Skip to content Skip to navigation

Stop Throwing Your Money Away on the Electronic Medical Record (or start creating value)

November 21, 2009
by James Feldbaum
| Reprints

It is official. According to a Study published today in the American Journal of Medicine there have been no documented financial efficiencies from computerizing medical records even when studying the “100 banner hospitals that are most wired.”

"The idea from this administration that we're going to pay for health reform out of savings from electronic medical records is baseless propaganda," according to David U. Himmelstein, MD., associate professor at Harvard Medical School and former director of clinical computing at Cambridge Hospital. "It may be politically attractive, but it's nonsense."

"Coding and other reimbursement-driven documentation might take precedence over efficiency and the encouragement of clinical parsimony," He explained that electronic medical records have the capability of allowing billers to scan patient histories for items that might result in justifiable reimbursement. "Hospital information systems help you do this, to find every co-morbidity that helps you jack up the charges,"

On Monday, The New York Times reported on a presentation by Ashish K. Jha and Catherine M. DesRoches of Massachusetts GeneralHospital. They compared 3,000 hospitals at various stages of adoption of computerized health records, and according to the article "found little difference in the cost and quality of care" between those that had adopted and those that hadn't.

I am not surprised. I have long complained and blogged that the EMR is missing the mark. I have a hard time convincing doctors, whose engrained workflow and thoughtflow will be completely disrupted, that the pain is worth the gain. The EMR is a repository of data. We need to tap into the power of the computer to mine out information and trends that otherwise might be too subtle to reach the attention of a provider in the normal course of rounds. Studies have clearly demonstrated that there are marked differences in cost and outcome between hospitals and that expensive care does not correlate to better care. The EMR needs to address that disparity and close the quality/cost gap between institutions. Maybe then we will see some real value. We need to put meaningful effort into Clinical Decision Support and help elevate the quality of care that we deliver. Only when we bring real evidence-supported information to physicians at the point of ordering will we modify performance, safety and outcome (and maybe even cost).



Interesting post. I'm absolutely sure that turning on an EMR assures nothing. Per John Glasser's quip "you dont buy ROI, you manage to it."

Purchasing and installing an EMR gets you a capability. Similar to buying a book. If you don't read it and apply what you read, then you just have a book.

I would be interested in your take on the Premier / CMS HQID work here:

Unless they completely fabricated the data (I dont think so), then the results (quality and cost improvements) are true of something. Several of the hospitals they featured are known to have very successful EMR implementations that contributed to the result.

As support for your argument, I would just mention my post "Obama Asks Bloggers For Help With Health Care System Overhaul." The gist is, using Dr David Eddy's Archimedes work (Kaiser, CDC, and employer clients), Eddy shows that closing the inferential gap (Evidence-based Medicine term) does not guarantee lower population costs over time. Cholesterol management in some patients absolutely does / Hypertension management not so much. This suggests that an EMR is a kin to a 2 edged sword that can cut for good or ill.

You started off with "financial efficiencies." I hope I didn't take you off track to a related but different topic.

Here's another take on why the AJM study may have turned out as it did ( These authors posit that healthcare exists in an ecosytem (not too different from Christensen, Grossman & Hwang's value chain) and that EHRs, by analogy, would be necessary but not sufficient to enable the expected improvements in cost and quality. Without other sufficient and/or necessary changes to other nodes in the ecosystem, the EHR by itself is largely effectless.

Excellent post Jim, and honestly, not too surprising - Joe's analogy to a book is apt. An EMR is just software, software is an enabler, it is nothing in and of itself. To look only at the EMR without also evaluating all of the organizational, behavioral and process changes that must simultaneously be effectively and correctly implemented in order for success, is to frame the problem so as to find failure.

Furthermore, the EMR is probably only one piece of the necessary software. You can buy, read, and understand an Introduction to French textbook, but you will hardly be fluent until you have bought, read, understood and practiced several more textbooks.

Jim you also hit the nail on the head with the emphasis on CDS. The role of CDS shouldn't be to replace the provider, but rather to reduce unnecessary and extraordinarily costly, variations in care. The huge store of clinical information necessary to elucidate and elimiate variation requires an EMR, but without CDS, CPOE and workflow-oriented interfaces, EMR is a cost sink rather than a value source.

Finally, there is also a necessary economy of scale and interactivity. For one hospital to implement is not the same as for a critical mass of hospitals to implement on an interoperable platform or platform(s). What's more, the functionality and interoperability must extend well beyond the hospital and into the clinics, offices, centers and ambulatory practices.

In the end, I agree, it is merely an assertion, and a vague one at that, that "HCIT will lower healthcare costs and improve quality". We can't know this to be true until HCIT is fully implemented. But based on everything we know and have seen from the adoption and implementation of IT in other industries, this assertion is well supported by existing evidence.

Thanks for the responses. I have to admit that the results of the two studies seem to be in radical conflict. I strongly agree with the book analogy. I would add that sometimes the pages are in no sensible order, the print can be too small and the book can frustratingly be missing pages.
Intuitively and experientially the EMR does have value. We have only begun to tap into its capability and we are embarrassingly struggling with implementing functionality that will add value. Clearly there has been too much emphasis on the non-clinical components which I would deem the low hanging fruit. Do we lack the programming skill or the implementation fortitude to attack the issues of CDS?
I would love for those institutions that have been so successful (yes, I am jealous and proud of them) to weigh in. Logically, evidence-based standardized care has to be a step up from the free-for-all that we have now.

Electronic Health Records Linked to Improved Quality in Primary Care Practices

Routine use of electronic health records may improve the quality of care provided in community-based primary care practices more than other common strategies intended to raise the quality of medical care, according to a new study by RAND Corporation researchers.

Studying 305 groups of primary care physicians in Massachusetts, researchers found that practices that used multifunctional electronic health records were more likely to deliver better care for diabetes and provide certain health screenings than those that did not.

While quality differences discovered in the study were modest in size, the study is one of the first to demonstrate a link between use of electronic health records in community-based medical practices and higher quality care. The findings are published in the Oct. 6 edition of the Annals of Internal Medicine.

"Overall, we were surprised by how few strategies to improve the quality of care were linked to measurably better performance," said Dr. Mark W. Friedberg, the study's lead author and an associate natural scientist at RAND, a nonprofit research organization. "The strategy that showed the most impact was use of advanced electronic health records."

Electronic health record systems were linked to higher quality care when the systems included advanced functions such as electronic reminders to physicians, and if the systems were used routinely by a medical practice.

Studies by RAND Health and other groups have documented problems with the quality of health care in the United States, including gaps in the delivery of preventive and chronic disease care.

To address these shortcomings, primary care physician practices are encouraged to invest in many types of structural changes intended to foster improved quality. These include giving physicians feedback on their performance, sending reminders to physicians and patients about needed services, having language interpreter services, offering appointments on evenings and weekends, and adopting electronic health records.

Researchers were able to examine whether such strategies could be linked to higher quality of care by making use of a unique set of information about physician practices collected by the Massachusetts Health Quality Partners.

Researchers surveyed 305 medical practices in 2007 to assess whether they had put into place any of 13 structural capabilities that are aimed at increasing the quality of medical care. That information was linked to the results reported by each practice for 13 commonly used measures of quality in four clinical areas—diabetes treatment, depression care, overuse of medical technology and common health screenings.

Primary care medical practices that used multifunctional electronic health records performed better on five of the quality measures—two involving diabetes care and screenings for breast cancer, colorectal cancer, and chlamydia.

Medical practice groups that had frequent meetings to discuss quality reported better results for three measures of diabetes care. Practices that reported high physician awareness of patient experience ratings reported higher performance on screenings for breast and cervical cancer. No other structural capabilities were associated with more than one measure of quality, and no capabilities were associated with better performance on depression care or overuse of services.

Researchers say their findings are relevant to ongoing discussions about the potential benefits of broadly adopting electronic health records across the nation's health care system. Recent federal legislation has called for new incentives for physicians who make "meaningful use" of the technology.

"Electronic health records with advanced features are uncommon nationally," Friedberg said. "Our results suggest that increasing their adoption may help improve the quality of care in important areas of preventive care and chronic disease management."

The study also has implications for "medical home" demonstration projects—ongoing efforts to improve the quality of medical care by investing in the capabilities of primary care practice groups. Researchers say their study's findings may help guide expectations about the magnitude of quality improvements that may be possible from these investments.

Support for the study was provided by the Commonwealth Fund. Other authors of the study are Kathryn L. Coltin of Harvard Pilgrim Health Care, Dana Gelb Safran of Blue Cross Blue Shield of Massachusetts, Marguerite Dresser of Massachusetts Health Quality Partners, Alan M. Zaslavsky of Harvard Medical School, and Dr. Eric C. Schneider of RAND, Brigham and Women's Hospital, and the Harvard School of Public Health.

RAND Health, a division of the RAND Corporation, is the nation's largest independent health policy research program, with a broad research portfolio that focuses on quality, costs and health services delivery, among other topics. RAND Health is the developer of COMPARE (Comprehensive Assessment of Reform Efforts), a one-of-a-kind online resource that provides objective analysis about national health care reform proposals. Visit to learn more.

About fifteen years ago, I was introduced to a graph that showed three curves that represented potential cost savings from technology.

Curve one - drive down administrative costs in healthcare through EDI, reducing clerical FTEs, etc. It's the earliest peak at it's 10% of $500B potential.

Curve two - drive down the medical management costs peaks next, maybe 2 to 5 years out, and valued at 40% of the opportunity, or $200 B.

and finally, Curve three - drive wellness, last one to bear fruit at 3 to 10 years, and 50% or $250 B of the opportunity.

This was an oft touted vision for HCIT as stated and quantified in 1995, by some of the same people who had just finished a decade's work on the economics of rolling in DRGs.

It would be helpful if the EHR value proposition was broken down in some comparable way, with contribution and time frames stated.

Presumably, an installed but hardly implemented EHR would deliver on curve one but not two and three.

Type of Contribution
Time of Initial Impact
Economic Impact
1-3 years
50 B$
Medical Management
2-5 years
200 B$
3-10 years
250 B$

James Feldbaum

Jim Feldbaum is a physician consultant specializing in clinical transformation, CPOE, and...