When EHR Critics Jump To Bad Conclusions

What's your favorite sport? For some health IT critics, it's jumping to conclusions. And a recent study published in Health Affairs will certainly give them lots of exercise.

The study found that physicians with electronic access to patients' previous test results are more--not less--likely to order additional tests, which presumably would add to overall healthcare spending. Naysayers then used this finding to attack the current push to install electronic health records (EHRs) and related electronic tools at more hospitals and practices.

For example, Merrill Goozner, former chief economics correspondent for the Chicago Tribune, said the Health Affairs study "belabors the obvious." In a recent blog, he concludes that EHRs will have little impact on reducing overuse or lowering healthcare costs until they are "linked to payment and delivery system reforms ... and greater incentives for physicians to deliver evidence-based care."

[Is it time to re-engineer your clinical decision support system? See 10 Innovative Clinical Decision Support Programs.]

Goozner is right about the link between the overuse of diagnostic testing in the current fee-for-services world and healthcare expenditure, but he's wrong about the Health Affairs data itself.

Had Goozner taken a closer look at the study and put it into context with others that have evaluated the cost implications of health IT, he might have come to a different conclusion.

To reach their conclusions, the Health Affairs authors, Danny McCormick, MD, of Harvard Medical School, and his associates analyzed data from the 2008 National Ambulatory Medical Care Survey, which covered more than 28,000 patient visits in more than 1,100 doctors' offices.

When they compared physicians who had access to computerized imaging results to those who did not, the authors found that the former group was 40% to 70% more likely to order additional imaging tests.

The authors speculated that the convenience of having the test results so readily available electronically may have enticed clinicians to order more tests. But even assuming that's true, there's no way of knowing whether the additional imaging tests were warranted or not. And in fact, the researchers admit in the report: "We could not assess whether the increased imaging associated with electronic access to results helped or harmed patients." And that's the critical issue!

If the patients in the group seen by physicians with e-access were sicker than those seen in the offices that didn't have such access, or their care required more complex diagnostic reasoning to determine the cause of their symptoms, the additional imaging may have been fully justified. And if that were the case, those additional tests would have likely reduced healthcare expenditures in the long run by pinpointing the most appropriate treatment plan early on and subsequently improving clinical outcomes.

Unfortunately, there's nothing in the data set about the severity or complexity of the patients' conditions, so there's no way to know if the tests were a smart decision or not. Similarly, without any follow-up information on all the patients in each group, we don't know if those in the e-access group fared better or worse than those seen by doctors without electronic access.

Other studies have tried to determine if giving physicians electronic data on previous lab test results would reduce costs. For instance, David Bates, MD, and his colleagues at the Center for Applied Medical Information Systems in Boston spent 15 weeks providing doctors with e-alerts when they were about to order a test for inpatients that had already been performed. The doctors canceled 69% (300 of 437) duplicate tests as a result of the electronic reminders, saving $35,000 in lab charges.

Of course, while these and other studies suggest that health IT can reduce lab costs, keep in mind that most of the positive data comes from very large healthcare organizations with sophisticated and often custom-designed systems. The research done by Dr. Bates and his colleagues, for example, was done at Brigham and Women's Hospital in Boston.

Similarly, a systematic literature review of health IT's effects on quality of care and health services utilization found that computerized physician order entry (CPOE) systems that gave clinicians point-of-care details on previous tests and test costs did in fact reduce utilization. Most of those savings were in the form of fewer lab and imaging tests. But once again, the main caveat to remember is the fact that a disproportionate number of studies evaluated in the review came from clinicians working at large research institutions.

So putting the new Health Affairs report into perspective, the only thing we can say for certain is that on its own merits, it does not support critics who claim that health IT investments are a waste of money. There are just too many other ways to interpret the data. Unfortunately, the larger context also suggests that you may have to be part of a mega-healthcare organization to see significant savings.

Healthcare providers must collect all sorts of performance data to meet emerging standards. The new Pay For Performance issue of InformationWeek Healthcare delves into the huge task ahead. Also in this issue: Why personal health records have flopped. (Free registration required.)