MAYS: I guess I want to pick up on where Helga started and then where we got in terms of jaunt. Because what I was sitting here struggling with is the notion of the scope, and it was really good in terms of when you went through the layers. but there are some other pieces of it, and I think this may be what Bruce is saying, is like are we going to do our regular who’s our customer, and then kind of go that way, or are we going to sit and use the expertise that we have and really do visioning.
And I think this is one of these instances where we need to do visioning first and really figure out what are we trying to solve for whom, because we’ve heard a lot over the last several days. And I mean I’m just trying to figure out like where are we going to go. And I think where we need to go is not to respond to what we’ve heard in the room, but where we need to go is to use what we’ve heard to spend more time on creating a vision of what we think will make a difference.
Because I think we still have players not at the table that would be useful for us. I think we’ve had new players that we may need to learn more about, but I think to me the next step for us would be visioning exercises, and in that visioning exercise I think we will get more clarity, I think we will understand better what our sweet spot is, and what we hand over, just like we handed over our roadmap and somebody could do it in a way we wouldn’t.
I think we need to really figure out what’s the best contribution, but it needs to be the big picture. And the only other thing that worries me is that I like the way that Bill was uncoupling things, but I’m also trying to deal with the fact that the whirl of births and the whirl of deaths, they have some very different contingencies on them.
So I don’t know if we can do it as vitals or whether we need to do it as births and deaths and then see what we come up with as vitals in terms of the vision, because I don’t know, the two sides, maybe I don’t know births as well, I’ve learned a lot, but it just seems like either we commit to uncouple and figure out how to do it on a very basic level, or we vision and then we get into the complexity of it.
DR. ROSS: I am impressed with the work that our committee and staff did in leading up to structuring this agenda because according to my records we are exactly on schedule, this is unbelievable, for actually now opening this discussion more for a public comment. Before I do that though let me reach out on the phone, any committee members still out there, if you’ve been able to hang in, do any of you on the phone have any questions?
MS. STRICKLAND: I do not have any questions. This has been a very interesting, very fascinating topic.
DR. ROSS: Anyone else?
DR. ROTHWELL: I am going to pick up on what Chesley opened up with. I really don’t think we would be here right now if our Division of Vital Statistics, and with the states, hadn’t done some miracle works in improving the timeliness. However, you look at the future, I’ll be very interested. What I want to let you know is that NCHS is committed to supporting the vital statistics system in any way that we can. So if that future involves us or doesn’t involve us we will continue to support it as we are now as best we can, to continue to move us forward.
So what I’m trying to say is that I’m not trying to discourage you from looking at the future and doing a leap. And if that’s what you feel is the most appropriate thing to do, then go do it. But I will say that I think the states have done a great job, and there’s still room for improvement. And anything that you could do in the interim to help them would be I think greatly appreciated by them and all their customers.
DR. ROSS: Personally, I come at this from a perspective of first do no harm. And yet acknowledging that much can change, and a great deal needs to change. And I think we’ve heard it over and over there’s a great deal of opportunity, we figured out the right changes to make, and I think this is one of those unusual opportunities for the National Committee to offer forward a thoughtful roadmap that will start some of this in motion.
It’s been done in the past by the National Committee, thinking back to the late ‘90s I think when John Lumpkin was chair some reports came out that really propelled some of this digital future forward. I think there is great opportunity. So now we’ve got comments.
DR. SAADI: This has been very fascinating, and we really appreciate the dialogue. I will say the states are extremely committed to assuring that the nation gets what it needs. Our Robert Wood Johnson project was census tract. We’re able to turn around those datasets in less than a year, providing the nation with information to be able to do census tract analysis, very localized. It’s de-identified, that’s easy for us, and we were very excited about the opportunities for that. So a great example of how we can pull together.
We participate in a cooperative program together. And I would say many of my colleagues in the room probably say that what we get from NCHS, it doesn’t pay the bills but it does allow us to do some special things. I have to say the best thing I’ve used over the years is an email from Charlie Rothwell a few years ago that said you can use this money however you want.
Well now I buy food for coroners training, birth cohort conferences, that kind of thing. Nobody else can do that. Of course, now people are knocking on my door and it’s like go away. But those little things are valuable to us. I mean we’re talking about a big scope and we’re thankful to hear the visioning.
But there are a few little things this committee might have some opportunities, like I was speaking with Delton about hearing from our funeral director partners about how can we get doctors to sign off on these death certificates. They tell me or tell others, you probably hear it, we don’t get paid for that. Dang you. That’s part of your job. And we kind of shame them into signing death certificates and helping to facilitate for the families, et cetera.
But is there some mechanism through CMS who they all listen to to say that’s part of your quality reporting or whatever the reporting process. You have to sign a death certificate. And maybe you need to sign it accurately. That’s what I understood from talking with Delton is this committee has the opportunity to make those small recommendations that could go a long way for many of us with our partners.
So I would say take some opportunity to look at what you already have in place to move us in some good ways. Birth side is the same way, we are becoming, we are a public health reporting system. I’ll add three programs to what Dr. Lumpkin added, CCHD, pulse oximetry readings, et cetera. We did all that, we’re adding all that to our birth reporting process.
But I get the opportunity to see syndromic surveillance, that’s under me. I wish I had immunization registry like Rich McCoy, that would be awesome to be able to do. but we feed that system. I have infectious disease surveillance under me and I see all the innovation that CDC has put together in those areas, and then I ask why not for us, for vital records.
And we do have some of that initiative that’s coming forward through the PCOR and a number of the analyses. But then you see the HIT infrastructure with the money going to hospitals. And are they going to be there for us in the future? They’ve got all this technology, but then we had to receive all that. Were we ready? CDC did pay for some things through our ELC funding, and we do get quite a bit.
And Digital Bridge, we’re a pilot state. We’re really excited about that process. But I look at what we do for vital records and for public health, and it’s very frustrating because we don’t get to benefit from those initiatives. And I keep saying to Delton, and I haven’t had a chance to say to Charlie, let’s work with what’s already in place at CDC in the vision you have for surveillance and see how we can improve our systems.
MS. WEBSTER: I want to speak a little bit about the idea of building the business case for the business model. And just a warning with that sort of direction of thinking, because I’ve worked with vital records at NCHS for, this is my sixth year I think, and ever since I started at NAPHSIS we’ve done business model after business model after business model after business model.
And sometimes the smaller set of things around one particular system, but sometimes for the everything, the all. In the vital records system, the whole NVSS. And I know NCHS is doing some of this work with Mitre right now. Mitre right now is working on the business model, particularly for the NDI.
But you can’t make a business model for one part of an enormous system that includes all of the 57 vital records jurisdictions and registration. So I would caution you not to go too far down that road without coming to either NAPHSIS or NCHS to do sort of a summary, here’s what we’ve done year after year for the last six years in creating these various business models, and trying to put them into practice.
I think the challenges that we’ve run into have been enumerated several times over the past couple of days. I think where the committee can be most effective is in the area of helping us with the statutes. Now, I know it’s hard, and I totally agree with whoever it was that said that it’s not going to happen in our lifetime, I totally get where you’re coming from. However I think with the right kind of leverage it could happen much faster. And that’s what we haven’t been able to get, at least at the NAPHSIS level, is the right kind of leverage.
And you may be the right kind of leverage. I don’t know. I don’t know if you are or not. But you could be the right kind of leverage to help promote the model law, to help the federal agencies promote the model law, to help change the model law maybe. To get that endorsed in a way that’s really meaningful and will have some kind of leverage and a ripple effect into the states.
You also may be able to influence the business community that we heard from in a way that will encourage them to go directly to certain states who right now can’t participate or share data with non-government agencies. We’ve been trying to get that done for three or four years now. And we haven’t been able to get anybody to really do it for us. So I think that’s a really important way that maybe this committee can help out in a very real way. So I hope if you go into a visioning process that that part in particular will be a part of your visioning.
DR. HOGARTH: I had just had one quick comment, thank you. And that it’s odd to me that the NDI can be used for research purposes, so there was some degree of beneficence from an ethical standpoint that made that argument stick with all the jurisdictions, but yet quality improvement is not in the same realm. So I’m just curious as to how you pulled it off, because you did it without a change in law, I’m assuming.
DR. ROCKWELL: I think if you write your proposal correct, quality improvement will pass the NDI test. It’s when it gets into more the administrative activities within the hospital that it would not. I think this whole area as far as how the NDI can operate and how it can’t, again we made out of whole cloth in the late 70s, and it’s up to the states to look at it but I think there is some wiggle room there in my opinion. But that is up to the states to do, and if they wish to change some of the access rules we’d love to do it.
DR. ATKINSON: Let me say, in addition to that, over a year ago this whole issue about NDI access, use, can and can’t, was one of the reasons that I wrote a small proposal to PCOR to provide us the funding to give us an opportunity to give a step back and evaluate NDI. As a part of the particular evaluations we have provided a contract with NAPHSIS to help us look at some of the state laws, statutes, rules and so forth, look at that compared to what we are now operating under.
We plan to have a workshop sometime in the March-April timeframe of 2018 where we bring together the states, where we bring together the researchers, and some of you around the table will be individuals that I want to have around the table where we can talk through some of these kinds of issues and to see whether or not there’s some opportunities, some of the work that MDR is doing for us will be used as part of that discussion, and also some of the work that Terry Deutsch who is on loan to me from CMS, as you see I go around and get resources from wherever I can, on loan to me from CMS to help in the evaluation.
So we’re doing some strategic interviews with a number of the agencies, NIH and FDA and others, where we talk about these particular issues. But again, it’s all going to culminate in a workshop around NDI and the issues of access, use, and can we get to that point of that goal that was talked about in my presentation of minimizing the cost to researchers while at the same time assuring the payments to the states.
DR. ROTHWELL: Maybe we drop back. I’m going to be really interested in seeing what comes out of this, because a possible solution will be that we don’t have a complete national NDI anymore, what we have is a really good functioning NDI with 35 states, and shame the rest, and it will change over time.
So in other words who knows how this might come out. We might decide that we can do more business under different rules for those states who are willing to allow us to operate in a more open environment, with different charges and with different access. It’s hard to know right now, and maybe that’s the right way to go. I don’t know.
DR. RIPPEN: It goes back to scope and what is the purpose. Is the purpose to get a snapshot so that we understand the trends are happening in the United States, or is it to actually have the knowledge and details in the ability to have a more granular?
DR. ROTHWELL: Okay. I am not talking about the vital statistics system, I’m talking about the National Death Index. So for a researcher who wants to do mortality assessment of their clients, of their patients, then maybe what we have is a system of 30 states, that operates more cheaply and more efficiently. I don’t know, we have no idea how this could turn out.
DR. ROSS: Charlie, how does that dialogue with the research community happen? Is it something you all do and they just react, or is there a forum that they —
DR. ROTHWELL: Lillian Ingster is the Director of the NDI. Would you mind?
DR. INGSTER: This has been very interesting. I do a lot of outreach because I go to a lot of conferences to promote the NDI, and I talk to more people on a daily basis than are in this room, and I hear a lot of conflicting viewpoints. The language for research is in our contracts with the states, it’s very explicit.
It says in the contracts that NCHS has with the states that the NDI data is limited to helping medical research, which by definition then excludes things like surveillance, because surveillance, and folks like FDA will tell you point blank surveillance is not research, which means therefore I can’t have them do an NDI link, and so we have a perpetual battle of wills over that.
Insofar as having 30 states, that’s a terrific idea, except I also get complaints from the researchers even when we’re just missing a few of the territories. They say well your data isn’t useful if it isn’t complete, because then we can’t find all the snow bids that moved here and there and wherever. I have all the people with for argument’s sake HIV in New York City, well they all die in Puerto Rico.
So if I don’t have the territory I’m missing them, and I’m not of use to the researchers. Well not me personally, I think highly of myself, but the data isn’t. So these are lovely discussions, but from my perspective they’re academic, they’re great. But I’ve got to sell it now. So it has to be comprehensive, it has to be accurate, it has to be timely, and god help me if I know how to make it cheap. If you’ve got other questions I’ll answer them, but I hear this every single time.
DR. ROSS: I have a reaction to the word cheap. I think what we are confronting is being able to quantify the value in the ways that those who benefit and get that value can value it. We argue price before we understand value, and that’s something we need to work on.
DR. INGSTER: I just wanted to add real quick, I did have one client say to me at one of my many, at booth spaces where I talked to them, I’m not advocating this, I’m simply telling you what they said, they thought that the NDI ought to be like a toll road where it’s a pay as you go sort of thing. Except then at least for those of us who live in the Maryland-Virginia area, I won’t speak for other states, we all know that the toll roads that are privately owned and charge very high fees, nobody uses them. So that’s a thought. But this was just what somebody said, and I’m not suggesting it, just telling you.
DR. ROSS: Tragedy of the commons. Denise?
MS. LOVE: I hate to bring up my world, but so much of it is colliding here, and the conversations. And so I’m of two minds, and I’m just going to throw out a wild idea —
DR. ROSS: For those who don’t know your world, you might just say what you’re talking about.
MS. LOVE: I work with states that have hospital discharge data, and all payer claims databases, and not all states are at the same place, nor do we expect them to be. And we’re doing pretty well with the leading states, and I have had some success going to states, the have-not states in the past, and displaying what a have state can actually do with emergency department data and with hospital data and now with claims data, and trigger their interests, because I find that states don’t like to be at the bottom. They aspire, many of them, to be at the top. So I think there is a little bit of shaming that isn’t a bad thing.
So my wild idea is you could have two NDIs, you can throw things at me, one that has some limited statistics but counts on the BIG, but really those leading 30 states, maybe get a little more money but also are in the complete NDI, but have two levels so that you are meeting some of the researchers needs but you are elevating the floor, because I do think a little shaming may be in order. I mean I’ve seen it work, it doesn’t always work. But researchers always want everything, and with hospital data I hear the same thing.
And HCUP, the Healthcare Cost Utilization Project hears the same thing. And you’re never going to have 100 percent. And with all payer claims we’ll never have 100 percent. But 70 percent is a lot better than zero, or noise, depending on the scope of what you’re trying to do. But I’m just throwing it out there as sort of a random question as having two cuts of NDI, one with rough counts and big numbers and one that’s the gold standard.
DR. ROSS: We’re just about out. Calvin, you get the last word.
DR. YU: I did not want that pressure. A comment which ultimately is a question. So the CDC has been very successful in influencing public policy and mandating reportable healthcare required infections. Antibiotic stewardship is now on the joint commission. And shaming does work. So all you really needed was a couple of states to go all-in on that, California was one, New York was another, those happen to be two states that I practiced in.
I guess my question is when healthcare groups like mine are faced with those regulations, we really socialize it and sort of believe in it ultimately because we know that there is a return on interest afterwards, if we prevent a certain number of hospital acquired infections you save hospital days and so forth.
And so I guess my question is since state governments are responsible for paying for the care of the uninsured, and since they are always struggling with resources in terms of the state funded healthcare systems, wouldn’t it make sense when you’re linking mortality data to mitigating downstream consequences of certain disease states, a couple of which I showed, if you can show that you are mitigating hospitalizations, utilization of healthcare resources, that turns into dollar signs, and that essentially funds itself. And that’s been done by the CDC for healthcare acquired infections. And so what is wrong with that model?
DR. LUMPKIN: Just a real quick anecdote. The chair of my appropriation committee once introduced me at a party, and he said this is John Lumpkin, he always makes a lot of sense, we just never give him any money.
DR. RICHARDS: Thanks for bringing that example up. We started on that road when I was a deputy director of the division that deals with hospital infections at CDC, and it wasn’t we that started it. Do you know who started it? Consumers. Consumers Union went around and got mandatory laws for hospital reporting, when we told them not to do it, in 30 states. And they created a movement. And that movement I can remember having a vivid conversation with Lisa McGifford at Consumers Union where she said Chesley, if you guys don’t get off the duff and start doing it we’ll put you out of business in terms of your systems.
Now fast-forward, CDC is able to do what you’re talking about because we partnered with the Consumers Union and we drove a national effort when we had a lot of resistance to that. And I think it has paid off in terms of higher quality care for people and patient safety. But you may think about in this context what’s the role of the consumer in all of this and where does that partnership get developed.
MS. LOVE: The states are great copycats, I remember those days, and the state would bring it up in the legislative session saying we are one of the have-nots.
MS. WARNER: I have just a quick comment to follow up, but changing the subject a little bit. Sorry to do this before the break. But I wanted to pick up on something you said about the differences between vital births and deaths. We’ve been talking about them as one lump, but there really are quite different ways of data collection frankly.
And a lot of the discussion earlier was about medical records and what we can get out of medical records, and I think for births we can get a lot for medical records, because most births happen in the hospital or in some kind of healthcare setting, whereas with deaths many don’t happen in the hospital, particularly my area is inter-epidemiology, and over 50 percent don’t happen in the hospital, and it really varies by type of death, for instance for drug overdoses over 75 percent don’t happen, any kind of healthcare intervention at all before the death.
But it also affects other kinds of death where people die suddenly, unexpectedly. They die at home. Some people don’t like to interact with healthcare at all, so there’s no medical records, it’s difficult to find those medical records. And it’s a lot of different people. So I think that we really have to think differently about how those data collections are done.
MS. ZIMMERMAN: First of all, I wanted to say I appreciate the task the committee has before them. I’ve been on some committees through the HIT policy committee workgroup, and have been on your side, and having heard the conversation for two days know you have a lot of challenges ahead in terms of just trying to get scope and where to go.
So in that light from my perspective both from where I sit in the state and from my previous experience being in your shoes on a different topic, I might think about suggesting maybe sort of short, medium, and long-term recommendations. Because I think the visioning session was great, but I think we can’t throw baby out with bathwater, we are where we are and states can’t move that quickly, hey just can’t be as flexible to sort of say okay, we’re going to start getting everything from electronic health records. And I’m certainly a proponent of much more interoperability and integration, so that’s one thought.
The other thought I had was with regard to comments on model law. And I fully support model law. And I think getting NCLS and NGA involved in that is important, but I do just want to say that having been on some amount of committees through NGA and some other things, model laws are great, but not all states are willing to go there, because privacy in particular is very variable in different states.
We’ve learned that way back through immunization registry work, through our HIE work, through our APCD work, I’m coming from a state where we’ve been a little oxymoronic on how we deal with privacy, in some ways public health was very broad, our HINAPCD is very narrow, and so I wouldn’t put all my eggs in that basket. I think it’s great to go forward and do it and promote it, but I just caution that I don’t know that that strategy always works, but I think bringing in the NCLS and NGA would be important in that.
And I know they’ve done other work around privacy and sharing in that regard. I just wanted to share those thoughts. And my last thought is around sort of value in business case. Again, I live a lot in the HIE world, and business cases for HIE have been discussed for a very long time, some have made it, some haven’t. Some have made it with business cases and some haven’t.
But one of the things that I’ve learned also through the work we’ve done is that sometimes you think you have a business case or your partners tell you they are ready to do something, if you produce a product or a service they are going to be there to buy it, but then they don’t really think on their end when you’re ready, they’re not. I think there’s a little bit of customer dialogue that happens. Sometimes it’s a leap of faith that you’ll get the foundation and you’ll build services on top of that. Sometimes it works out great, sometimes it doesn’t.
I’m not sure that that’s a really helpful comment other than a practical lesson learned comment that I just thought I would share. But sometimes you do have to get the foundations and build the services, but if you’re doing that or what the customer wants, you have to really work with them to be ready to accept the data, take it and do what they said they want, because sometimes you are ahead of them, and I’m more speaking on HIE side.
But there could be some relationship here in terms of just thinking about the dynamics of business case. I also lastly think, again I think return on investment is important, and I think in the fraud, waste, and abuse area, with deaths there is the return on investment to be gained, and a small amount of that could come back into the infrastructure if things were set up properly.
DR. ROSS: Very good. We have a new visitor here, I believe our new Surgeon General, is that correct? Please introduce yourself.
SURGEON GENERAL ADAMS: For those of you who don’t know me my name is Jerome Adams, I was sworn in last week as the Surgeon General. I hit the ground running with the mini-hurricanes we’ve been dealing with. I heard you all were here, actually ran into my friend Dr. Lumpkin in the hall and said what’s he doing here, and found out you all were here, and I wanted to come by and just say hello.
Thank you all so much for being here. Those of you who know me know that I am an immediate past State Health Officer from the State of Indiana, and dealt with vital statistics quite a bit. Two of our big priorities were infant mortality, lowering the infant mortality rate, and certainly vital statistics very much plays into proper tracking of our infant mortality rate.
And then obviously dealing with the opioid epidemic, and one of the big concerns that we had in regards to the opioid epidemic is we still have no clue what the true magnitude is, because across the country we do not know who’s dying, what they’re dying from, and Indiana, as in many other states, we are a home rule state. So we can sit here in this room and have all the best ideas we want in the world, but if we can’t convince the person who didn’t finish high school who ran for coroner and got elected, to implement our plan, then we’re still not going to get the statistics that we want and need. So what you’re doing is critically important.
One of my major themes, my motto for my time as surgeon general is going to be better health through better partnerships. Some of you have heard me speak before, we need to step outside of our silos and we need to rethink what our silo is. And I mean even beyond our health and health oriented silos. I don’t know if they’re represented in this room or not, but I appreciated the comment about the legislature.
We can come up with the best ideas here in the world, but if we can’t convince the legislatures to enable us through privacy laws to actually do it and to fund it through the budget, we’re not going to be successful. The law enforcement community, there’s a lot we can do by leveraging our knowledge about vital statistics with the statistics that they have from a low enforcement point of view, again particularly in regards to the opioid epidemic, which is one of the Secretary’s priorities.
And so with that, I just wanted to say I look forward to working with all of you all over my tenure. I think what you’re doing is critically important and under-appreciated, and so that’s why I’m so glad that you’re all here. And if there’s anything that the Office of the Surgeon General can do to assist, please let us know.
I’d also like to introduce you all, I hope all of you know, but to our immediate Past Acting Surgeon General, Admiral Trent-Adams, who is going to remain on as our deputy surgeon general. And if you need either of us please let us know.
DR. ROSS: Thank you for coming in. Welcome to the National Committee for Vital and Health Statistics, Population Health Subcommittee. Our Chair Dr. Bill Stead is Chair of the NCVHS. Bill, anything you want to say.
DR. STEAD: Thank you for stopping in. The work that has been done in combination between the National Center and the states has gotten, as Charlie said, a system that’s almost working that is drawing attention to the art of the possible, if we could actually produce a system that went the next mile. And so we’ve had two remarkably informing days, and now we’ll start the hard work of trying to figure out what the next steps might be. So we very much appreciate your interest and support.
SURGEON GENERAL ADAMS: The other thing that I would say, I don’t see anyone here that I recognize from ASHTO, but if you don’t have a state health officer in the room you need to have a state health officer in the room, because they’re where the rubber meets the road in terms of implementing the policies that you all put out there, they’re the ones that are on the state level that are going to have to stand in front of the legislature and say, this is why you need to change the laws to allow us to do the things that we want to do.
And so please, please, please make sure you’re including them in your discussions moving forward, and you’ll also find out some of the more practical barriers. I will tell you as a State Health Officer from Indiana, and I don’t mean this at all in a pejorative way, but a lot of the discussions occur on the coast, and they occur with people who don’t realize what a state health officer from Indiana or from Kansas or from Arkansas or from Texas has to deal with when we’re standing in front of a legislature.
So please include state health officers and a variety geographically of folks from across the country in these discussions so that we really can move forward. And thank you for giving me a little bit of time to speak with you all. I didn’t mean to jump in and hijack your meeting, but my office is just down the hall so I figured I might as well.
DR. ROSS: Thank you for taking the time and effort and trouble to pop in on us, it’s a great surprise. We will take you up on your offer. We’re always looking for another good experienced capable pair of hands to add to any recommendations we make. This is a recommending body, and we understand that, we’re trying to make sure that we make recommendations that actually can have some traction and lead to something better. So it’s going to take a lot of deliberation, it’s been a really good two days.
I think we, to your theme of partnership, have managed in this couple of days to bring in almost all possible partners, some of whom I was not aware of at all in the private sector, who benefit hugely from these data and can be enormous advocates to help public health if we help mobilize them. And I think we’re seeing that now and looking really creatively at trying to examine what the future potentials are. So we will be talking with you sir. Thank you so much.
SURGEON GENERAL ADAMS: The business community too, the one other aspect I’d bring up, and you brought up the business case and ROI, a lot of times we think about our health business partners, but I had an interesting conversation with a high level executive at General Motors, and I told him a story that I frequently tell folks. The number one cost for most Fortune 500 employers is salary. The number two cost is healthcare. He actually stopped me, he said no, the number one cost for General Motors is healthcare.
And so bringing in those business partners who can leverage the vital statistics information that you all can help us provide is certainly going to allow you to better make that case in front of the legislature, again particularly a legislature such as the one we have in Indiana, who is all about jobs, jobs, jobs.
And I think what you all are doing is important to jobs, jobs, jobs, but we’ve got to bring those partners in the room and make that case. So thank you again, and I will leave so I stop taking over your meeting. But I again appreciate the opportunity and look forward to working with all of you all.
DR. ROSS: One quick question rom Helga, and then we’re going to do a break.
DR. RIPPEN: This is really detail stuff. If you’re going to compare the quality and kind of the, I would say the validity of the data, and compare it to let’s say social security that’s also collecting the data, are they comparable as far as quality, completeness?
DR. SAADI: It is the same data. They just want FACA death. They don’t want all the additional information that takes us quite a bit of effort to provide for public health.
DR. RIPPEN: I figured it was the same source, I just wanted to have it on records because the implications of that.
DR. ROTHWELL: There are many strengths to the vital statistics system. The real strength is the demographics of it. If you take a look at the demographic information you get out of medical records, I’ll just leave it at that.
DR. SAADI: Lillian reminded me, they do purchase data from us, and we have frankly been in negotiations with them to kind of get an upgrade in the amount of money, they say billions every year, we know that. So they do get information from families for example, but they don’t have to, funeral directors through our EDRS do not have to submit their form 721s, but some still do, some families still call up, they’ll put on their website, SSA does, you have a responsibility to call us up.
But that record goes into the DMF if they get it from families, versus what they get from the states. But it is, our contracts we send daily data to them. And it is the fact of death, just basic, every state participates. And on the birth side absolutely the same thing, that’s been going. I don’t know was that what you needed to hear.
DR. RIPPEN: It was just a question of is there any difference in truth between multiple federal datasets, that’s all.
DR. COHEN: All the data ultimately comes from the state vital statistics systems. They percolate up.
DR. RIPPEN: But you have to remember with CDC there was a method that they used to enhance the data to correct for example for Native Americans. So again, I’m just trying to figure out, that’s all.
DR. COHEN: We can break and talk about it.
DR. INGSTER: Just a quick note. Some people in the room will remember the big 60 Minutes spiel on the Social Security data and the fact that one half of one percent are alive. That doesn’t sound like much until you figure 2.5 million people, and one half of one percent is somewhere between 12,000 and 13,000 people.
Now, because they get all the data from the states, eventually that gets corrected. But in the short term, when the only data they have is from sources other than the states, there’s a certain degree of error built in until it gets corrected by the state data as it comes in. So that’s what happens to the Death Master File, especially in the early years, the first year or two.
Similarly, and there’s papers out there that you can go find in PubMed if you look for it that shows that there’s an age bias, you’re missing a certain percentage of young people because people don’t think to report that to social security, because they’re like oh well they’re 20, they’re not getting social security. So you miss that group until the state data comes in.
DR. ROSS: Tell you what. We are going to take a 15 minute break, which does put us 10 minutes behind schedule, but I think given the hour we need a 15 minute break. Be back at 4:00, please.
DR. ROSS: We are going to need to reconvene. Committee members please. We appreciate everybody staying the course. In planning this, we asked our friend and colleague, Dr. Gib Parrish, to take a stab at synthesizing two days of a huge range of ideas, thoughts, contributions. So Gib, I know you’ve worked miracles before. Stream of consciousness or whatever works, please go for it.
DR. PARRISH: Thank you very much. I think we need a vision, and maybe next steps. And then we need a few short, medium, and long-term objectives or recommendations. That’s my summary of what I think I’ve heard the committee saying in the past hour and a half or so. In any event we have about half an hour, and the goal of this half an hour is to try for the committee to talk about and think through what to do next.
I’ll give you a chance to exchange ideas about that, build upon what we’ve heard, I think somebody put it well, we’re going to build upon what we’ve heard in the last couple of days, put it to use, and try to think through what might be a way to move forward here in terms of the report that the committee is going to have to put together related to this particular project. The discussion has been very good, and I’m a little reluctant to try to just lead you in some direction.
Vickie was commiserating with me over my role here during the break, but I think the discussion was going well, you were putting out I think some good ideas and thoughts about ideas of how to proceed. The one thing I would say that I heard in the last session and I think is a key point here, we talked about business models, we talked about funding, we talked about a number of other things, different possibilities for the whole system.
I think it is very important in this kind of situation, I’m going to go back partly to the scope issue, but to remember what vital records and vital statistics are used for. I mean ultimately those uses remain no matter what kind of system you’re going to try to define. You still have to have a way to establish people’s identity using birth records, you’ve got to have a way to establish fact of death for various legal purposes.
You’ve got to be able to facilitate various planning within the health realm. You need to help facilitate various kinds of commercial activities that utilize mainly death records to do their jobs. So there are certain uses that are going to be there no matter what, and that you’re going to need whatever you think about in terms of recommendations and a future vision, that those will need to be served.
By some system that’s here in the United States and that has the force of law, that’s going to be really key to what goes forward. So with that let me just stop and see what people have to say about how we should proceed, and I’ll try to moderate the discussion. And this gives also an opportunity to Dave and Bruce to speak without having to worry about their role as moderators too.
DR. COHEN: I am just going back to our guiding question, how do we transform today’s vulnerable vital records data collection network into a system that produces accurate and timely information, supporting a breadth of local, state, and federal data needs. I think we heard lots of explanation of what those needs were.
And I really think Amy had a really great suggestion about focusing on, as a committee, what we can do in the short term, develop a work plan over the next year, and then intermediate and longer-term visioning kinds of activities as we discuss what vitals might look like in a broader context and in the future. So I really think we need to develop a short-term strategy and a long-term term strategy. And I’m particularly struck by a couple of things that Charlie said. It’s a perfect example of how this federated system needs support, the states need to do stuff, and the feds need to do stuff.
The states really need I think to focus on expanding their ability to share data for the multiplicity of uses that we heard are needed, and figure out how to provide the data in a timely fashion, if not for all 57 jurisdictions then for whoever can do it. And there are potential strategies that can focus on that. A couple of my NAPHSIS colleagues suggested that accreditation, which was briefly mentioned, could be a real leverage point.
If vital statistics systems were credited through that kind of process, and that accreditation acquired timely sharing of data with a variety of users, and included certain required elements, state vitals folks could go to their legislature and say we need an accredited system and we need to meet these criteria in order to be accredited. So that’s an example.
Certainly, we’ve heard, I think one thing that we need to act on as a national committee is to endorse the development and incorporation of medical examiner data and support medical examiners in an integrated death system. So that’s from the state perspective.
My federal takeaway is I really need to, and I think the committee needs to better understand all of the users, all the federal users of these data and how to develop a cost model from the feds that incorporates the supplying of vital statistics data to all the federal agencies that need it. So I think these are kind of short-term things in addition to longer term visioning, which I’ll save for later from me.
DR. PARRISH: Other thoughts, Dave?
DR. COHEN: Why don’t we go around the table?
DR. ROSS: I am after the two days compelled to also support the notion of short, middle, and long term. To me on the short-term I think we have to understand who benefits and who pays. We’ve heard repeatedly it appears to be an imbalance in how this fairly jury-rigged system is paid for, but with no clear understanding of what we would change. So understanding who benefits and who pays is a start, leading to then some better more thoughtful analysis of a business model, which ultimately then does propel a new model.
But we’re not at all close as I say it to leap to say what should be changed. I think there’s also I would urge that we take advantage as public health often has to do of current problems. If you look at CDC, Clifton Rhode, all that new building came about after 9/11. When I was head of operations in the mid ‘80s we put forward a $600 million building proposal to Congress, and that sat there for a decade and a half until there was a big event, and then new buildings happened.
We know right now we have several crises that are very relevant to the vital records community in being able to help inform. The opioid crisis, Congress will do something. And better death data would be hugely informative. Antibiotic resistance. Pandemic preparedness. There are issues where better vitals will be relevant and relevant in the current funding context.
I think that leads to then funding maybe the current approach, but at least it starts to accelerate innovation, which also gives us the breathing time to put forward ideas on transition to modernization. So we’ve got a lot that we’ve heard, a lot that we can learn from. It has been an incredible education opportunity for a green bean like me in this field. I thought I knew something walking into this hearing, I realize how little I knew. So that’s my two cents.
DR. STEAD: I resonate with all three of your remarks. And I do think starting with, I like Gib’s suggestion of let’s get a discrete set of uses, to the degree we can let’s subdivide them if in fact it may be that we want uses of birth, it may be that we want uses of death, so that we can see opportunities to decouple.
And if we keep the uses and users straight, and I think that would provide a way to let us make choices about scope, almost like the environmental scan let us do in the measurement framework. Then I think where the money is currently flowing we need that diagram. It sounds like Charlie we have it or at least the central pieces, and maybe what you have, what NAPHSIS has, we put those together, maybe we would have where the current money is flowing.
And then to maybe take a few urgent cases, and opioids is obviously an example, I think the idea of pension and insurance, there are a few key cases. Then in essence your short-term recommendations can be designed to address some of the root cause pieces we have with those as the justification for the why it matters and why it’s important.
And it seems to me that’s probably a reasonable starting place, and it may well be that whether we want to make that round without actually putting together a next generation vision, it may be that it’s useful for us to get one round of work done that’s more ground level, and then put the vision together and say how do you go — that gets some of our own critical path making.
I think that we do better as a committee when we operate at the two extremes. We’re pretty good when we’re visioning, and we don’t have to be anchored into okay, what do you do with that tomorrow. And we’re pretty good when we’re saying what do you need to do that’s doable, actionable, in the next six to 18 months. We tend to not be very useful at the big range in the middle. So those are a few thoughts.
MS. KLOSS: I don’t know where to begin. I think in addition to kind of defining, as you said starting out with a clear definition of the space, I do think it would be useful even if we’re just doing a preliminary report to do some kind of description on the current state, and possible future state.
I mean I don’t think we have to lock into it, but I think we need to do enough to say current state doesn’t meet our definition of next generation, there’s opportunity that needs to be exploited. So I think we can stop short of doing a full vision. But doing something to say we’re looking forward somehow.
And then, I think the importance of that, at last from my experience, is it allows you to ask the question when you’re laying out short-term recommendations how do these fit with where we think we’re going, and it may help prioritize those short-term. For example, we had a lot of recommendations come through about standards. I do think there is probably a short-term standards opportunity that could frame some suggestions.
We certainly heard a lot about timeliness, and we can say that that’s going to be an issue no matter where the vision takes us, and there might be some short-term things that can be done to really help accelerate the good work that’s been done on timeliness. And I do agree with taking advantage of the current state. The one that also came out in addition to opioid was fraud mitigation, and I think that probably will resonate broadly, and it’s more of a business rather than a clinical thing. And I like the idea of doing some description of where the money is flowing.
And I suspect that when we think about the future, what fundamentally starts to shift a little bit is where the data is flowing, and the emphasis, shifting away from putting so much of our energy into getting the data, but repurposing data that exists and allowing resources to shift to making the data better and making it fully available.
It seemed to me, I made another note, that we really heard from all of our terrific panelists, it was like a stakeholder analysis, and I think it would be possible to do some kind of table that would lay out the stakeholders. And I agree with Bill, I think we need to do it for deaths versus births, because that will kind of suggest again who has a stake in this when we start thinking about some public-private discussions and other opportunities for maybe improving the current business model.
I think it would help us really identify are there any stakeholders that aren’t here, but boy we’ve really had the people who’ve had a stake in this here the last two days. So I’m comfortable with doing short and long-term, but I would suggest that we give a little bit of time to the long-term to help guide what short-term decisions we put forth.
DR. PARRISH: Vickie, now you are on the spot.
DR. MAYS: But I don’t have to do what you do. I’m agreeing with kind of what we just said. I think that we should talk about short term fixes and long-term vision is kind of the way that I would think about it. And in the short-term fixes I would actually use the disaster as the thing that I would anchor around, because I think it has the greatest momentum right now about fixing things to understand those fixes will help rebuild.
I worry about trying to take on the opioid, because the quality of the data, the things that we need, I think they’re expensive, unknown, and difficult to get for a short-term fix I think it may be for the visioning, and a long-term kind of approach we could do that. Substance abuse, it has suffered for a long time, and there are a lot of things that are needed. And again, it’s mortality in general. When we start moving into that I think we get into kind of a bigger area that needs a lot more fixes than just the vitals itself.
I think about that, but I like the fraud, because I think it really gets us far, it’s really within some of the stuff we’ve done in other hearings, I think that’s one of our strengths if we do fraud, and I think for disaster may be the other, because we kind of heard from people. The thing that I’d like to also suggest that we think about is when we do this visioning is that we do something similar to the way in which we did the roadmap, and that is making sure that we have the stakeholders like today.
I mean I think the things that we heard today, if these stakeholders hadn’t been here we would be in such a narrow place. So I think if we’re going to do the visioning we want to spend some time saying how can we involve these individuals without asking them for a huge amount of time.
The sense I got in talking to several people who are here is they would be willing to help us out. I’m seeing nods. I think we want from a process standpoint to figure out how to involve who was already here as presenters, and to make sure that we ask ourselves who else do we need to have at the table with us.
DR. PARRISH: Alexandra.
MS. GOSS: I did not know it was so formal. I am not used to being called that. Lots of different thoughts. I do think a current environmental scan sort of report is really something of service, that we provide in a number of prior experiences, and I think they become good references for people so they can kind of learn from the journey and not reinvent the wheel.
I do think the aspect of a disaster is an example, whether it’s opioid or environmental disaster helps to capture the hearts of citizens and legislators and funders, and so I think that’s an important thing to leverage, and I really do commiserate with your concern about the opioid aspect, but it is front and center so much it’s kind of going to be hard not to tiptoe into that one.
This concept of partnerships between the feds and the states, it’s very interesting as we’ve gone through this dialogue over the last two days, I hear both sides of the coins when we talk about things, and they’re almost conflicting at times and hard to figure out, so if I want national standards but then I don’t want to get in the middle of states’ rights and laws, there’s a big pushing and a pulling.
I also feel this sense of when we start to use the data in ways we didn’t initially envision, as technology is taking off, as health issues are taking off, and it’s really hard to plan for what you don’t know yet. But this repurposing of data, some of the themes that I’ve heard from quality to timeliness to confidentiality and protections, linking the right people together, the hammer for consistency, the federal rulemaking to make everyone kind of get in line.
There’s a lot of commonality across the NCVHS body of work as a whole, and I made that comment I think on day one, and I think to the longer term visioning comment from Linda we’ve got to think about the predictability roadmap for national standards for administering financial transactions, how that converges with clinical data, how does that fit in with the privacy protections and the beyond HIPPA work, how do we get the data quality and terminologies and vocabularies down pat along with these public health purposes.
And it feels like a great opportunity for a perfect storm to weave a good story and a long-term vision together, and that’s going to take us a little time, but I think we do have to almost start there and then almost work backwards and make sure that we aren’t forgetting any key players at the table. I’m sure I have some more thoughts but that’s enough for now.
MR. COUSSOULE: I will try not to repeat what the others have said that I agree, and I’ll just focus on a couple points that I think were relevant for me. One is it has been a couple of incredibly educational couple days for a lot of information I frankly just did not know and understand, so it has been very educational. It would be pretty easy after listening to the testimony the last couple days to kind of view this system as broken.
I’d argue it’s not, I’d argue it’s more of a work in progress with lots of value being generated, but more work to go. So you have a tendency, I think many of us have a tendency to try to fix things, that’s why way we view the world. I think I would be remiss to try to push back and say that without recognizing all the good stuff that’s actually really happening today. So I think we need to keep that in mind.
I’m also a big believer in the plan big but execute small kind of world. So understanding both structurally are we trying to accomplish things different over a long time horizon, but by the way I’m also an operator so I need to figure out what differences can be made in the short term. There’s an interesting piece to this where there’s a significant private and public benefit to what’s happening here, which is relatively unique in my experience.
And so I think the value propositions are varied, and I think it would be interesting for us to try to keep what we’ll call the users and the uses straight in this one. The data collection side is probably easier to get your hands around, although probably the most difficult thing to solve, but the use cases can drive a lot of that I think.
The other thing I would raise that Alex talked about a little bit is that I think we need to be very cognizant that the regulatory environment may have a tendency to stifle the creativity and use of some of this. And that came up a few times in the meetings, and it doesn’t mean that we want to ignore privacy and security and confidentiality issues, but we need to be thinking that those are not to be hindrances and how do we get through that, as opposed to those are roadblocks to stop some of these things from happening.
Because it didn’t come up a whole lot, there were bits and pieces of it that came up the last couple days, but I think we could get caught up in the restrictions on us from a statutory perspective could stifle a lot of good thinking. I’d like to be creative about the thinking process and then realize how to work within that framework to make sure that we are diligent and thoughtful about the protections at the same time, the uses and use cases need to drive the game.
MS. LOVE: Again, I will try not to be repetitive. Good comments all around the table. I’m just trying to translate it into my world. And we heard about disaster. But I think we have a money disaster coming into public health and surely into state health departments, and that has to be thought of very carefully. And so what does that mean?
And we’re struggling with similar things of how to communicate to the public and to legislators that this investment is worth it in a time of money crunches. And so one of my business coalition folks used the term, you know we need to communicate in terms of dashboards, not population health.
I mean if this was another time, but I’m bringing this up in a brainstorming sort of way, so you can all say she’s way out there, but what is a public health dashboard. And he likened it to we spend so much money in healthcare in this country, and population health, public health, and saving the chump change for our data systems is like going out and buying a Bentley and saying I don’t want a speedometer because I’m going to save some money. And he was saying why can’t public health communicate that way.
I’m kind of throwing it out there that isn’t, and in my world we’re thinking episodes of care, this is short-term. But as I sat here the last few days I’m thinking birth and death are the ultimate episodes of care, they’re the bookends. But what’s in the middle could be creatively populated as I listened to the bookends to cancer, what happened to these people when they got diagnosed and how they died, or this circumstance of birth and these kids, this group of kids.
So the episode is depending on the use case the middle, what’s happening, and how much did that cost for this child that didn’t get the prenatal care, and how much did it cost to the system, and do some predictive analytics and really get in there, because ultimately, I hate to say it, but legislators and funders think in terms of money.
I know health is great, but they’re thinking how much did this kid who was born early cost us in that burden. So I’m sort of shifting, so those use cases are huge, but in terms of episodes. And again, intermediate, I heard accreditation, and NAPHSIS might be another intermediate step to improving our data systems, including vital statistics where you set some thresholds and some expectations and link it to training and improvements.
DR. PARRISH: Helga. Last but not least.
DR. RIPPEN: Yes. Actually, I made a mistake, everyone had all these great comments, stole my thunder, what can I say? Except I guess I just want to reiterate what is kind of important that Linda highlighted, which is I do think we need to have a certain, it can be a fuzzy vision, but I think we do need to know which direction we’re going.
And the reason for that is because as we start doing the short and the medium-term kind of things, if we kind of miss opportunities to build in that direction, it will be in a limited resource environment, you can’t afford a rebuild. And especially when I think that there might be opportunities for funding, and it’s a double-edged sword, for example for opioid, where we know that there are significant challenges with the quality.
However, this is an important issue, and there might be funding to actually improve, as opposed to building a siloed surveillance system to be collected independently, which now actually causes increased problems for future, because now you’re not actually incorporating it into the death record for example. So I think those are really kind of important things.
The other things that weren’t actually brought up that I think we may want to consider is this concept of the standard minimum from a federal level that everyone can agree with on the state, because every state has different needs, and because states use it from what I’ve heard for different purposes, that they have the ability to do that, but if we have kind of a common baseline I think that’s important, which then implies you also need governance, that there might be an opportunity to have conversations across the states and territories to then say well we added this, what do you think, and that it’s really kind of a state collaboration, which I know they already have the organizations and everything, but to really partner with the federal government.
And then even within the federal government the users, it’s federated to some degree too with regards to who uses it for what purposes and things like that. I agree with regards to the importance of environmental scan, I think that’s an important component. We should also include all the federal users, because I know we just hit on a few as far as where their data sources are and things like that. And that of course is a whole list of things, but the day grows long.
DR. PARRISH: Dave reminded me we need to check, is anyone on the phone that would like to add any comments to those that we’ve heard?
MS. STRICKLAND: Deb Strickland is still on the phone.
DR. CORNELIUS: I’ve been on the phone the past few days. We have so much work ahead of us and such great input, and so it requires quite a bit of reflection to figure out how to move things forward.
DR. COHEN: I really like the idea that several folks suggested around use cases. I think we should do at least one from births as well as deaths. So if we choose opioids or disasters for deaths I would suggest for births we could choose either Zika or early elective deliveries, because that really highlights the current value of birth data collection systems. So I just wanted to make that addition.
DR. STEAD: I think the other thing we should think through is that we need to figure out how we can activate the stakeholders who were here or not here to break up and do different parts of this work. I think some of it at least they may already have in hand, we just need to ask for it. Because having spent a certain amount of time in our charter, our role is to catalyze the community to do what as a community we need to do to address the problem. So I think we need to sort of as we get this list maintain the connections and try to divide and conquer to the degree practical.
DR. PARRISH: I know that we are essentially at the end of our time for this particular session. Bruce reminded me that we need to take just a few minutes to see if there are any comments from the audience as well as here. Charlie?
DR. ROTHWELL: First of all, thank you all very much. Why we have a FACA committee is to get an outside look at what we’re doing. At times it’s painful to hear, but it’s necessary. And so when we push back a little bit at you it’s just a normal defensive type of thing that happens. During a race in Indianapolis you would like to have a car of the future, but what you want right then is to make sure that damn car that you have right now is operating as good as it can while you plan for that future car. And you can help both with the current car and what the future is.
And by the way, things have changed in CHS. When we’re talking about the rapid release program of the vital statistics, we would never have done that before. We always waited until the end of the year, and we massaged, we made sure that the data was as good as it possibly can. We can’t do that anymore, because all of a sudden the data we’re holding is useful for surveillance purposes.
And yes, there’s great problems with the drug overdose deaths and how they’re reported. But the past CDC director, we were talking, and it wound up that I’m holding the most current information that government has on that, and that what I have is better than anybody else, and I’d better damn well report it. And he was absolutely correct, and the current CDC director feels the same. And so now all of a sudden NCHS is in the surveillance activity. And so I think that’s an example of what can be done.
And finally, I’ll just say next week there’s going to be something called a festival, which all the federal government gets together and they talk about the futures and how this current administration is operating, and the last day they call Bold Friday, and they went out for 200 requests for innovations in government that really made a difference. They picked four. Guess what one of those four were? It was a mortality system of the United States, and they want to use the opioid overdose deaths as an example of what can be done and what should be done better in the future.
And so I’ll be giving a TED talk next Friday on that. I’m a little uncomfortable doing it because there is so much more to be done, but it also shows what has been done through Delton and through the states. And that’s what my point is, we’re not as bad as we think we are, but we could be a hell of a lot better. And whatever you can do to help us with that vision would be extremely helpful to me and others who come after me.
DR. PARRISH: Thanks, Charlie.
DR. SENS: Mary Ann Sens. This will be short. I just want to make my rural advocacy. There’s no question the opioid epidemic, I use it myself, it’s a tsunami. But it’s in the larger context of very poor mental health, and widespread addiction. And in the native communities that I serve that takes the form of methamphetamine. It takes a different form in many communities. I think we need to latch onto the largely Caucasian frankly opiate epidemic.
And I’m not minimizing it, it’s awful. But it’s not the be all and end all of addiction, and I think as public health people and certainly as a medical examiner and physician we need to recognize the addiction, mental health problems, and the struggles that our minority communities have, that our veterans have, that our brothers and sisters have, and not just limit it to one segment. Use that segment, it’s got their ear, but get broader services.
DR. PARRISH: Last comment, and then I’m going to turn it over to Dave because he’s got something he needs to do.
MS. WEBSTER: You bet. It will be super short. I just want to reiterate how valuable this kind of gathering really is, and not just for the intended purpose, but for the 40 other purposes that get handled while we’re all in the same room together, for the first time or the third time, and just the happening of it causes all kinds of connections that will bear fruit, maybe not tomorrow but eventually.
And in that way things like the Joint Public Health Informatics Taskforce, JPHIT that Denise and I are on, and that era, so we already have these connections because of JPHIT, and those kinds of collaborations are so important, and I just wanted to reiterate that, these gatherings, the other collaborations that are happening, the AASHTO Affiliate Council, all of the things that we do together as an association community to represent our constituents, there are economies of scale that can be taken advantage of there, so that we’re not reinventing the wheel, so that we’re using the same systems in a really efficient manner. And I just wanted to bring that up before we closed. Please consider that when you’re creating a vision.
DR. ROSS: Alright. It’s late in the day, a highly productive two days in my view, I think everybody agrees we’ve learned a lot, Nick, unlike you, walked in. I thought I knew a little bit about this, realize I’m a slow learner, I’ve been at it for a long time, and I didn’t know but a fraction of what all this is about. So my hat is off to the teen that planned this, Kate and Rebecca and Bruce, you really put together a wonderful group of stakeholders to inform us, and I think we now have our plate full about the next steps. There’s something else we want to do first though. Let’s go up front.
This seems to be the right time to acknowledge that our colleague Delton Atkinson who announced he is going to retire at the end of this year I think, this is kind of a grave blow to the vital records community. But we, the National Committee, Delton I want you to come up here, I wanted to acknowledge a career very well spent, and in honor of that career the National Committee wants to present you with this token of our appreciation, thank you for all you’ve done, continue to do, and I know will continue to do, even if you do retire, because we won’t let you. And thank you.
DR. COHEN: This is a sculptural rendering of a circular plaque on a rosewood base commissioned by NCVHS inscribed for Delton’s lifelong commitment to public health. As Delton so eloquently said he’s been in the trenches from everybody’s point of view and has spent a lifetime of service to vitals. So we all appreciate it.
DR. ATKINSON: For one time in my life I am speechless. My wife should be here to hear that. But let me say that all of my years in public health and in health statistics I have enjoyed every role that I have been in. And a lot of it is due to the people that I’ve had the honor to work with, to learn from, and to grow from.
And I tell you it’s been a wonderful career, I probably won’t do December, Charlie and I are talking about March, my wife is talking about something else, but I do appreciate this, and I really appreciate the opportunity to have the opportunity to work with each and every one of you, and I hope to stay around to be a nuisance if anything else in terms of it, because I’ve just come to love this area of work, and that’s kind of why I’ve stayed with it for all this time.
DR. COHEN: The wrap-up is pretty straightforward, not only should we be thanking Kate and all those in the planning committee, Helga and Vickey were, I addition to Dave and I, were o the planning committee. And everybody has done a just phenomenal job recruiting folks.
I really need to thank the participants who were incredibly thoughtful and just opened up all of our eyes to lots of different world views and who have given us a lot to think about as we move forward. And moving forward will be developing a report, and certainly will be before we issue any report share it with all the participants so they’ll be able to make sure that we quote them and reflect their ideas accurately.
One of the other things that the National Committee does is provide a letter to the Secretary with our recommendations. We’ll certainly consider providing recommendations to the secretary. All of the slides that you’ve seen over the last several days will be posted on our website. There’s a lot of really wonderful information in them to review.
And I think this is such a huge topic, we’ll also look for other ways to publicize and promote the learnings from these two days. So we’ve got quite a future ahead of us in terms of what we need to do. And as Bill said, there are tremendous resources here in this room, and we hope to call on you to help us with these activities. I want to thank everybody for making these last two days just an incredible learning experience, and hopefully one that will move vital statistics forward for a long time. So thank you all.
(Whereupon, the meeting adjourned at 4:50 p.m.)