Actual Oral Testimony May 26, 2004   by Tom Doerr, M.D.

Thanks for the opportunity to present my views.  My name is Tom Doerr.  I am a physician with a part-time Internal Medicine practice that is limited to Medicare beneficiaries.  I am also one of the two physicians who founded Wellinx, which is an electronic prescribing company.

I read the March 31 meeting notes and the April 14 eHealth Initiative report.  My goal here is to offer oral testimony that is complementary to these sources.  I will briefly mention and underscore important issues that these sources addressed, but I really want to address two issues that these sources have underemphasized.

Near the very end of the March 31 NHCVS meeting there was discussion about the Medicare Modernization Act’s directive to provide the patient’s medical history with prescriptions.  Most of my prepared oral testimony will address the benefits of writing prescriptions associated with diagnosis codes.  The other topic I’d like to address is the barriers to creating real-time interfaces between Practice Management System interfaces and electronic prescribing software.

The bottom line is my recommendation that, at a minimum, prescriptions should be accompanied by a list of the patient’s diagnosis codes.  Optimally, in addition to the electronic problem list of diagnosis codes, each prescription would be accompanied by a diagnosis code that indicates the condition being treated by that prescription.

The value that can be created when physicians write electronic prescriptions and associate them with diagnosis codes is enormous.  This value deserves emphasis and consideration.  With diagnosis-driven work flow, the safety and the cost-effectiveness of prescribing can be substantially improved in four ways:

1) We can automate checking for drug-disease interactions and omitted drugs at several points in the prescription process.  Ideally this will happen at the point that prescribing decisions are made.  That way, prescribing decisions can be reconsidered and changed most efficiently.

2) We can communication: PBMs, pharmacists, and payers will be able to better understand the patient’s health conditions.  This can facilitate understanding and improve the appropriateness of communication among stakeholders.  Some unnecessary phone calls can be prevented when various interested parties understand the context in which prescribing decisions are made.

3) We can lower treatment costs: Intelligent messaging about high-quality, cost-effective pharmacotherapy can be delivered at the time of prescribing.  The Medicare Modernization Act clearly calls for electronic prescribing standards that accommodate messages about lower cost drug therapeutic alternatives.  It will be more effective if messages are delivered to doctors that are specific to the diagnosis being treated.   To be specific, its optimal to give physicians objective, comparative information about the efficacy of treatment alternatives and about the tradeoffs between cost, efficacy and adverse effects for that patient’s particular diagnosis.

We physicians need something like Consumer Reports for Doctors integrated into our work flow.  The content must be transparently evidence-based.  The most logical way to do this is by promoting a diagnosis-driven work flow.  And the most effective way to promote diagnosis-driven work flow is with a requirement that diagnosis be associated with prescriptions.

My company implemented diagnosis-driven electronic prescribing at a 70-physician group in Missouri several years ago. We measured a significant improvement in prescribing behavior in that group.  So, we studied the impact of this system in a controlled trial that included 38 physicians practicing at an Integrated Delivery Network in Wisconsin.  Compared to the control group, the costs for new prescriptions and their refills decreased by 7.5% in the intervention group during the first six months that they used the system.  During the next six months of use, the costs to the payer were reduced by 16% compared to controls.   Almost identical improvements were observed after the system was implemented in the control group in that IDN.  Overall, the payer in this study measured their return on investment to be around six to one in the first year.

These results were subsequently independently confirmed in another group of 100 physicians sponsored by a payer in Maine.

These studies are limited by their small size.  And they focused on commercial populations because the sponsoring payer was not CMS.  Nonetheless, there is a remarkable consistency of results.  When we consider all the prescribers using this system with its diagnosis-driven evidence-based work flow and decision support  for all patients, 65% of all prescriptions are written for generic medications.

We have three different manuscripts describing various aspects of our system that are currently pending publication.  The salient results of the use of this system are summarized in Appendix of my written testimony.

Requiring diagnosis codes will not impose an undue burden on physicians.  We  doctors are routinely associating diagnosis codes with our orders for blood tests, radiology tests and with our bills for professional services.  Many Electronic Health Record companies already have diagnosis codes associated with their prescriptions.  Three vendors on the market already generate prescriptions associated with ICD-9 diagnosis codes in wireless handheld computers[1].

Less than two months after the Medicare Modernization Act was signed into law, we learned that the OMB estimates of the cost were about a third higher than those estimated by the CBO.  We have got to figure out how to reel in the costs without compromising quality.  Nationwide adoption of diagnosis-associated prescriptions enable the promotion of fiscally-responsible, high-quality care.

The interfaces needed between electronic prescribing systems with Practice Management Systems will probably be the greatest hurdle to widespread adoption of electronic prescribing.  As far as I can tell, this matter has not been discussed very much before this committee, and it was only briefly mentioned in the April 14 eHealth Initiative report.

If you don’t have a real-time interface between the scheduling system and the ePrescribing system, the office staff has to do double entry of patient data before being able to write any prescriptions.  That is usually too time-consuming to be practical.  My written testimony describes some alternatives to real-time interfaces that are sometimes used, but these are not optimal alternatives because of the work flow problems they cause.

The current standards for these real-time interfaces are inadequate.  Even when the Practice Management System and the electronic prescribing software are both compatible with HL7 version 2.4, it takes one to two person-days of labor or even a week for each of the two vendors to complete the interface.  Version 3 of HL7, when available, will reduce the labor required, but it won’t address the other aspect of the problem—the exorbitant expenses that are often charged to clinicians who need the interfaces.

The total cost of these real-time interfaces typically run from $5,000 to $10,000 but can be as much as $30,000.  The higher figures, between $10,000 and $30,000, can arise due to anticompetitive business tactics, price-gouging, or other reasons that are detailed in my written testimony.

But even $5,000 to $10,000 is a huge barrier to adoption for electronic prescribing for the 1 to 2 doctor offices where about half of U.S. physicians practice.  So we have got a technological challenge and then some unseemly business practices that are major barriers to widespread adoption of electronic prescribing.

Both Dr. Teich’s testimony on March 31 and the April 14 eHealth Initiative included several additional standards-related barriers to the vision of seamless electronic prescribing and prescription fulfillment.  I would like to briefly reinforce his comments on three issues based on our experiences at Wellinx.

The first additional issue involves code sets for drug names.  The government has encouraged the existence of several competing code sets for drugs at a higher level of abstraction than NDC.  We saw this when they forced the divesture of Medi-Span by First DataBank (FDB). This encourages competition, but it also causes some barriers to interoperability, and thus to rapid adoption of e-prescribing. The NCPDP protocol currently does not resolve these interoperability issues.  This is perhaps the greatest existing barrier to smooth communications among stakeholders, specifically prescribing applications, pharmacy networks, pharmacy management systems, PBMs, and Practice Management Systems.  Ideally, there would be one well-defined hierarchy of numeric identifiers and associated text descriptions that was supported by all payers and commercial drug database vendors.

A second concern that I want to emphasize is the welter of state requirements for the details of printed, faxed, or electronic prescriptions.   Each state pharmacy board has its own approach to the same essential needs.  The result is a bewildering set of sporadically changing requirements that act as barriers to rapid implementations and adoption of electronic prescribing.  These need to be standardized at the national level.

Finally, the need for standardized tables for managing formulary status information is become increasingly apparent.   Insurance companies, PBMs, application vendors, and formulary data vendors do not share any widely accepted standard.  I am specifically referring to the definition and sharing of formulary lists and the special “policies” that go with them to define a formulary.  Setting standards for describing formularies will significantly reduce the costs of managing this irregular information.

In conclusion, the challenges here are formidable.  But pharmacotherapy is so well-positioned for improvement that the potential to create value is really exciting.

Evidence Based Medicine is often admired but seldom practiced.  In order to get Evidence Based Medicine out of academia and into patient’s bodies, we need to give doctors informational tools and incentives to embrace fiscally-responsible care.  We are at a crossroads with electronic prescribing standards.  We have an opportunity to create standards that include a framework for seamlessly weaving context-specific decision support into prescribers’ work flows.   That’s half the battle.  There are additional comments about incentives in my written testimony.

Thank you for considering my opinions.

[1] Allscripts, WebMD and Wellinx