[This Transcript is Unedited]
Department of Health and Human Services
National Committee on Vital and Health Statistics
December 3, 2014
U.S. Department of Health and Human Services
Hubert H. Humphrey Building, Room 705-A
200 Independence Avenue, S.W.,
TABLE OF CONTENTS
- Plenary Session – Call to Order – Bruce Cohen, Ph.D. and William Stead, M.D.
- IOM Committee Report on Including Social and Behavioral Domains in Electronic Health Records – William Stead, M.D.
- OMB Guidance on Administrative Data Sharing – Susan Queen, Ph.D.
- Working Group on Data Access and Use – Vicki M. Mayes, Ph.D., Damon Davis
- Committee Discussion – Priorities and Plans – James Scanlon
- Committee Wrap-up
- Public Comment
P R O C E E D I N G S (9:00 a.m.)
MR. SCANLON: What we will do this morning is we will go around the table. I am going to ask just the members to introduce themselves and whether they have conflicts, then we will review the agenda and plan for the rest of the day.
(Introduction of Committee Members)
MR. SCANLON: Good morning everyone. We are going to start off this morning with Bill Stead briefing the full committee on a report I think just released last month from the Institute of Medicine, dealing with social and behavioral domains for content of electronic health record. It is something the committee has always been interested in. We are kind of looking at how we can improve that whole side of it.
There is a lot of interest. I remember we talked yesterday about the SES dimensions in terms of healthcare. There is also quite specific interest in thinking about making risk adjustments for SES factors, in terms of quality of care, such things as hospital readmission and so on. Of course, that depends on us having the information to do that. Bill, let’s turn it over to you.
DR. STEAD: Thank you, Jim. As this slide summarizes, we had a remarkable committee that spanned public health, informatics, social and behavioral science, and clinical practice. The expertise also ranged from pediatrics through geriatrics.
Our charge was to identify domains that could be considered by the Office of the National Coordinator for stage three meaningful use to determine criteria that should be used in selection and then to identify domains and their measures that should be included in all electronic health records, consider the implications of incorporating them, and then identify issues and opportunities in terms of linkage with other data systems.
Because of timing, we did the study in two phases so that we could release an initial report that could be earlier in the deliberation phase of the stage three work and then a second report that was released November 13th with the final set of information.
The work started with identification of conceptual frameworks that would allow us to systematically scan for domains that should be considered. The committee, in the end, selected three as the basis of that work, one by Ansari et al., one by Kaplan et al., and one by the MacArthur Research Network. These frameworks looked at both upstream and downstream factors and let us, therefore, have a holistic view. As we worked through that, we identified five levels of domains that should be considered. It showed demographic, psychological, behavioral, social relationships and related living conditions, and then the built and social environment.
We identified six criteria, but we used two in identifying candidate domains. One was the strength of the association of – the strength of the evidence of the association of the domain with health. The other was its usefulness. By usefulness, we considered three areas. The first was improving health for the individual patient, including decision making between the clinician and the patient for disease management and treatment. The second was in describing and monitoring population health to inform both policy and intervention. The third is the value for conducting clinical and population health research in order to accelerate discoveries regarding the causes of health, outcomes of treatment, and interventions at multiple levels.
Based on these criteria and within the five levels, we identified 17 candidate domains as being best suited for inclusion in all EHRs. That was really the ending of the phase I report. With these two, we then applied the other four criteria. The candidate domains had already been rated for association and usefulness. In phase II, as we assessed the measures for those domains, we looked at whether there was a standard representation of a reliable and valid measure that was freely available, i.e. no intellectual property issues, second, the feasibility, considering burden placed on the patient, the clinical team, and the health system of incorporating the information, third, sensitivity or patient discomfort regarding revealing personal information or increased legal or private risk, and finally, was the accessibility of the data from another source. If information was already available in another source, there would be a lower priority for including it in the EHR.
This slide summarizes the cascading decision of our work. I already mentioned how we went from the five conceptual frameworks to 31 domains based on those frameworks to the 17 candidate domains. Then we looked at those 17 candidate domains and we identified measures that would meet the criteria. We focused on 31 measures. From that, we then constructed a parsimonious set of measures that was concise and complementary. That turned out to be 12 such measures for 11 domains. It takes 25 questions to assess them.
This slide shows the way that we compared the measures to each other as we tried to construct the parsimonious set. The Y axis is the readiness of the domain for inclusion in the EHR. That is the minimum of the committee’s rating on availability of a freely available standard measure, feasibility, and lack of patient discomfort. We used minimum rather than average because if the measure was problematic for any one of those three, we felt it wasn’t ready for inclusion in all EHRs. For usefulness, we narrowed the three perspectives I showed you as we were assessing domains and focused on the usefulness of having the information that resulted from the measure in all EHRs, which really meant usefulness in the clinical setting.
The colored symbol that is in front of each measure, the blue triangles, the green, gray, et cetera, represented the committee’s overall judgment of the priority of including the measure in all EHRs. The number in front of the Q in parentheses represents the number of questions that were in the measure. Chapter four in the report lists all of the questions in each of the measures that we recommended. The bolded items are items that are already routinely collected in EHRs.
After considerable deliberation, the committee decided to include all of the measures in the right hand column in its recommended panel. That means we weighted usefulness over readiness. It does not mean that the other mean that the other measures that were not selected are not important. However, it does mean that further work is needed, either to refine the measure to something that is feasible in the clinical context or to clarify the intervention that is feasible in the clinical context. In essence, the things that we didn’t select set up the research agenda going forward.
This is a slide that Bob showed us yesterday. It summarizes, again, the measures at the top that are already in EHRs, but not standard. One of the key differences here is we are actually proposing standard questions that everybody would use in obtaining this information. It is not just asking that people have the information and letting them figure out how to do it and then figure out how you make whatever that result is interoperable. We are actually trying to standardize the questions at the front end. It also shows in the right hand column that these measures – some of them are captured at different frequencies, some only on entry into the program, some that need to be done more regularly.
As a basis of this work, we came up with several finding sand recommendations that I will walk through briefly. The first is that four social and behavioral domains of health that are already frequently collected in the clinical setting, but that the value of this information would be increased if standard measures were used in capturing the data. Based on that, we recommend that ONC and CMS should include in certification and Meaningful Use regulations the standard measures recommended by the committee for social and behavioral determinants that are already regularly collected. These are race/ethnicity, tobacco use, alcohol use, and residential address.
The second finding was that the addition of selected domains together with those four domains would constitute a coherent panel that would provide valuable information on which to base problem identification, the clinical diagnosis, treatment, outcome assessment, population health measurement. Therefore, we recommended that ONC and CMS include in their recommendation the standard measures for those eight domains, which you have already seen.
The benefits of painting this information are in three categories. The first is more effective treatment for individual patients, for example, by identifying depression and inadequate physical activity, can improve the treatment for patients with diabetes. Second, more effective population management for healthcare systems and for public health agencies, for example, a health system which finds a high rate of socially isolated patients could provide group visits or patient support groups. Third, enhanced opportunities for research that can inform new treatments and interventions to improve individual and population health, adding social and behavioral data to clinical and genetic information, for example.
Now, as we looked at implementation issues, these recommendations obviously involved modification of the EHR technology. However, the more challenging issues relate to the expanded view of the determinants of health and adaptation in the way clinical teams work and how patients participate in their own care and wellness.
Most of the recommendation measures rely on self-reported data. In other words, there are things patients experience so they logically have to report it. That could be done on paper or via computer. The practices in health systems will need to consider workflow design to figure out when it is most appropriate to capture it, how to review the information with the patient, when to intervene.
Data can be shared to reduce the data capture burden, but there has to be transparency about information sharing. Who will it be shared with and why is critical to make sure that both the patient and the clinical team know the answers.
Links to community agents in public health can enable concerted intervention for individual patients, but may require two way consent. While additional time is needed to collect such data and, more importantly, act on it, the committee felt that the health benefits of addressing the concerns outweigh the added burden to patients, providers, and health system.
Based on that, our next finding was that standardized data collection and measurement are critical to facilitate the use and exchange of information on these determinates. Therefore, at this juncture, EHR vendors and product developers lack harmonized standards to capture such domains and measures. Therefore, we recommended that the certification process be expanded to include appraisal of a vendor or product’s ability to acquire, store, and transmit and download self-reported data germane to the social and behavioral determinants of health.
The next finding was that the addition of social and behavioral determinants to EHRs will enable novel research. The impact of this research is likely to be greater if guided by federal prioritization activities. Therefore, our next recommendation was that the Office of the Director of the NIH should develop a plan for advancing research using these determinants collected in electronic health records. It seems logical that the Office of Behavioral and Social Science Research should coordinate the plan and ensure input from across the many NIH institutes and centers.
The next finding was that advances in research in the coming years will likely provide new evidence of the usefulness and feasibility of collecting these determinants beyond that which the committee is recommending at this juncture. Therefore, there needs to be a process for considering this additional information as it becomes available. Therefore, we recommend that the Secretary of HHS convene a taskforce within the next three years and is needed thereafter to review advances in the measurement of social and behavioral determinants of health and to make recommendations for new standards in data elements for inclusion in EHRs. The taskforce members should include representatives from ONC, CMS, CMMI, in particular, AHRQ, PCORI, and the NIH and also research experts in social and behavioral science.
This study was unusual in the breath of the sponsors that supported it. There were nine that represented different government agencies and foundations. We think that really shows the readiness of this topic for action. The full report is downloadable as a free PDF off the IOM website. With that, I will take questions if Jim wants to give us time. We are over a bit because we started a bit late.
MR. SCANLON: We have time for questions. Very nice discussion. Nice report. Any questions for Bill?
DR. QUEEN: Aren’t the Meaningful Use standards for race/ethnicity using the OMB guidelines?
DR. STEAD: The current stage two recommendations are the OMB guidelines. You will see that that is on our chart in addition to the census questions that we recommend. We judge the Census recommendations to be more useful in the clinical context. Therefore, we are recommending that the measure be changed.
DR. QUEEN: Did you include gender identity along with sexual orientation?
DR. STEAD: We included gender identity and sexual orientation. They did not make the cut primarily on the basis of usefulness. That information is in the report.
DR. QUEEN: I think in March, didn’t the ONC – their committee had a recommendation, preliminary recommendation for certification of EHRs just having the functionality to collect both gender identity and sexual orientation.
DR. STEAD: You are correct. I think the committee had the unique opportunity to consider this idea of a concise – complementary parsimonious panel. If we had a magic wand, we would change that to certifying the capability of collecting self-reported data such as each of these domains. Then we would recommend including the core panel.
Institutions, the VA, for example, is already collecting that data. Vanderbilt is going to be interested in collecting it. We put the questions in chapter four. They are there. They just didn’t make the parsimonious panel. We are hoping the places that want to include them will go on and include them with these standard measures.
DR. SUAREZ: Thank you. This is a terrific report. I have two or three quick questions. First, I think some of the things you are recommending have already been – I think you mentioned have already been implemented. The very first recommendation on page seven about four social behaviors, including race/ethnicity, tobacco use, alcohol use, and residential address, that is already a capability of EHRs.
DR. STEAD: Yes, and we mention that those items are there, but they are not being collected in all EHRs with the same standard measures. What we are recommending and we have put in chapter four are the specific questions that would be asked to obtain that information.
DR. SUAREZ: So there is the expectation – what ONC does is they establish certification for EHRs, meaning an EHR must have the capability of –. That is as much as they say. Must have the capability of –. Then someone else, CMS, says now you have to activate that, use it, and start capturing that data. What you are saying is that the report shows the definition of the data and the type of questions to capture the data – the definition of the data as in what are the actual code sets to be used to code race and ethnicity first and then the questions, how to ask the questions. Are those the two additional things? ONC has to define the standard and the certification. CMS has to define the requirement to capture it. Then there is this additional process. I am trying to understand.
DR. STEAD: Understood, and it is a good question. This slide may help get at the answer to the question. Right now, the stage two measure for smoking is the one I have the example on. It uses the SNOMED code set. Over time, it has used the evolving CDC Smokestack. What the committee recommends is two scientifically validated questions that have been used together and scientifically validated together. You can then – if you ask those questions, you can create the answers to SNOMED. What this is attempting to do is give you a scientifically validated way to ask the questions. It doesn’t change what code set you would use. That can be whatever you wanted to use. If we ask the questions in a scientifically valid way, as the code set changes, we will actually maintain the consistency of the measure, itself.
DR. SUAREZ: Just a quick follow up on that. On the right hand column, it is not a measure. That is actually a standard code set. That is not a metric or a measure that CMS defines. That is a standard that ONC defines for certification. What is missing from here that I don’t see is the actual measure, meaning what is the percentage that we are required, as providers, to capture tobacco use electronically in electronic health records using the standard code set that you have under the stage two measure.
DR. STEAD: Very helpful clarification. In our first meeting, we clarified that the threshold for Meaningful Use, the percentage, was out of scope of the committee’s work. Our job was to come up with the standard way of obtaining the information. It wasn’t our job to come up with what the threshold should be.
DR. SUAREZ: You are adding, actually, another level of standardization, if you will, which is the question, how to ask the question. That has not really been conceptualized in a regulation in Meaningful Use as I understand it. In other words, providers are told – you must use a certified EHR. The EHR must have the capability to capture race and ethnicity. The code set for race and ethnicity is the OMB standard. CMS defines the metrics saying 15 percent of the people, you must capture that or 20 or 30. You are moving it one stage farther by saying and on top of that, the way to ask the question should be this.
DR. STEAD: Very close. We are saying the question that should be asked is this. The way you ask it remains to be worked out in the workflow. There is still another level beyond this that we didn’t get into. You are absolutely right. We are adding the fact that we should be using a consistent, scientifically validated question or question set to obtain the information. That is the major add from an interoperability point of view for the report.
DR. SUAREZ: Thank you.
DR. NICHOLS: Bill, really nice job. I applaud your work. I think you should be co-chair of everything. I was struck at how adroitly you slid over, well, you know, the committee considered the time cost and we decided that the benefit was greater. Was there an estimate of the average time added to a typical encounter?
DR. STEAD: No. We debated long and hard if we could do that. Our belief is that since the information should be self-reported, there will be no or little time added for obtaining the information. The expense is going to be on intervening with the information when you have it. We could not come up with a way of estimating that.
DR. NICHOLS: When you say self-reported, do you mean we are going to hand them a form and they are going to fill it out while they are waiting?
DR. STEAD: That is one way you could do it. If you look at the work out of Promise, there are now very well worked-out adaptive testing for people that use computers for how you would obtain this information. It could be a form. It may be that there has to be a person that is interviewing you, depending on the literacy of the individual. It will be highly variable how you should actually obtain it. That is why we said this is the question, but not how you should ask it. The committee – we didn’t have the expertise to model that in a useful way.
DR. NICHOLS: It is going to be different in a patient setting, clearly.
DR. COHEN: Great work. A couple of comments. Are these minimum – the definitions minimum, maximum, or required in the sense that in Massachusetts, for instance, we feel that the Census categories don’t really capture the information we want. We will require hospitals to provide more detailed ethnicity information that can be up-coded into the Census codes. One concern is that establishing minimum criteria sets a bar too low for folks who want to use the data for other purposes, more detailed purposes. That is one question.
The second issue, in the work that I have done around developing standard recommended questions, the key that always gets overlooked is training. These data are only going to be as good as how well we train folks to collect this information, particularly self-report information. When we implemented our race/ethnicity guidelines, we spent an enormous amount of time developing training for the admitting clerks to respond to questions and training for WIC intake workers in a variety of different settings. I know it probably wasn’t in the scope of this initial report. I would love to see follow up work that focuses on requirements for training around data collection.
DR. STEAD: Both points are very good. This panel is the minimum set necessary to obtain, if you will, a balanced psychosocial vital sign. Several of these say screen and follow-up. If this very brief screen indicates a need – indicates that the person has a likely risk in an area, then additional questions come into play. In many cases, these additional questions are already represented in our matrix in the less ready for inclusion in all cells. In many cases, we have the information in the report of the additional questions you would use. There is no question about the fact that the initial panel is that brief set you need to obtain for everybody.
We recognize and put formally in – the clinical teams will need significant education around the implications of social and behavioral determinants and when you act. That has not been a central part of clinical training. That was called out as a requirement and noted in the table. It is an extremely valid point and concern.
DR. MAYS: Bruce started the race and ethnicity question. It was one I was going to comment on as well. I guess the question is for the committee, whether or not we comment on some of these things. I think from NCVHS’ standpoint, the race and ethnicity question needs not to be the standard way it is done, and then to say somebody will follow up. Whether or not it is something that we think is important to have in the maximum – well, not the maximum, but to have it in a detailed versus a short version.
I want to ask about two others. I didn’t understand it all the way through. One is you all chose education. The other thing you chose is intimate partner violence. In Education – we even said this in our hearing – the issue of knowing the educational attainment is not turning out to be very predictive in terms of health behaviors, et cetera. I can have a kid that says I have an eighth grade education and they went to some bourgie school and I have another kid – so I am assuming that if we take this out of the context of research and we put this in the context of usefulness in the healthcare environment, we really want to know something like primary language or your capacity, your proficiency with English or something like that. We are trying to use it as opposed to research.
The intimate partner violence – it seems very limited. One of the things that – Larry is still here. Good. I think it was the family medicine doc – somebody came up and said that we need to start assessing trauma. It is not just that aspect of it. It is really about trauma. Trauma, actually, is turning out to be more predictive, I think, than other things. I am just trying to – it is out and it is done. I am just trying to get a sense of why those two and not the bigger kind of picture things.
DR. STEAD: Very good question. Education – the association of education with health, forget behavior, is scientifically well established. The education was not there as a proxy for health literacy. It was there as a proxy for your ability to function in the world in ways that are associated with health outcomes. That is one. We also, you will discover, did not include health literacy because we felt that it was more effective for every health system to be taking a preventive approach, if you will, to the patient not understanding what they needed to do with things like playbacks, one part of which could be assessing health literacy.
Health literacy did not actually make this cut. Now, I can tell you as an individual, Vanderbilt is going to collect health literacy. The standard question of how to do that is in the report. It simply didn’t make the same grade for readiness and usefulness as other things.
Intimate partner violence is narrow. The committee was very careful about using established evidence. The recommendation of the U.S. Preventative Taskforce is narrow to intimate partner violence. It is clear that it is important in other age groups. The questions is – it, therefore, did not rise to the same bar as we cut ourselves down to this parsimonious set. My co-chair is an expert on education.
DR. MAYS: I know. I think it is a research question not a clinical question the way it is done.
DR. GREEN: I would like to back out of the details for a minute and make two comments about this. This report is a harbinger of what is to come that grows from two things. I can illustrate the first. The flood, the overwhelming flood of measures, metrics, and data. There are 26 states out of the 50 in the United States that by themselves, require clinicians to report more than 11,000 measures. There is a scream from the delivery system for relief from data reporting. It is a major source of waste. These data are collected and reported. EHRs are constantly being turned on and off and required to do stuff without achieving desired results, without the data getting used.
Just imagine every community in the United States being filled with hospitals and practices of various sorts and individual personal health records stored there. You could extract just that list of measures. That is a game changer. It is a major game changer. There is the cry for parsimony. There is more coming. There is another IOM report that is over this same target. It should go under review Friday. It may never come out of review.
This is tough territory. It is really important, I think, for the committee to recognize that the work of NCVHS can actually align with the thrust of this report and more reports to come. Data, data everywhere and not a drop to drink. Data, data, everywhere and we have 13 ways to measure this, but we can’t really collect all of them on anybody, therefore, they are not useable. That is where we are. We know this.
The other critical contextual thing, I think – Bill, you might push back on this. This work and its connection to NCVHS work is actually floating on the notion that you cannot improve what you cannot measure. It is a fundamental breakthrough in human thought. The converse is not true. Just because someone can measure something, doesn’t mean you can improve it.
We are looking for a harmonization of a parsimonious set of measures that the delivery system, the public health community, and local communities and individuals can unite around to stop the march of the United States to being dead last in so many important population health measures. This is really important. It is not so important to argue, in my opinion, about the details. We need to see everyone make some sort of compromise, make some sort of adaptation that says I can live with that one. I will do my part.
This has been really hard territory. I want to applaud Bill and his co-chair. These types of reports basically alienate everyone. They fail to completely please anyone. If you need a shield, I am looking for new work.
DR. SCANLON: Well said. That is a good point to end the discussion on. Thanks very much Bill. We want to turn now to Susan Queen. Susan is going to update us on some OMB guidance that directs agencies to share the administrative data, another NCVHS issue that we have been pressing for quite a while now, to make available the administrative data that they gather for their programs, claims data, for example. All federal agencies have received this directive. Susan will tell you how we are implementing it within HHS.
DR. QUEEN: Let me know how quickly I have to talk. I know we are a little over time here.
The OMB published in February of 2014 a memo and a guidance that was sent to all the heads of all executive departments and agencies on providing and using administrative data for statistical purposes. The memo was to focus on the strategic use of data within the federal government as an asset to enhance statistical uses. This is not for open data or for public access, but within the federal agencies to share, themselves.
It focused on data that cannot be made publicly available due to statutory/regulatory policy protections. The focus also was to use these administrative data that can’t be shared publicly as high quality and reliable sources of information that could be used to enhance statistical data. The notion is, and it is not a new one, that you could take the administrative records that you are already collecting for programmatic and administrative purposes, link, combine, match, supplement the survey data to make the end product more valuable.
Part of the guidance was to help agencies, particularly program agencies manage their data more effectively so that from the outset, you would be thinking of a statistical – potential statistical purpose for your data, which would then permit you to set up your Privacy Act system of record notices with appropriate routine uses and disclosures.
The memo build on previous and referenced these previous OMB memorandum, which did create a little bit of confusion amongst HHS and all of the other entities that were trying to respond to the memo. You see they have the Open Data Policy Memo. That was one of the memos this was building on. The point of this particular memo and guidance was not to be making data available to the public, but to use some of the activities that were being conducted under the Open Data Policy, such as developing inventories and enterprise architecture for your data systems and becoming aware of all the data that agencies produce for their own purposes, that you could use that in responding to this memo.
There was a lot of work and clarifying with OMB as we were responding to the memo, the fact that this would not be information that would then be released to the public, but that it would retain its protections, retain the privacy and confidentiality requirements, but could be shared for statistical use within the federal government.
I have always been told you never use the word that you are defining within your definition, but OMB described administrative data as administrative, regulatory, et cetera. It included information collected by contractors on behalf of the government or grantees. The data are usually carried out for purposes of administrating or operating a program. You are talking about enrollment information, application information, payment information, service information. It relates to – it can be individual respondents. It can be businesses, facilities, et cetera, institutions.
What was important to clarify for everyone trying to respond across the government to the memo was the statistical purpose because it was a very specific definition. They used language that was under CIPSEA, the Confidential Information Protection and Statistical Efficiency Act, to define a statistical purpose and use. I think the critical key is it couldn’t be that second bullet. Anything that related to determinations or decisions regarding benefits, privileges, et cetera, is not a statistical use. It includes the usual types of activities that you would associate with a statistical agency, but not anything that would be a determination.
For the purposes of the memo, there were three key agency programs/units that fell in the scope of this. Statistical agencies, these are under the executive branch and they are covered by CIPSEA. Within the Department of Health and Human Service – that determination, by the way, as to whether or not you are covered is something that is made by OMB by the chief statistician. In HHS, the only statistical agencies are the National Center for Health Statistics and the Center for Behavioral Health Statistics and Quality in SAMHSA. The second bullet, statistical components that are units conducting statistical activities within a program agency – that would be if you look at MEPS in the Agency for Healthcare Research and Quality. They are conducting statistical activities. In the Administration for Children, Youth, and Families, there is the Office of Planning, Research, and Evaluation that does statistical activities. It is not covered under CIPSEA, but it is doing statistical work. The third type of agency is the program agencies. These are units within the federal government that are either administering or helping to administer a program that does make determinations about rights, benefits, and privileges.
Within the memo, itself, OMB gave examples of the type of statistical activities for which the administrative data would be used – building sampling frames. NCHS has to spend a lot of money developing the lists for the HIS. They can’t get it from Census, et cetera, so they have to pay to compile their own list to draw the sample. That would be a use. Inputting and editing missing information. Comparing outcomes for program participants. The definition of statistical activity, while they didn’t clearly delineate this, I think, within the memo, it is in CIPSEA. It does include other activities that you might typically think of as being research or evaluation. The key is the focus had to be on aggregate or groups. It can’t be information on individuals.
The policy addressed four key areas: fostering collaboration across program agencies, recommended data stewardship policies and practices – this is a critical aspect of managing administrative data for statistical purposes. They want agencies to be formulating and implementing these data stewardship practices. Within HHS, since we do have quite a long history of sharing administrative data for statistical purposes, a lot of this was in place already. Particularly for some program agencies, this was a new way of considering how to use their data.
It outlined considerations to assess data quality. This was another critical component. The recommended – the memo contained a recommended quality element that needed to be assessed by the statistical agency when obtaining the data before use. The Federal Committee on Statistical Methodology had done some work in the past I think with the Bureau of Labor Statistics and had something like a checklist of what key quality elements would be important to consider when reviewing the administrative data systems.
The last bullet contained within this guidance, OMB provided some sample model agreements like an interagency agreement with all the critical elements that would designate all the roles across the agencies in sharing the data, disclosure, use, transmission of data, security, confidentiality, protections of the data. The guidance also referenced things like the Fair Information Practices, Privacy Act, the exceptions to the Privacy Act requirements that might be applicable for statistical uses, and then recommendations if that didn’t work to amend system of record notices going forward. There was a lot of practical information contained within the guidance on how agencies could manage this data.
When the guidance came out, OMB had been working primarily with the statistical agencies across the government. However, the memo went to the heads of all agencies and departments. ASPE was asked to coordinate the response for the Department, working in close collaboration with NCHS. The first thing we did was to use the vehicle of the Data Council as the departmental body that would coordinate and convene all of the responses in responding to the memo. The members of the Data Council were briefed on the memo. We had a series of conference calls and meetings.
I think, for the statistical agencies, this is not a new thing. This is something they have done. For the program agencies, this was new unless you were working with somebody in a particular program who was accustomed to being requested for their data. There were concerns within HHS and other federal agencies on who is going to be asking me for my data. That was sort of the first. Who is going to be asking me for my data? It is protected by statute so I can’t share it. Those were some of the typical things that you think of.
OMB asked everyone to think outside the box so that if you thought that NCHS could benefit by asking for the Numident file from Social Security, go forth and make the request. They really wanted to encourage careful review of the statutes, regulations, and policies, potential revisions to those policies, statutes, and regulations to permit statistical use. I think what was key, also, is for agencies not to review their statutes in the traditional way. If a statute said you could share the data, you would share it. If it was silent, you wouldn’t share it because you thought you couldn’t share it. There was encouragement on a different interpretation of how you read the statutes.
One area for HHS – it was easier, I think, for some places to respond to this memo because we have had such a long history. NCHS has its record linkage program, the use of the National Surveys being linked with Medicare data, Medicaid data, Social Security data. The use of the National Death Index. There is quite a tradition of being able to share within the department and also obtain from other federal agencies data that could be linked. Census has also a long history of doing this. There has been a lot of collaboration within HHS between CMS, NCHS, et cetera, but also across agencies, extending to Census. I noticed that on the NCHS website that the National Death Index actually clearly stated that it was available to investigators solely for statistical purposes. It is a clear – they have a clear, a long history of doing this and it is well documented.
Another example was the Medicaid Undercount Project. I don’t know how many of you are familiar with it, but it was a multistage, multiyear effort between Census Bureau, ASPE, AHRQ, NCHS, CMS – I don’t have to recount this. I can look at the slide – and SHADAC. It is known as the SNACC project. It was specifically – this was a large effort to look at concerns related to the Medicaid undercount. We rely on our surveys to provide information as a source of – for key policy information on groups that we may not otherwise have information about, such as the uninsured, people who are eligible for Medicaid, but they are not enrolled in Medicaid. In looking at the administrative counts, you see more people that are being counted in Medicaid and yet the surveys were showing significantly lower counts. There was an undercount of Medicaid population by some of the survey estimates.
This took a lot of cross-agency collaboration and used data from Medicaid, National Health Interview Survey, the MEPS survey, the household component of the MEPS survey, the current population survey from the Census Bureau, their annual social and economic supplement, as well as the Medicaid Statistical Information System to just explore why this would be the case, how we could potentially either impute or resolve some of the issues related to the undercount. That was a long-standing effort.
Within ASPE, someone in my office, Joan Turek, has done a lot of work working with the Census Bureau and IRS data, detailed earnings records, to explore how imputation techniques may impact the results in comparing them to the actual administrative data, recognizing that administrative data also have their own issues. You have to ensure that it is up to date, people who have been removed from the rolls are no longer getting X benefit. There has been a lot of work between HHS and the Census Bureau and IRS.
I think HHS was well positioned in responding to this memo. This is just the start. There was a reporting requirement that was due this past summer on – all agencies were asked to provide OMB with information on either a new process that was established or an existing process that was adapted in order to start doing this on a more routine basis. Agencies were asked to identify a number of datasets that they would potentially fund of high statistical value.
When OMB received these reports, they were reviewing them and asking questions and getting feedback from the agencies. They did say that there are potentially going to be more reports down the road, but they haven’t implemented any sort of routine timeline. They did make a request for considering this to be included in budget planning. They will be getting back to the agencies to let us know what the timeline is going to be. OMB plans to use the information that was provided this summer to inform them on the uses of the guidance and what potential revisions they may need to make in providing technical assistance.
That is it. Was that fast enough?
MR. SCANLON: Questions for Susan?
DR. COHEN: A couple things. I am trying to understand the impact of this in very practical terms. Does this mean another federal agency can request all the vital statistics for NCHS?
DR. QUEEN: There is a difference when it comes to states. This is for the federal agencies. Vital records have a slightly different twist. I would say one of the requests that was made is Census is planning on making a request for some data from HHS. Please forgive me that I don’t remember what it was. Maya would have a lot to say about it. It is one that is very, very protected. It is from the Administration for Children, Youth, and Families. It would be for a statistical purpose.
Yes, we are expecting to be receiving and making requests for data that traditionally you would not think of being made available. It is not for the public, but it is, again, within the federal government.
MR. SCANLON: But it is the statistical agency that would be requesting. NCHS or MEPS or SAMHSA in this case or Census would be requesting the administrative data. It would be adjudicated. There is a lot of linkage that goes on already. I think many of you know that. In fact, this committee did a nice workshop and analysis of the various linkages that occur and what would make it easier to do – I think a couple years ago.
In this case, Bruce, this is not open data generally. This is a statistical purposes. You are starting with an NCHS or a Census or a BEA. The request would come from them about, okay, we would like to link our survey data to claims data, which we have done to death data to income data to things like that, nutrition assistance data. It would come like that. The goal here is it starts as the statistics agency requesting. Once the data are linked, then they are available for analysis. Now, most of the time, it would be in a restricted data access – once the data are linked, it is awfully hard to —
DR. QUEEN: And Bruce, it wouldn’t necessarily be the case that it would be – that you would be providing the data physically. The linkage may very well be done by either the statistical entity, itself – so the data don’t necessarily have to move. The administrative data and the survey data could be – there are a variety of mechanisms for how the linkage you are matching would actually occur.
DR. COHEN: And aggregate results would be provided rather than the detailed files?
DR. QUEEN: You would be using it for a statistical purpose so it would be aggregate or group results.
DR. COHEN: That was one issue. My other hat, as you well know, is the states sell data, vital statistics data, to a variety of federal agencies. I want to make sure that this isn’t an effort to get around the agreements, for instance, the states have with the SSA to provide them with —
DR. QUEEN: No. No. Absolutely not. We haven’t received a request for the vital records.
DR. COHEN: It will happen, I am sure. I am concerned – and expanding the use of NDI? NDI clearly was created for the intent for research rather than any statistical administrative uses. That would be a huge change if —
DR. QUEEN: Except as it is on the current NCHS website, it has to be used for a statistical purpose.
DR. COHEN: But one could argue there are administrative statistical purposes as opposed to research administrative purposes. That is a huge potential loophole for access to NDI data.
DR. QUEEN: I think the safeguard is going back to CIPSEA then. Exactly how is a statistical activity or a statistical purpose or use defined under CIPSEA and restricting it to that.
DR. COHEN: Has this been shared and processed with NAPHSIS at all, making them aware? It would be nice for the states to know at some level through NCHS, I think, about this. Even though it is an internal federal process, I think it is – in terms of transparency of changes in data sharing – data sharing amongst states through vitals is a huge issue. I think it would be valuable just to provide the states and NAPHSIS with a summary or do a webinar.
DR. QUEEN: I can verify this later because the person I was working with is Jennifer Parker at NCHS, the vital records are different.
MR. SCANLON: They are special.
DR. QUEEN: I hate to say this – they are different.
MR. SCANLON: They are not easy.
DR. QUEEN: There is absolutely no intent of trying to run around the existing protections and statutes that safeguard state data. There is a different mechanism there.
DR. COHEN: Thanks.
DR. NICHOLS: I guess I am really asking about where the puck is going because I was struck at the instruction that came to think outside the box. Maybe that was a cover letter. I am going to think outside the box like I usually do and say so what about CMS and the thing that I think may be most relevant to kind of what we have been talking about for a while is all-payer claims databases getting access to Medicare. There has been a real firewall set up that only three states have partial access for some things. It is beyond my personal ability to comprehend why three got partial access and the other eight have nothing. It is a deep frustration on the part of U.S. states because a fair chunk of spending is in Medicare and a lot of practice patterns you care about are in Medicare. You want to see Medicare versus non-Medicare. Is OMB —
DR. QUEEN: This is only to executive branches.
DR. NICHOLS: I know this, but is there a movement afoot to try to loosen some of this?
DR. QUEEN: What is interesting is within our office, we are – we have one of those CMS Virtual Research Data Center seats. CMS has permitted some use of their Medicare data for some projects working with states.
DR. NICHOLS: For specific projects, but not for making it part of a normal – that is kind of what the states want.
DR. QUEEN: That is definitely a CMS question.
MR. SCANLON: I think that the creation of the new office at CMS that Nile Brennan is heading, the idea there is to put under one roof a lot of the data dissemination programs. As you know, they have designated various programs in the states or not-for-profits can apply to be designated as – they call it a designated entity. They can get the Medicare data. Their role has to be one of analytics and some association with healthcare, itself.
I know what you are talking about. A number of states – I have heard about this before – have asked for access to all of the Medicare claims data. It has not been that easy. I think discussions continue to go on. We have, actually – remember this. We and ASPE, using some of the PCOR money that we have access to, we actually invested in feasibility and testing of an all-payer claims database. We were – everybody thought it would be very easy to do. Everybody wants to share their data. It was really very difficult, both architecturally, in terms of how do you set up such a system and people being willing to really participate. We were willing to – we actually asked CMS if they would be willing to take our – when we finished with the pilot study, to take what we had done into their data center. I think for legal purposes, they just were not interested in doing it.
I hear you exactly. I think the pressure is continuing to make that happen.
DR. NICHOLS: I just think in the spirit of the notion of community data, there is very little that can be more actionable in the short run than for counties to be able to say this is what I am spending on birth versus everybody else in the world. Birth, I guess, you don’t need Medicare data for, but you get my point. There is a lot of stuff that would be incredibly enhanced. I just think, Jim, we have kind of been stopped at the door of CMS saying oh, no, no, no. We have to get beyond that. This committee could perhaps play a role in pushing —
MR. SCANLON: My own view is someday we will get systematic data from EHRs on a wide-scale basis, standardized. You can tell a lot about healthcare and the health in a community from good administrative data, from the claims data. It took a long time to get that available. It is pretty much what we know about healthcare now from the claims data, expenditures and variations, what does the healthcare system do —
DR. NICHOLS: Yes, we know that it is highly variable, but communities don’t know how they stack up versus others without having their own —
MR. SCANLON: Yes. That would be the next step. You either have CMS do it, which I don’t think they want to do particularly –
DR. NICHOLS: I am pretty sure they don’t. It may be that we need to push. That is all I am saying.
MR. SCANLON: I think the all-payer –
DR. NICHOLS: I don’t think they can to be fair to them.
MR. SCANLON: They have a difficult job. Every time they take a step forward, they get – there is a whole other side of the world that differs from us in terms of what they do with the data. I think the idea of an all-payer/multi-payer claims database that includes Medicare and all the commercial and others as well, it has really been the Holy Grail, at least for administrative data. We are still pursuing it. It is just amazing how difficult it turned out to be. We will keep pushing and see what we can do.
Some vendors are trying to do this for their areas. I am not sure that will ever solve the problem. They have even linked some clinical data to their claims data. There have been vendors for analytics around a long time. That is probably not the solution in the long run. We hear you. A lot of folks are pushing in that direction.
DR. FRANCIS: This is a completely different question. You mentioned stewardship. I wondered whether there was anything for us to learn from what the stewardship recommendations were or was it just a general be good stewards that they might learn from us.
DR. QUEEN: The second half of the memo was a couple of appendices. One was related to the Privacy Act. One was sort of a model agreement, whether that would be an interagency agreement, data use agreement, et cetera. It is half of the guidance, itself, with the elements and what they would look like for good data stewardship.
DR. FRANCIS: What I am actually interested in is how much does it look like or not look like or should we look at it vis a vis our toolkit.
MR. SCANLON: They do have a model data use agreement that is probably overkill for a community. We could certainly share it with everyone. I think our agreement –
DR. QUEEN: I think I actually heard a request for staff analysis to give us a recommendation on how close it is or not. Did I mishear you Leslie?
DR. FRANCIS: That certainly could be possible. We could just have a look at it and see whether we need to.
MR. SCANLON: I think the first thing is to see what it looks like. Again, I think it is probably more than is needed for a community, but it is sort of a standard data use agreement. We did manage to get – OMB was trying to do a – when they started a one-size fits all, sort of assuming nobody was doing anything right so they had to tell us. We really pushed back and we gave them language that some agencies have been doing this longer than OMB has existed. We would like to be able to use what we have already perfected. To their credit, they took our language.
MR. SCANLON: Thank you very much, Susan. We have reached the stage now where we could take a short break and then come back. Why don’t we do that? Come back at 11. Then we will discuss the subcommittee and overall work plans at a very high level and see what follow up we need to do to keep moving.
DR. SUAREZ: I put together a table summarizing all of the work plans that might be helpful.
MR. SCANLON: That would be great. We are having some nice themes emerge in terms of convergence and data availability. There is a fair amount of interaction among the workgroups. We can look at that as well.
DR. MAYS: Is it possible to do the workgroup work plan before we do the big one? Can we just flip it?
MR. SCANLON: Yes.
MR. SCANLON: We are going to turn it over to Vicki to bring us up to date on the workgroup on data access and use. We are still having some trouble with our microphones so stand by.
DR. MAYS: Thank you for letting the workgroup move up a bit in time so that we can see if our agenda also fits in with the larger priorities. As you know, the workgroup this year is proposing to do something a little different to try to be a little nimble, to try to be as responsive as possible to HHS’ work with the various agencies. I am going to let Damon actually describe kind of the space that will be helpful to HHS, what the expectations are, and then I will talk a bit about how it is that we are going to get there, which is number one and two. We are actually going to move up the issue of the agencies and what the agenda would be.
MR. DAVIS: Thank you, Vicki. For those who don’t know me, I am Damon Davis. I work in the CTO’s office or the IDEA Lab, focused on the open data agenda represented by healthdata.gov. I think one of the things that is going to be important, in terms of what we consider here is the fact that the agency, itself, is so focused on data writ large. We heard the administrative data piece a moment ago. It is only a component of the overarching data agenda.
With that said and with the backdrop of things like a chief data officer at CMS and folks like that at NIH, you are starting to see a much more significant, more robust set of data-related activities across the agency. Therefore, many of the operating divisions, staff divisions, and others are thinking more critically about their IT infrastructures, their data collection and curation mechanisms, and then the ways in which they are making data available, trying to advance beyond beautiful graphs and things along those lines that are part of an online report to actually making the data components behind the report usable, accessible, et cetera.
With that said, one of the opportunities that I think the workgroup has is to be contributory towards some of the data systems that are going to be either developed or advanced. There are existing data systems that do need to be updated. A lot of times, the folks who are running those systems, running those programmatic activities actually do need some level of feedback mechanism. We are often challenged by the means by which we can get sort of feedback from the community, from the users, from the experts in a specific domain as to what is going to be valuable in the advancement of that particular system, the advancement of that dataset or what have you.
I think that there is a real opportunity for the workgroup to be very contributory towards some of the feedback loops that are going to be part of these systems development. If you think about the desire to get these opinions and these expert inputs, PRA can often be a roadblock in terms of asking too many people what their opinion is in order for it to be formulative for the development of these systems and their data collection activities. I think the workgroup really has an opportunity to be really participatory across many more agency data-related system development activities.
I am not sure if you want me to expand at all more beyond that. I think that really is going to be the crux of what I hope to rely upon the workgroup to do. We have – the Health Data Leads Organization is a set of individuals across HHS who the CTO’s office, the IDEA Lab relies upon to sort of find data, find the data sets, the systems, and look for the places where we don’t currently have open data. I think there is going to be an opportunity to allow the workgroup to gain greater access to some of the Health Data Leads meetings to understand a little bit about what is going on behind the scenes of the department’s data activities and also introduce the Leads to the fact that the workgroup can be a feedback loop, a sounding board through which a lot of the opinions can be pushed back toward the organization doing the development.
I think I will stop there and just sort of leave it with a highlight on the fact that we need to have a little bit more cross-collaboration between our organizations.
DR. MAYS: The charge to our group really is to deal with data access and usability. Part of what we have tried to do in the workgroup is to start framing how to best do that. One of the things we are trying to do is to really determine what agencies at this point are in a state of readiness, looking for this kind of consultation. This is the type of thing that Damon is talking about because they are coming to him. We are going to plan – in 2015, one of the things we will do, in terms of agencies, is begin to plan a schedule out of who we are going to meet with.
Prior to doing that, there are two things that we have talked about as our work plan going forward. One is instead of kind of doing this on an ad hoc basis – if every time an agency comes in the only thing we do is respond, but that as we continue to learn about what the needs are within HHS that we start to develop some kind of guiding principles. I was thinking about using the term frameworks. I used it a couple times. I have pulled back a little bit. We are between principles and frameworks.
I don’t know quite where it is going to be, but I think it is probably better for us to talk about some guiding principles that HHS, in general, can think about when it is trying to figure out what is the best way to make sure that this data is accessible and useable. We are doing that within the context of kind of three levels of meaningful use cases. We are talking about the ability to be able to access the data if you are somebody who is a data entrepreneur. You have resources. You are really very focused on putting it out there for a very specific reason. You are a data warehouse. Our second group that we are talking about is really those people who actually have skills – researchers, departments of public health, other areas where people kind of have a skill level to be able to use the data and then we want to make sure, kind of following up on some of the things that we hear the Secretary’s interested in, is also talking at the level of the consumer.
Now, it is not in our wildest imagination that every agency and every dataset is going to have equal use and usability, but I think helping agencies to improve at each of those levels would be one of our goals – to make sure that the information is getting out to individuals who might see value in trying to push the data further, making sure that at the level of the consumer, there is some voice out there that can try and make sure that they understand the information. It may be tweeting it. Twitterthon – that is what I was trying to think of. It may be that is what is employed at the level of the consumers is that you ask these agencies that are developing this data to also do that.
Those are the kinds of approaches that we are going to take in terms of trying to develop, first of all, just a set of guiding principles as we go along that, in general, can be there and that – I don’t see it as static. It would be something that would be refined.
The second bit of work is if we are going to call agencies in, we really do need to decide what we are going to do. What is the set of questions? We want to begin to standardize that. What is the set of questions we want to ask that particular agency about their data and how it is used? The second thing that we want to do is to have some sense of what our responsibility is going to be. Are we going to follow up with something written? Are we just to give them off the cuff in the room? We have to determine how to be the most useful that we can be within a timeframe that is useful.
The other thing is we differ a little bit from the subcommittee. We should be a lot more nimble. We should be able to have the capacity to do quicker feedback. We just want to standardize that process. That is what we are planning, in terms of 2015.
The third piece that is there is to hear from the subcommittees ways in which you think that we can be helpful, in terms of the agendas that you have. That was part of why I wanted to see if we could go before we did the whole plan. It was also to say tell us ways in which we can be helpful. On our agenda later, we will actually talk with privacy about their launch and give suggestions. If there are other very specific things, it would help us to put those into our work plan so we kind of know what we are doing. While we are going to move nimbly, we also have to kind of have a plan going forward so we know what the workload is.
Damon, anything else? Okay. Let’s stop here and see if there are comments, questions, advice.
DR. NICHOLS: Vicki, I was thinking how our agendas could overlap a little. I was just going to ask how do you prioritize which agency is called before the inquisition first? I vote for CMS, but you heard that.
DR. MAYS: I think you are going to get a joint response here. First of all, it is an invitation. Second of all, I think that Damon could talk a little bit more about the state of readiness. He has a sense of who has already completed their activities and who hasn’t.
MR. DAVIS: That is certainly going to be a qualifying factor for what will sort of be the invitation is where someone is in their cycle of development. These are austere times. There is not funding for every single program to do something new and creative with their data assets and tools. Not everybody is going to be in a position to be receptive to feedback because they are not going to be in a position to actually do anything with it versus those that are significantly further down the line in terms of wanting some level of feedback. I think it is going to be keeping a finger on the pulse of where any one agency is in terms of their desires.
It is also going to be a little bit of seeing the need and inviting them to understand a little bit more about what the opportunity is. There may be agencies that are not necessarily at all thinking about the challenges of access to their data and its usability. The opinion might actually spark some level of – it might catalyze something within the agency that wasn’t already in the plans. There is going to be a little bit of push and pull. It will be those who are ready, interested, and available, clamoring to come and get some opinions. It will be those that are not necessarily even thinking about this where we are shining a light on the fact that this is a really valuable thing that you can probably spend some resources on. We would love to offer you some opinions as to how to make this better.
DR. COHEN: Item two here, data dissemination assessment, I think there is perfect alignment between what the Population Health Subcommittee is trying to do in understanding data provider literacy with assessing the dissemination strategies of the HHS agencies. I see the work that the Data Workgroup will do in this area that I am going to be involved with as actually a precursor feeding into the data literacy workshop roundtable that we are going to have later in the year. I think we should make sure that both groups are in sync around this activity.
DR. MAYS: I think there what we are looking to do is develop some kind of assessment tool that we would give to the agency.
DR. COHEN: Here is my draft right now.
DR. MAYS: Oh, okay. If it turns out that that actually is useful, in terms of something – again, it is kind of there as part of maybe the guiding principles of these are a set of questions you should ask yourself. I think that would be great.
MS. KLOSS: For number one, given how much data has already been made available on data.gov, my assumption is that you have some guiding principles that you are starting from. Is that true?
MR. DAVIS: Absolutely.
MS. KLOSS: You are not starting from scratch on that. You are looking to refine or embellish or make them more publicly useful.
MR. DAVIS: And make them, obviously, specific to health domains and social services domains. When you think of data.gov and healthdata.gov, there is a set of common core metadata, for example, that the agencies are required to put out. You could potentially see how, say, healthcare research data would have a specific set of core data elements that would be relevant to only healthcare research kinds of things versus a space agencies research agenda. I think it would be appropriate for the department to sort of refine some of the common data elements, some of the usability requirements to make them specific to healthcare so we can have an infrastructure of data sharing linkages and things along those lines that is going to allow for more efficiency. I think we need to help the agencies to understand sort of what is at the forefront of data’s usability so they can catch up with a lot of what industry, private sector entrepreneurial endeavors are already doing so that our data can be usable and valuable.
MS. KLOSS: Do you see those guiding principles being arrayed by the three kind of markets you described?
MR. DAVIS: I don’t mean to speak for Vicki at all, but I could certainly see how that would be appropriate. One of the things that I think we need to be better about is an understanding of our audiences for our various datasets. For example, on healthdata.gov, I think we are challenged to point to any one audience. Obviously, entrepreneurs are coming there. That is one of the things that we very much focus on with activities like Health Datapalooza and other challenge-related activities. I think it is a little bit more difficult to sort of look at the platform as it currently stands and point to specific audiences that we are attracting. I think that is indicative of something that we could probably do better across the department and all of the operating divisions.
DR. MAYS: I think, for example, one of the things that will get recommended will be an issue of metrics and knowing who is using it and how difficult it is for them to use it, something as simple as who is staying and who isn’t staying. In the background, when we started talking to the different agencies, it would be like if we did, say, user case three and it is the consumer. It may be that putting into place a series of data collection that would allow you to know something about the difficulties may allow you also then to increase their usability by addressing some of those issues.
MS. KLOSS: And that is built into number two.
DR. MAYS: Yes.
DR. SUAREZ: My question was actually going to be about just the last point that you brought up, Vicki. I think the data dissemination assessment tools is great, but there is probably one of the more critical elements of this, which is laid out in the goals of the guiding principles three-fold. The very first one is make easier the use of the data.
My comment, my suggestion, which again goes to what you were saying, is the need to develop metrics for measuring usability, but not so much from the perspective of the provider of data, in other words, the federal agencies or agencies, but somehow from the perspective of the user of the data to see how usable the data and unuseful the data have been. I think developing those metrics and a mechanism to assess that from the user perspective would be helpful.
DR. MAYS: Thank you. We are good.
MR. SCANLON: Thanks Vicki. Thanks Damon. We are going to look across all of the work plans at a fairly high level. We don’t have to go into detail because we are now meeting as a full committee in most of the three areas as well. If you could focus on sort of the higher level or any major directions – I know there were a couple areas where we have to sort of get back to the workgroup with some feedback from HHS about roles or other things like that, just so we can do that before our next meeting. I am going to start with Standards, Walter, if you are okay.
DR. SUAREZ: I might need to say a few words about the table that you all have. We emailed a table that I just created out of the three work plans that we discussed yesterday and today. I wanted to just say very high level, briefly, that this is not complete. I tried to capture the essence of the main points being made by the subcommittee. I might have missed a few things, particularly in the population health side since I was just typing as they were presenting it. I tried to capture the core elements.
The other thing is the alignment – I didn’t change any of what was being described at all in terms of moving a hearing here or moving a letter there. It was just basically following what was discussed and presented. What you see in the tables are for each of the quarters and for each of the subcommittees, of course, what are the topics that during that quarter, the subcommittee will work on, topics like in discussion topics during conference calls and prep work and all that, activities like hearings, workshops, other things, and then products, outputs, reports, letters, things like that. Those are the three things that I tried to capture for each quarter, for each of the subcommittees. Again, I didn’t have the workgroup on Data Access and Use, which can be added as another role as well.
That is the caveat I guess. The other thing, of course, is I didn’t stretch this to 2016. This was just purely a 2015 gathering. I know there could be something that could spill into 2016 or even beyond. Those are the elements. I am sure there are going to be adjustments and corrections.
From the Standards Subcommittee point of view, I think we tried to highlight basically the main activities. This time around we are looking at doing a hearing in February on operating rules and ICD10 maintenance process. We won’t have a letter or report to produce between now and February to present to the February full committee for approval or action or anything like that.
In quarter two, we are going to be holding our – at this point, as we evolve, as you see during the first six months, we are going to be preparing for the review committee and developing those metrics and all those things. If that works, we would then hold a hearing in June of the review committee. In June, as well, or May, we will present a letter from the February hearing to the National Committee.
In September, we are going to focus – between after the June meeting, in the third quarter, we are going to prepare drafts from the review committee recommendations and plan for possible activities in the November area like November hearings or workshops on different topics. That is what we looked at focusing on in quarter three.
Quarter four, we are reserving time for a possible hearing or workshop on these topics of transparency or population health management or public health standards. That is basically it for the Standards Subcommittee.
MR. SCANLON: Are there questions for Walter?
DR. NICHOLS: Thank you, Walter. I just wondered – I didn’t really see the overlap with the ACA evaluation. Where does that work?
DR. SUAREZ: It is called the ACA Review Committee. You might be talking about more of the larger National Committee, not Standards Subcommittee. The Standards Subcommittee had the responsibility of the Review Committee. That is the ACA Review Committee for Standard Transactions.
There was a larger point that was made yesterday about NCVHS looking into the evaluation of the Affordable Care Act, itself, not the standards, not the transactions, but the larger Affordable Care Act health reform. That is a National Committee topic that is not included in here. That is something I did not include. That would be something that we would be needing to consider adding, but as a full committee activity. Maybe there is a separate role that is called full committee, actually.
DR. NICHOLS: I was just going to ask. The one issue that I could see you could put either place would be this issue of claims adjudication and standardization, which has been – when a provider group hears the words administrative simplification, in addition to the extremely important work on transactions and attachments and all that that makes actually claims adjudication and standardization possible, prerequisites for doing this, they really want the claims adjudication standardization because that is where they spend all their money.
MR. SCANLON: And that was one of the goals.
DR. NICHOLS: That is what I thought, but I have not seen anything in the implementation that would suggest – except for you and me. We remember these goals. That is why I am trying to – is that under your purview under standards?
DR. SUAREZ: You use very specific terms. Claims adjudication standardization process. That is very big. It is an interesting question. We have over 1,500 health plans – I mean insurance companies. Each of them has different policies and products. Each of them go through different adjudication processes. I am not sure if the question is the efficiency of the adjudication process to achieve payment or if you are talking about standardization of the adjudication, itself, which would make it look more like a single payer system in the United States. There is some sort of a —
DR. NICHOLS: Walter, let’s be careful about words. What I am –
DR. SUAREZ: I am just wondering about the –
DR. NICHOLS: Single payer is very different than claims adjudication. I am suggesting that when the IOM study the question of how to save money in the United States healthcare system, there were some guys from Harvard – I forget their names. I am sure somebody can remember them – that studied the theoretical savings that would occur if everybody used Medicare’s methods. He did it using their own data inside Health Partners or something.
What I am talking about is moving toward a standard for a claims adjudication algorithm that would indeed give providers serious relief. I don’t think that is single payer. I think that is efficiency in administration.
DR. SUAREZ: Let me make a point about that. Every payer system, the adjudication process follows a number of expectations and actions, including the policies and the practices and the benefit packages and everything else. A single, standardized adjudication process – that is why we need to be very careful with the words – could lead to – as you say it, if we all used the CMS adjudication algorithms that means that we are all going to be paying like CMS.
DR. NICHOLS: No.
DR. SUAREZ: That doesn’t —
DR.COHEN: Can we take this offline?
DR. NICHOLS: I am just trying to find where is the issue in the jurisdiction. Is it a Standards question or is it a broader question? I do seriously believe it is a serious question that our policy community is supposed to take up. A lot of us thought when the law passed and certainly the intent of the Senate Finance Committee, was for that to be part of the implementation of the simplification words that were in the ACA. Somehow this seems to have been, oh well, this is just too complicated. No, it is not.
DR. SUAREZ: Let me suggest the following. I don’t want to continue this debate. It is an important debate. It is an important topic. We will bring it up into the Standards work for February. We will come back to the committee with some discussion about it. I think it is an important topic.
MR. SCANLON: We will look into it in more detail.
DR. STEAD: I would like to just advocate in terms of process. A, we had add a row up on top for the full committee. We get topics we really want to resolve. Whether we need to have clearly on this plan when we are going to resolve, how we are going to measure the effect and effectiveness of a transaction as part of the review committee, that needs to be in Q1, either as full committee or Standards. Standards certainly needs to drive it.
Second, I would advocate that in Q3, under full committee, we enter in a draft of a three to five year planning horizon plan for the full committee so that we work at that there. Then we put in each of our things for Q4, the 2016 work plan. In Q3, we could look at the overarching horizon. In Q4, we could draft our work plan. We would flip this thing around and we would be working from our planning horizon instead of reacting to the plans that are coming up from the committee. I think that might help us.
Under Population, move our ACT Platform Alignment Roundtable to Q4. We are not going to do that in Q3.
DR. SUAREZ: What was that again?
DR. STEAD: The Platform Alignment Roundtable will be Q4. It won’t be in the July to September. It will be in the Octobery sort of window.
DR. SUAREZ: I only presented the Standards subcommittee. Are we talking now about Population?
DR. STEAD: I am just talking about the framework.
DR. SUAREZ: You are correctly already things in the Population one. I just want to be systematic about how we go about discussing.
DR. STEAD: Okay. I am back to a macro topic. I would move everything out of Q4 that we haven’t decided to do, out into 2016 and beyond, and only leave in Q4 the things we know we are going to do. We can always pull things back in. I believe we need to have an Executive Committee call before the holidays that closes this thing down so we go into January with a work plan. We have 20 minutes to sort of agree on how we are going to get to that. That is the reason I am being sort of – talking process. Thank you.
MS. KLOSS: I was going to suggest as I look at the Privacy that, even if we are ready to go forward with the hearing on HIPAA privacy and security state of play in Q3, I doubt we would synthesize everything we had learned and be able to generate a letter or report in Q4. I was going to suggest that gets moved out.
When I step back and look at the whole, we essentially have two hearings or roundtables/meetings each quarter. I would just ask the question of whether we have the staff capability to do that. Have we ever done two per quarter?
MR. SCANLON: That would be hard to do. Realistically speaking, we are not going to have more staff.
MS. KLOSS: I think that before we have that Executive call, we need some assessment on what our staff capability is.
MR. SCANLON: I think that would be really hard to do. I want to be realistic.
DR. SUAREZ: Can I ask one quick question? Linda, speaking on Privacy, are there any other changes that you see besides moving the letter from Q4 to Q1 2016?
MS. KLOSS: I think from what I know today, I am okay with the rest of it. We need to have a Privacy Subcommittee call before Christmas to see how far we can get in framing this February 1179 hearing. If we need to do more learning and understanding the issue more thoroughly as a subcommittee, then I think I would be really comfortable just pushing that out. I would just look at it as working based on what we know now.
DR. COHEN: Walter, I think you have done a great job. Thank you so much for putting it all on one piece of paper for us to consider. It helps me a lot to see how things integrate. It is a process we need to go through.
I suggest that the co-chairs of each of the committees go back and either talk with their committees or decide amongst themselves to sort of tweak the rows for their particular committees given what they know. I agree that we really need an Executive Committee call to finalize the whole schema and to decide whether we can support hearings and meetings on a quarter by quarter basis.
I would also suggest that we all – that everyone on the committee comes up with some general ideas that work for the entire committee, broadening our perspective. I think Len’s idea of thinking about the whole process of unifying or standardizing claims adjudication, the idea of what is our role and evaluating the ultimate impact of ACA – there are a bunch of big ideas. This is our work plan, but we shouldn’t lose sight of the big ideas to help drive us if not in 2015, down the road.
I also think as we become more visible in our activities, we are getting asked to work with more and more groups. I would like to see us in the row that focuses on big ideas for the full committee to really do more proactive engagement with ONC, with IOM, with the Health Data Council, with a variety of other FACAs to help support our activities and really capitalize on the streams that converge across the department. Our homework for everybody is to think about these big ideas for the committee. I think the homework for the chairs of the subcommittees is to just refine the 2015 plan.
MR. SCANLON: That is a great idea.
DR. FRANCIS: This is a little bit of a swan song comment, but there is obviously going to need to be some prioritization between or among the subcommittees. If there are not going to be eight hearings, somebody is going to need to give. I would simply suggest that as you all go forth into the future thinking about this, you continue to use as the touchstone, the kinds of criteria that this committee has used in the past to decide which things really matter for this committee to do.
Some of those criteria include, of course, what are our legal responsibilities, but they also include what is it that somebody else isn’t doing and what might we have a particular role to play in our capacity as advising the Secretary. As I look at some of these, they are much more within either our legal or our advisory scope than others. Some things that are here on this list look more like, well, this would be neat information for somebody to have, but don’t look so much like potentially actionable recommendations for the Secretary or something that is actually – that we could make a difference with respect to our advisory role.
That is my parting shot as you look at these. I am happy to tell you where I think – the one I am the most – well, why don’t I just say the one I am the most worried about is community data literacy from this perspective.
MS. GOSS: I really support a lot of the input that I have heard from the last couple folks here. I think that the chart that Walter put together for us really becomes the beginning of our own strategic plan, but would think that the Executive Committee should take this and that from a staff leadership perspective, something like this should start to be maintained to help us have appropriate scheduling at meetings. We have a lot of things to cover. To Leslie’s point, we need to be a lot more honed in what we are taking on and how we are commenting on things within our own deliberations tied back to that vision of where is our sweet spot and what we can really influence and accomplish with the fact that we only have so many hours of the day and we only have so much staff support. I give a nod to getting ourselves a little bit more process-oriented and focused in on really where we need to be spending our precious energies.
MR. SCANLON: Good point. Anything else? In terms of the committees, the full committee agenda setting, in terms of issues, when we plan – when we start planning the February agenda – do we have a question on the line?
DR. CHANDERRAJ: I just wanted to compliment and thanks Larry Green who chaired our committee for so long, as well as Leslie Francis, who contributed a valuable amount. I just wanted to thank them both. I am sorry I could not attend.
MR. SCANLON: We are going to have a salute to Leslie and Larry in a minute. That is very nice.
I think in terms of looking at ACA or health system change and impact and so on, maybe what – think about this – what we could schedule at the February meeting is sort of our own overview from HHS about how we are approaching this. Again, it is a convergence. It is a convergence approach. Our role here is the nexus is the data. If not policy, necessarily, it is the data part of it. We actually have some different streams in terms of evaluation and monitoring that I think would be a good place to start. I think since it includes all of the different streams of data, it would be helpful for the committee. We have metrics. We have domains that we are monitoring. We are looking at modifying some of those since we are now entering the second enrollment period here. That might be a good plenary discussion for the full committee maybe in February. It would be a place to start.
In general, I think when we have big ideas, we would like to start with the full committee, starting it as a plenary discussion with some fact finding. Some place to start is probably a good idea. We can start planning that. ASPE can take – we can help deal with that.
Before we wrap up, again, I want to personally thank Larry and Leslie on behalf of HHS and the full committee for eight wonderful years of service. You have hit right up against the record, in terms of service. It has been a pleasure having you. I hope you can still affiliate with us, at least for the Data Workgroup. I think I reflect the ideas of all of the members and certainly HHS that it has been a pleasure having you and we appreciate your leadership. We will have some formal recognition from the Secretary a little later. We were just taken by a little surprise by the FACA audits.
At any rate, we want to thank you again. I wonder if we could all give —
MR. SCANLON: We are at the stage of the meeting where we can invite any members of the public in the room or on the phone to speak. Steve Lazarus —
MR. LAZARUS: I wanted to follow upon something Bill mentioned a few minutes ago about the metrics for the review committee and their work. Not only would CAQH like to see those metrics discussed at the February public meetings of the committee, but also to have them published in draft form before that hearing so that the public can comment to the committee on our feelings about what those – are those metrics valid or not or whether there are improvement to be made or things that have not been thought of. Thank you.
MR. SCANLON: Leslie, you have the floor.
DR. FRANCIS: I wanted to take the opportunity before I turn the microphone over to Larry to just say thank you to all of you. This has been an amazing experience. I have learned an extraordinary amount and I hope made good friends that I will stay in touch with. Thank you all for the privilege.
DR. GREEN: Indulge me for a minute. I really like this public thing.
DR. SCANLON: He gets to say whatever he wants.
DR. GREEN: These sort of moment stimulate reflection. What has happened to me over the last week since I learned that this was going to be my last occasion here, I did a little looking back. I just want to tell you some things that come to mind in looking back over the last eight years.
One is being on the committee is more like taking a graduate course than being on a committee. One suggestion I have is that the Secretary think about granting members a degree. I don’t know if HHS has a degree grant, but it is just – when I went back and looked at some notes, it is stunning what I did not know eight years ago that I know now because of the privilege of being part of this committee.
I want to assure you the pleasure was not always mine. Learning and growing as human being often comes with pain. That is a special feature of being on NCVHS that comes to mind.
Another is I am just so keenly aware of all the shoulders we are standing on. For me, I can highlight this with just the chairs that I worked with on the committee, Simon Cohen, Harry Reynolds, Justine Carr. Without exception, they taught me and I learned from them as individuals. It is true for the members of the committee. I was just saying to Leslie – Leslie and Maya and then joined by Linda later on, pretty much 95 percent of what I understand about privacy, security, and confidentiality, they taught me. I am very, very appreciative.
We are also standing – I don’t want to carry the standing on shoulders metaphor too far, but there are products of this committee – I think it may have been Linda last night over dinner that said they have shelf life. I really think it is important to not lose track of that 63 year trajectory. There is stuff there that just keeps going. I think that is an example of leading by the wake. When you start hearing Leslie’s voice about you got to decide what you are really going to do in 2015, you are not going to do that 2015 work plan, in case you don’t know this. It is – it defies the second law of thermodynamics and other things. Now, the Standards Committee may do everything on their list because Walter will be there.
The vision work – that 2001 vision of the health statistics still stands. I think that continues to be a great piece of guidance for this group. I think the Stewardship Report is still standing out there as the unreplaced strategy for what it is really going to require to use data to improve health and healthcare. It is going to require stewardship framework. That is a really good one. This CHI Report and the Toolkit that is just coming out, these are useable tools for the American people that are worth having and worth knowing about. I feel really good about that when I look back, having been part of that tradition and that sequence. It is a real privilege.
I want to tell you some things I have learned. One is you are an impossible committee to manage. It just doesn’t – this committee defies everything that is taught about managing a committee. You are absolutely a bunch of irascible folks, which is probably why the committee works.
Another thing is I have really solidified realizing that health is a community affair. It is quite clear it was. I think the work of the committee lately has confirmed that it still is. I predict that it always will be. Staying close to this idea of helping communities learn how to learn, learn how to be healthy and create healthy environments, work with clinicians in those communities to help them learn how to work together, that is a sacred cause that this committee would fit in Leslie’s criteria for something that it does not appear to me anyone else is really taking up.
Another thing I have learned is how much obsolescence fills the airwaves and the space, our universities, our institutions, our insurance companies, and our governments at a state and federal level. We have spent better than a century breaking everything we can up into as many part as possible, commoditizing it if possible, owning part of it and making a living off of it. Now, we are at a point where we have discovered that we are going to spend this century putting it back together again. In my view, data probably are the glue that are going to put things back together again. I don’t understand why we want to spend a whole lot of time continuing to sculpt and improve obsolescence.
Basically, everything has changed about data and the data world for the information age except the way we think about it. We still think about it the way we thought about it in the 1800’s. You guys have the mental reach and capacity to break through that. You just really have to do that.
I want to tell you something I am really worried about. I think the greatest risk the entire enterprise that we are about here is not earning and maintaining the public’s trust that their data will be used properly. That has not yet been earned. If it is not earned, what we are talking about is not going to happen. If it is lost, it will be our kids and grandkids that have the next opportunity to restore it. I think that is just hugely, hugely important for you guys to watch.
I want to tell you what I would pursue if I were arriving as a new member. If I were coming in now, knowing what I know now, which I didn’t when I arrived, the first thing I would do is intensely own board your new members. Don’t just let them show up for the first meeting. This is much too complicated. There is too much to do. You need to do it efficiently. I think – one of the most important calls to you remaining folks who will still be here, even though there is going to be a big turnover in 2015, is you got to own the board people. Pay attention to that please.
I would just do everything I could think of to sustain the nascent linkages that the committee has with ONC and try to build them toward CMS. I would try to build them toward the philanthropies that are trying to help communities become learning systems for health. The committee can get some partnerships there that could really help. I hope you will pay attention to that.
That is it. I want to say good bye, happily so. I wouldn’t trade it for anything. I think I am considering after last night adding to my curriculum vitae under honors that I was Marjorie Greenberg’s eighth chair. I really appreciated the kind words that some of you have spoken individually and that were spoken last night over dinner. It has been a real, real privilege. I think I may erratically audit the course by lurking on the public phone lines without your knowing that I am there. If the spirit moves me, I may send some of you an email just to be annoying.
MS. JACKSON: My final comment is that this is the meeting that almost didn’t happen. It wouldn’t have happened had it not been for all of you and all of staff and members who are on the phone and called in. The last two weeks have been like nothing in my life, but this is quite a testament that we are here. We are going forward with thoughts and hearing the words and the sounds and the voices of Larry and Leslie inside of us as we go through 2015. Thank you all.
MR. SCANLON: It is a good time to adjourn.
(Whereupon, the plenary session adjourned.)