• Will health IT increase productivity in health care?

    This is a TIE-U post associated with Jonathan Oberlander’s Political Dynamics and Policy Dilemmas (UNC’s HPM 757, Fall 2011). For other posts in this series, see the course intro.

    In a 2005 Health Affairs article, a team from RAND, Hillestad et al., [1] estimate massive savings from widespread adoption of electronic medical record (EMR) systems.

    EMR implementation and networking could eventually save more than $81 billion annually—by improving health care efficiency and safety—and that HIT-enabled prevention and management of chronic disease could eventually double those savings while increasing health and other social benefits.

    Productivity in the health sector is relatively low and raising it could go a long way toward addressing the rapidly growing US health budget, as the authors illustrate.

    The authors advise caution in interpreting the estimated effects of EMR adoption, stressing they are reporting “potential savings,” by which they mean those “assuming that interconnected and interoperable EMR systems are adopted widely and used effectively.”

    I’ll add my own word of caution: the study is not based on anything like a comparison (randomized or otherwise) between systems that have and have not implemented EMRs of the type assumed for analysis. This is, as is essentially admitted by the authors, a thought experiment about a world not like the one in which we currently live. (There’s nothing wrong with that. But it does threaten the generality of the findings.)

    Here’s another cautionary note, this time from the CBO:

    [T]he RAND study was based solely on empirical studies from the literature that found positive effects for the implementation of health IT systems; it excluded studies of health IT that failed to find favorable results.

    In 2006, also in Health Affiars, Jaan Siderov [2] throws some more cold water on the idea that EMRs can produce the kind of productivity increases and savings Hillestad et al. suggest might be possible. Citing literature, he writes that

    • “EHR[s] increased documentation time among physicians by approximately 17 percent, while computerized provider order entry (CPOE) increased it by 98 percent.”
    • “EHR[s] can ‘auto-populate’ or scour the record to justify a greater intensity of service. Accordingly, ‘increased coding levels’ account for the return on investment.”
    • “EHR decision support has no effect on adherence to primary care guidelines for asthma or angina management; it leads to ‘variable’ and ‘limited’ adherence to diabetes and coronary artery disease reminders; it has no effect on evidence-based interventions for heart disease and heart failure; it causes no change in the care of patients with depression; it leads to ‘unwieldy’ tracking and monitoring of preventive health and chronic illness; and it has no impact on diabetic glucose control.”
    • “75 percent of physician respondents admitted ignoring reminder icons, and more than half seldom or never acted on the information.”
    • “‘Profit center’ laboratory or radiology departmentswill not necessarilywelcome EHR-based interventions that lead to fewer tests and less revenue.”
    • “More lives saved will come at a heavy price.”

    Siderov cites research in support of these statements. I’ve omitted them because there is an ungated version of Siderov’s paper available. However, all are just as plausible, if not more so, than the assumptions made by Hillestad et al. They suggest to me that much more than effective use of widespread, interoperable EMRs (or EHRs)* is necessary to achieve dramatic reductions in health spending and improvements in health. This sentiment is echoed by others as well (e.g. DesRoches et al. [3] in Health Affairs, 2010).

    A lot of what must be done has nothing to do with information technology at all. That’s not to say health IT can’t play a role. It can. It just isn’t so clear that it will be the cause of whatever success we achieve (if any).

    * I have glossed over the differences between EMRs and EHRs. One of the papers discussed uses the former, the other the latter. Do you think those distinctions matter to what they or I have written? If so, let me know.


    [1] Richard Hillestad, James Bigelow, Anthony Bower, Federico Girosi, Robin Meili, Richard Scoville, and Roger Taylor. Economics Of Health Information Technology: Can Electronic Medical Record Systems Transform Health Care? Potential Health Benefits, Savings, And Costs. Health Aff. September 2005. 24:51103-1117; doi:10.1377/hlthaff.24.5.1103

    [2] Jaan Sidorov. It Ain’t Necessarily So: The Electronic Health Record And The Unlikely Prospect Of Reducing Health Care Costs. Health Aff. July 2006 25:41079-1085; doi:10.1377/hlthaff.25.4.1079

    [3] Catherine M. DesRoches, Eric G. Campbell, Christine Vogeli, Jie Zheng, Sowmya R. Rao, Alexandra E. Shields, Karen Donelan, Sara Rosenbaum, Steffanie J. Bristol, and Ashish K. Jha. Electronic Health Records’ Limited Successes Suggest More Targeted Uses. Health Aff. April 2010 29:4639-646; doi:10.1377/hlthaff.2009.1086

    • If there are to be savings, it will come with EHRs. Every week I reorder tests because they have been done at another hospital and I cannot obtain the results. On call, we routinely get older patients who cannot remember their history and we cannot find family. We make educated guesses. Having real data should provide better outcomes, reducing costs.

      One other thing we should note, I think, is that in the past, most of these systems really sucked. As they improve they should take much less time to use, I hope. I just recently looked at a new system that should save me about 5-10 minutes per patient, assuming it works as advertised.


      • @steve
        when I went to china on business, I put my ATM card into the machine and it spit out Yuan. About 10 min later, my wife texted me and asked if I took out money b/c she saw the transaction online. I know nothing about EMR but it strikes me that technology is not the problem? I am unsure, but if we can figure out money I don’t get why we can’t do it for medical records? Maybe just a more complicated problem? I mean these as questions: I know nothing about EMR…

    • Don- The early EMRs were clearly not designed by people who were going to use them (I am talking about hospital based and OR based EMRs). They required large amounts of extraneous data, the kind of stuff an administrator would want, but drives busy clinicians crazy. They were also poorly integrated with other systems in the hospital. Only now am I seeing systems that cut out the extras, that let you integrate in data from the rest of the hospital. Forget about getting true EHR data that would include info from another network.

      Why this is so has been a mystery to me, but I suspect the large upfront costs have been a factor. Our competitor network is on its third system, having spent many millions that were wasted. Someone, I hope, will finally figure this out, and their sales will boom, but as of right now, these systems dont work very well, at least the ones for sale. Proprietary systems like the VA and those for Kaiser/IMG have good reputations, but those are not for sale. To be fair to these software companies, they do need to figure out how to get large amounts of data, including graphics, into the record quickly and reliably. I wish they talked with more practicing docs.


      • @steve
        interesting. One key in banking seems to be that all banks want their customers to be able to use the ATMs of other banks, which means they have an incentive to cooperate. This incentive may be missing in health care and could be a key obstacle

    • Two cents from a patient/consumber perspective.

      1. Having moved around quite a bit I have had the great joy of filling out health histories an uncountable number of times over the past 45 years. I am sure that my accuracy rate in each effort was significantly less than 100% – and will probably be lucky to be better than 70% the next time I go to a new provider.

      2. I have bristled a bit over the years about the lack of transparency on my health information. Most of my Doctors seem to feel that sharing my charts – data – info is some form of state secret so I often feel that I may not be fully informed – and that has led to a certain amount of paranoia on my part – a suspicion that providers prefer me to be less than fully informed.

      3. As a Libertarian I have no desire to have government do more than they already do – but I also wonder if the efficiency payoff may lie in finding a way to integrate all of my [or any patient/consumers] data accurately and seamlessly as it is acquired. The comments above on banking led me to think there might be a role for a “medical reporting” counterpart to credit reporting. I will confess that this causes me some pretty serious privacy concerns – but I do see potential benefits too.

    • Which electronic systems work? The claims payment systems! follow the money. You want functioning clinical systems…compensate providers for using them.

      My current provider owned by an academic medical center demonstrates the disconnect between billing and clinical perfectly. At the end of the encounter (or by the end of the day) the provider reviews the medical record, enters the proper codes for procedures and diagnoses and voila the recommend E&M code. A clerical staff then prints out the summary sheet, carries these sheets for all the day’s patients to the billing office, hands off to billing clerk who re-keys the E&M codes, cpt codes, and diagnoses codes to generate an electronic bill to the third party clearing house then to the third party payer, public or private. The patient’s paper summary sheet is then scanned and electronically attached to the submitted bill for storage. The patient’s paper summary is then shredded. How is this efficient…it does however create jobs.

      A recent trip to a hospital lab for basic outpatient lab work to be paid by patient in cash, at a hospital that advertises itself as paperlessly efficient, generated no less than 7 separate sheets of paper, granted two did deal with processing credit card. Reporting results generated three more sheets via fax.

      In a world where you can program you DVR via you smart phone health care is spending inordinate resources on security and privacy at the cost of functionality. Do they still believe the myth that a patient’s medical information is private…you’ve never delivered a baby, if you had you’d know their is no such thing and patient privacy or dignity.

    • A 2008 national survey of physicians found that only 13% were using “basic” EHRs (http://www.nejm.org/doi/full/10.1056/NEJMsa0802005f). If the benefits are so great, why is it not happening? This is a genuine question and suggests to me there are benefits for the “system” (aka patients!) but not the individual provider.

      On a related note, the CBO “caution” should be taken as more than a caution. Any review that selects on positive effects will find…positive effects. Seems like a fundamental flaw to me.

    • EMRs/EHRs have been around long enough that if they were really such a big winner, we would have seen that victory a while ago. I’ve adopted them in a number of clinics and would describe the benefits as uneven. I think they’re oversold.

      This is an area where some kind of highly disruptive innovation is going to come along, because despite all the fancy add-ons you get with going paperless, the whole thing is still conceptualized as “EMR=paper chart”. I remember being in elementary school in the 80s and trying to convince my parents to buy a computer. I used the argument other kids and salesmen always used then: “But mom, you can use it to store all your recipes!” That is where we are with electronic records.

    • @steve

      A minor correction: VistA, VA’s system, isn’t proprietary. It’s completely in the public domain: http://en.wikipedia.org/wiki/VistA#Licensing_and_dissemination