Blog Home


HEALTH IT: Fewer Than 1 In 10 Doctors E-Prescribe

October 12th, 2006

While about one-quarter of physicians were using an electronic health record (EHR) as of 2005, fewer than one in ten physicians were using EHRs with functionalities such as electronic prescribing, researchers say in a Health Affairs article published yesterday [2-week free access] online and reported in today’s Washington Post.

The data on hospitals’ use of information technology are more limited, but best estimates suggest that 5-10 percent of hospitals had electronic prescribing — or computerized physician order entry (CPOE) — systems in 2005. However, whether hospitals had stand-alone CPOE systems or comprehensive EHR systems with a CPOE component is unknown, the researchers say.

The article by Ashish Jha, assistant professor of public health at Harvard University, and coauthors is based on the first report of the Health Information Technology Adoption Initiative, a landmark public-private partnership between the federal government’s Office of the National Coordinator for Health Information Technology, the Robert Wood Johnson Foundation, and several academic research institutions. The Bush administration has set forth a goal of widespread EHR use by 2014, and the research by Jha’s team is meant to create a “reliable baseline” against which progress toward that goal can be measured.

“Health information technology in general and EHRs in particular could increase quality and reduce the cost of health care,” said Jha. “However, despite the potential importance of EHRs, there is a surprising lack of consensus about just how prevalent — or, more accurately, just how rare — they are in the current health care system.”

Jha’s team found a lack of data on the adoption and use of EHRs among safety-net providers. What evidence there is suggests that safety-net hospitals use less health IT than other hospitals. Given the potential of EHRs to improve quality, “ensuring access to these tools among all providers is critical to reducing disparities in health care,” Jha and coauthors state.

Email This Post Email This Post Print This Post Print This Post

 to the #1 source of health policy research.

No Trackbacks for “HEALTH IT: Fewer Than 1 In 10 Doctors E-Prescribe”

16 Responses to “HEALTH IT: Fewer Than 1 In 10 Doctors E-Prescribe”

  1. ashishkjha Says:

    Matt — to answer your October 30 question — my understanding is that it will happen if Congress approves an HHS budget. To do a high quality national survey with detailed questions, large sample sizes to allow for subgroup analyses, etc., we’re likely talking about something that ONC will need/want to fund. Other folks in our group know more about what exactly will happen assuming congress passes a budget but my best understanding is yes, it will. Hope that helps.

  2. Matthew Holt Says:

    Ashish–So this survey will happen should it get funded? Presumably it shouldn’t cost too much (low 6 figures is my guess) Can RWJ just fund it alone without needing Congress to act?

    Perhaps I’m being too hopeful as annecarrol says. But I’m asking for the easy part — counting and measuring what’s happening. All the stuff anne’s complaining about is the hard part of increasing adoption and interoperability in a marketplace with no incentives to do either, in the absence of a governement insisting on it or paying for it!

  3. ashishkjha Says:

    Matt Holt — indeed you are right when you say “ONCHIT idea was to set up a constant, regular survey….and became the gold standard.” The first annual report (and the HA paper that we’re discussing) was to get a baseline from existing data.

    I believe the plan going forward is (if Congress approves a budget for HHS) is for ONCHIT to fund high quality detailed surveys that will become the gold standard of where we are with EHR use. At least that’s the plan as I understand it.

    Hope that is helpful.

  4. annecarroll Says:

    To Matthew Holt: It seems that you are looking for too much from the Feds, at least under the current administration. Looking at the mission statement of the ONCHIT ( you will see that its main responsibility is to “develop” a “strategy” for interoperability among systems that collect health and medical data at the Federal level. That’s why the new Coordinator is from the VA, which has the most advanced EMR/EHR system in the country, widely accepted and used by its physicians. A minor ONCHIT responsibility, “to the extent permitted by law,” is to consult on a strategic plan for national interoperability in “public and private health care sectors”. Under Section 3 of the Executive Order creating the office, “Responsibilities of the Coordinator”, is the interesting clause that, to do this, the Coordinator shall “(v) Not assume or rely upon additional Federal resources or spending to accomplish adoption of interoperable health information technology;” ( In other words, they can’t spend Federal money to research what works or strategies to incent providers to adopt their recommendations, or to “force” anyone (presumably public OR private) to adopt their “strategic plan”, “standards for interoperability,” or any other deliverables that they are mandated to produce. In the Federal government, this means “no regulations.” Thus, even if they “develop” and “recommend” interoperability standards at the Federal level, presumably each agency can still design and implement its own HIT. What “ONCHIT funded piece” were you referring to?

    Elsewhere on the web site, it’s clear that the mission also includes an emphasis on interoperability with State health programs which are also funded by the Federal government.

    The most that the previous National Coordinator accomplished was to award three contracts to technology vendors to “develop” standards for interoperability. Given the history of the information technology industry, that is assured to be an exercise in futility, conflicts of interest, and trade secrets. Given the way the Office is hamstrung by the lack of funding and its constricted and inconsistent mission, it’s clear why Dr. Brailer quit. (They are also quite experienced in setting up conferences and meetings, talk-a-thons, and providing “keynote speakers” to a multitude of “professional” conferences. Well, they have to spend their budget somehow. “Your tax money at work”…)

    So, it appears that state and local governments and the health care industry and academics are on their own with regard to researching and developing HIT that works and that enhances the quality and safety of the health care delivery system. It’s clear that the application of HIT will be piecemeal, non-standardized, inefficient, and ineffective, with no clear public health advantages, although there may be evidence of quality, safety, and financial improvements anecdotally from individual institutions and providers.

    In the “glass half full” category, I suppose the data privacy community can be pleased that the effective integration of personal health care data is a long way off. I myself have mixed feelings about improved HIT vs. the civic value and responsibility of maintaining the security and privacy of personal information.

  5. Matthew Holt Says:

    Ashish Jha–thanks for your reply. Can you tell me (us) what you’re going to do next? Given that the current set of surveys are such a mish-mash and in general don’t get much below the “do you use an EMR question” I thought that the ONCHIT idea was to set up a constant, regular survey that captured detailed information on the evolution of all clinical computing use, and became the gold standard. If you’re not going to do that, what are you going to do?

  6. annecarroll Says:

    1. Re: concerns about integrating various sources and types of data, such as clinical labs and pathology departments, into EMR’s: the National Library of Medicine has done all the hard work of developing data standards and a taxonomy for medical terminology, and makes it available for free:( Standards enable data to be shared and integrated, while allowing user-defined content and formats. Use of such standards within organizations would also simplify the EMR format and limit physician “verbosity.” This isn’t rocket science; many industries that depend on data from multiple sources to run their business have created such standards for data shareability.

    2. Diffusion of innovation theory has shown that the characteristics of the object/technology to be adopted are just as important to successful adoption as the characteristics of potential adopters. I have seen a few small, anecdotal studies of limited applicability that actually analyze the characteristics of EMR and POE systems themselves as well as the characteristics of the target user population. It’s up to the users and subject-matter experts, not the vendors, to define what is required for this to be viable technology, and to ensure that the “right things,” not the “wrong things”, are measured. As a previous poster has implied, to reap the maximum benefit from information technology and to ensure that it is not disruptive but creates value, it is necessary not merely to automate current dysfunctional health care delivery processes, but to optimize processes in order to deliver the highest quality and safest care to patients.

    3. In the literature, one barrier to adoption is the difficulty for physicians who have admitting privileges at multiple facilities to sort out the differing requirements of the various systems they may be forced to use, posing a risk for patient safety. Therefore, a standardized EMR record with mutually agreed-upon definitions must be implemented; if the commercial vendors will not develop such a system, a non-profit or government entity should take on this role. This goal could be attained long before 2014; again, it’s not rocket science.

  7. ashishkjha Says:

    Two quick comments:

    To Matt Holt — I appreciated your “so what” comment. I was the lead author on this “rehashing” and actually — let me add a couple of points about it. If you look at all the surveys out there about EHR adoption, you see rates of adoption that vary from 4% to 65%. Such a range is clearly not useful. What we tried to do was apply some objective criteria (not developed by us, per se, but by external experts) to identify only high quality surveys and use their results to create national estimates. Not a meta-analysis and I agree, to a very careful reader of the literature, nothing is shocking here. But, sometimes, hacking through the brush to get at a single number can be useful.

    If you ask why doctors are so slow to adopt — I think the comments here are really right on. These systems are expensive, disruptive, and it is not clear that they are a good deal for the physician, at least not in the short run (though Robert Miller at UCSF has done some of the best work demonstrating a real value for EHR use over a longer time period).

    Finally — one quick thought for Richard Levitt — EHRs will not affect disparities in as much as they have to do with access. But, we also know that even accounting for access, blacks and whites, for instance, often receive very different care. I am not quick to blame doctors — but I am quick to say that our healthcare system should not allow this to happen so pervasively and in such dramatic fashion. EHRs can radically reduce variation in care and hopefully, make a real difference in reducing disparities — but this is an empirical question not yet answered.

    My first time blogging — this is fantastic.

  8. LFBaltrucki Says:

    Re: “Getting to the nitty-gritty of how clinicians use computing in their workflow”

    The weak data that has been generated up until now does not necessarily mean that the EMR doesn’t have the potential to help us transform health care delivery.

    If you really want to learn about the impact that the EMR can have, just talk to the people who have been using it. Ask them how it has tranformed their practice. Ask them, as experienced users, about the changes that they would like to see for improving the technology. After over 10 years of working with the VA’s EMR no one inside or outside the organization has asked me anything about my experience.

    Re: My experience

    Right now, no one understands how to integrate the Clinical Laboratory and Pathology Department into the integrated delivery system. The unfortunate reality is that there is no mechanism for “front-line workers” in “vertically-oriented organizations” to share their experiences and perspectives, and further that there are no incentives for this kind of innovation;
    (like next-generation EMRs with flexible user-defined content and formats and improved, clinically-relevant functionality), the kind of innovation that we need to “take things to the next level”.

    Not only do I know how to integrate the clinical laboratory into the integrated system, (using the EMR) I can also explain how it will save the VA millions of dollars and improve the quality of care at the same time. Such is the nature of the EMR’s potential when it is placed in the appropriate “context” for health care delivery.

    Despite the fact that most of the data that informs us about making health care decisions is derived from clinical laboratory testing, no one has made a serious attempt to “re-engineer” the laboratory for the 21st century health care delivery system. The major reason for this is our inability to “think (or work) outside the box”.

    While IT provides us with the opportunity for transforming health care, at the same time it also challenges us to think in new ways about health care delivery and to be open to the possibility that we may need to make some fundamental changes in what we do and the way in which we do it. The EMR is disruptive technology. Simply using information technology to support current processes and to maintain existing professional boundairies/activities will probably not give us the ROI that we are seeking (and that we deserve).

    Other issues brought up by contributors are also very important.

    Re: Productivity and Cultural Transformation

    Just like anything else, there is a learning curve, and of course “off the shelf” technology will never be as useful as that with “user-defined” features.

    The 7 page ER entry with the relevant information “hidden” inside the note speaks to the need for training existing (and new) clinical staff members, (prior to implementation or entering on staff).

    This also points to the need for some degree of standardization in terms of the content and format of notes in the electronic record, (perhaps with oversight from a medical records committee and IT-staffers).

    This certainly addresses the need for a “shift” in organizational culture where individual providers come to the realization that they are not the “center of the universe” and that the time for merely writing their notes for themselves, colleagues in their own department, or for coding purposes, has come and gone. There are several clinical settings in which communication, is equally important to documentation, as an essential function of the EMR.

  9. David Grant Says:

    From observation many in this field already know that very few physicains adopt EMRs, but having emperial evidence that support this observation is key. However, I believe there are better emperial methods for benchmarking adoption which I will address below. Before I go on it is important to note that EMR EHR CPOE. They are different concepts but often produce or capture similar data.

    The question we must ask is why is adoption so low. I believe the answer is three fold. First, EMR applications are too combersome and effect physicain productivity to be practical. The average EMR take 12 minutes to fill out with equates to adding $75 to each exam vist and dropping overall physician productivity by 10% – 20%. Second, dispite the promise of reduced errors EMRs have never been emperically show to reduce errors. Third, is with the execption of a few coding examples (not an EMR), EMR have never produced a positive ROI.

    That said there is potential value in EMR? Yes, if you could make them fast, connected and used by the physician – not their staff (a key point that is often overlooked), that error and cost could be avoided.

    However, there are aspects of EMRs that do gain adoption. Contrary to the many failed attempt sto launch electronic prescribing (note the only proven ROI for payers and providers to date) a few companies have gained significant adoption and maintianed that for many years. However, the early failures have scared investors who are reluctant to try again.

    My point of view; however is bias having done this successfully for over 6 years with better then 4000% adoption, thousands of physicians and millions for prescriptions per year; but consider the emperical data. Using electronical prescribing as an example, we know the theoretical average of the perscription per physican per day volume based on past history and speciality. Forget, contracted physicians, enrolled physicians, etc. They do not reflect adoption because they do not indicate usage. If we know the historical average and we track volumes that approach this limit, minus controlled substances which can not be electronically prescribed, then we can compute adoption rates based on the percentage of electronic prescription in relation to the historical average.

    Let me illustrate this another way, in 2002 the largest electronic prescribing company was routing only 2 MM prescriptions per year with a claimed 60,000 physicains under contract. Meanwhile a small company with a fraction of the capital routed over 2 MM prescriptions in 6 months with only 2,500 adopting physicians.

    Therefore, focus on transaction per physicians by physicians as a means to determine adoption. You will notice that the public companies purposely avoid answering this question. In addition, if a physicain delegates data entry to his/her staff, the benefit of many of these systems is lost because they lost the ability to shift behavior and are not only a expensive data collection tool.

    Now for the bias…we can demonstrate EMPERICALLY 85% adoption for over 6 years but have never been asked once to participate in these studies. Interesting…

  10. Matthew Holt Says:

    I’m not sure that this study isn’t a waste of money. It basically confirms a bunch of studies that were done over the years and tells us what we already knew. I was under the impression that the ONCHIT funded piece was going to both define, measure and track the adoption of EMR in a systematic way, doing original research.

    This paper seems to be a rehashing (I believe you academics call this meta-analysis) of every study ever done, with little new to say. As it turns out, the data is in agreement with the latest CDC stucy. So it’s a big “so what”.

    There are though plenty of interesting aspects that we should be looking at–really getting into teh nitty-gritty of how clinicians use computing in their workflow and their practice. No one has done a good job analyzing that and I speak as the author of two major surveys of physicians computing at Harris Interactive in 1999 & 2000. Is that type of detailed work upcoming or is my (and the RWJ’s slighly more singificant contribution of) cash being wasted?

  11. Peter Basch Says:

    To John Haughton’s excellent comments re EMR-generated verbosity – this is a huge problem, and one that will make meaningful health information exchange more difficult. Because we have framed much of the quality problem around illegibility, and the E/M coding system is based on documentation volume – EMR vendors have responded with systems that too easily create voluminous documentation (and satisfy coding requirements of any level), but actually convey no useful information. As this problem was caused by a market response to coding requirements, it will only be fixed if the EMR agenda includes discarding the outdated E/M coding requirements – and recasting documentation as “brief and structured is usually better” and supporting longitudinal, care, and not just care episodes.

  12. Peter Basch Says:

    >>Correction to yesterday’s posting – I inadvertently sent an early and incomplete draft.

  13. John Haughton Says:

    Why does nearly every provider have the ability to submit electronic claims and not many are using EMRs / EHRs? — We see stories about lost productivity – initial or ongoing; expensive systems with vendors getting bought, going out of business or changing course…
    In short – at this point, from an economic standpoint, the risk-reward balance does not point in the direction of reward – and additionally, the emrs / ehrs and registries themselves seem to be getting cheaper and more flexible, further rewarding those who wait…
    Perhaps sometime soon, technology and interoperability will nail the productivity issue – early and / or ongoing (note – clearly the specialist who documents focused procedures can and perhaps always will, get productive faster than the Primary care physician seeing patients with multiple co-morbidities.
    PS – another issue: synthesizing “information” and “knowledge” out of the longer, more formed notes coming out of EMRs. At a gathering of about 60 heads of IPAs / PHOs last weekend, the story of a 7 page ER note for a simple laceration surfaced. It took the reviewing physician more than a minute or two to get to the action of what was done and when to take out the stitches.

  14. Peter Basch Says:

    This is an important advance in research, as the definition of EMR / EHR is unclear and evolving. However, even measuring functionally advanced EMR use at 10% does not mean that 10% of physicians are using EMRs to move healthcare towards the IOM\’s goals of STEEEP care. That figure (reflecting healthcare transformation) is closer to 0%, except in integrated healthcare systems, where providers are incented to deliver care that goes beyond typical office visits (where care is mostly reactive and episodic).

    Rather, as Miller alluded to (in his article in last September\’s Health Affairs), and Sidorov inferred (in last month\’s Health Affairs), the evidence that EMRs of any type are actually being used to improve healthcare is weak to nonexistent. EMRs (even those with advanced functionality) are being used primarily to enhance coding opportunities — something that is unfortunately necessary to create an ROI for physician EMR purchasers, bur irrelevant to improving outcomes.

    Accuracy and clarity in measurement are important, but lose relevance when one is measuring the wrong thing. If the desired \’endgame\’ is healthcare transformation, then what should be measured are elements that reflect those changed processes, rather than simply looking for infrastructure (EMR adoption). However, it would be useful to also attempt to determine the role and cost of EMRs as enablers of those changes. In particular, the questions that should be asked are:

    • Are EMRs neccesry to achieve healthcare transformation; are there instances where providers have delivered transformed servivces without HIT or without EMRs?

    • What is the average cost of HIT needed to deliver transformed healthcare?

    • What is the average decrease in FFS income, when providers spend less of their time on traditional office visits, and more of their time on care coordination, prevention, and chronic care management.

    The answers to these questions will not just make for good research. They should help to guide policy makers and health plans towards appropriate reimbursement reform; and establish the basis for providers to see advanced EMRs as good investments, rather than onerous mandates. And the end result of such changes will hopefully be to dramatically increase both the rate adoption of advanced EMRs and their use for improving and transforming care.

  15. Nainil Chheda Says:

    There is an interesting link at explaining \”Facts and Figure of the EMR Industry\” and it\’s Adoptation. You may visit: .

    The various standards like CCR, CDA would help harmonization of various EHR\’s thus advancing applicability of EHR\’s in the healthcare domain.

  16. richard levitt Says:

    EHR will never reduce disparities in healthcare. The dispsarities have to do with access. Access is inhibited by lack of funds. Those without insurance seek no care, or ER care after the fact. Even the disparities from plan to plan for the inusured show a big difference in denial of medications and services/procedures. There are easily two classes of medicine in the US, the private sector and public sector. Introduction of EHR will do nothing to remedy this situation.

Leave a Reply

Comment moderation is in use. Please do not submit your comment twice -- it will appear shortly.

Authors: Click here to submit a post.