Monday, January 3, 2011

Here's a Whopper of a Prediction from National Coordinator Blumenthal About the EHR That Hasn't Quite Worked Out

'Tis the month when journalists, pundits, bloggers, wonks and the other denizens of the commentariat examine past predictions and make new ones. So, when a new research article on the merits of the electronic health record (EHR) came out, the Disease Management Care Blog couldn't help but recall this whopper of a forecast:

"The widespread use of EHRs in the United States is inevitable. EHRs will improve caregivers' decisions and patients' outcomes. Once patients experience the benefits of this technology, they will demand nothing less from their providers. Hundreds of thousands of physicians have already seen these benefits in their clinical practice."

That's from national coordinator for health information technology David Blumenthal writing in an August 2010 issue of the New England Journal. The Disease Management Care Blog can forgive him: as a Czar presiding over our national medicaltechnologic complex, cheerful blather is part of the job description and part of the Washington m.o.

Facts, on the other hand, are a different matter. The latest example that EHR adoption is not the Utopian panacea that the HIT weenies would have you believe is this RAND article authored by Spencer Jones, John Adams, Eric Schneider, Jeanne Ringel and Elizabeth McGlynn (yes, THAT Elizabeth McGlynn) titled "Electronic Health Record Adoption and Quality Improvement in US Hospitals."

The authors cross-referenced survey results (including having a basic or more advanced EHR system in place) from the Health Information and Management Systems Society (HIMSS) with the quality data from the American Hospital Association's Hospital Compare. The hospitals used for the study were the 3971 nonfederal general acute care hospitals. The quality measures assessed treatment of a heart attack, heart failure and pneumonia. If Dr. Blumenthal's above bluster is correct, more EHR in each of these US hospitals should translate into clinically and statistically significant increased quality, right?

Not exactly. From 2004 to 2007, there were 4-16 point performance increases in treatment of heart attack, heart failure and pneumonia in the 62% of U.S. hospitals that didn't have an EHR in place. Compared to these "no EHR" hospitals, there was no difference in the hospitals with a basic EHR or an advanced EHR - with one exception (treatment of heart failure was better among those with a basic EHR).

But wait, it gets worse. Heart attack and heart failure quality scores statistically improved less among hospitals that adopted an advanced EHR compared to hospitals that did not adopt an EHR. What's more, heart attack and heart failure quality scores improved significantly less in hospitals that upgraded their basic EHR compared with hospitals that maintained their basic EHR.

To their credit and to Dr. Blumenthal's minions' relief, the authors point out that their study was short term, was limited to a narrow set of quality measures in select hospitals; it didn't capture all possible quality measures in all health care settings. In addition, the authors note that there's a difference between having an EHR and fully as well as correctly implementing it. It's possible that a well implemented EHR does lead to real quality. It's also possible that measures of quality in heart attack and pneumonia may have reached their ceiling, making it more difficult to detect any meaningful changes one way or another. Last but not least, the authors speculated that the considerable work of installing an EHR distracted the hospitals over the short-term from continuing to work in quality improvement and that there will be a pay-off in the coming years.

Despite these limitations, the results should give pause to policymakers that have bought into the notion that EHRs will "inevitably" fix all that ails U.S. health care. In addition, this study may warn us that complicated systems lead to complicated problems; when it comes to EHRs, the truth may be that less (like computerized physician order entry or medications, lab review) is more. EHRs may eventually improve caregivers' decisions and patients' outcomes, but this study shows patients have yet to consistently experience the benefits of this technology for heart attack, heart failure and pneumonia. The truth is that there are hundreds of thousands of physicians have not seen these benefits in their clinical practice.

Dr. Blumenthal may wish to reconsult with his crystal ball.


Roger Collier said...

…and Dr Sidarov should read the RAND article more carefully.

While there’s certainly been overselling of the benefits of EHR, the article clearly shows that “more” EHR was associated with higher measures of quality at the end of the study period for every diagnosis cited. (Take a look at Table 4).

The article does also show that the measures of quality did not increase as much in EHR hospitals over the study period as in ones without EHR — but just possibly because they were doing a better job to begin with.

Aside from this confusion between absolute quality measures and increases in the measures, the whole study seems flawed by RAND’s selection of such simplistic measures. Non-performance of some of the care processes listed may certainly be indicative of quality problems, but surely most patients were receiving more extensive treatment — something that might well be influenced by EHR capabilities. The study neither addresses this, nor does the article mention the possibility.

Jaan Sidorov said...

Hi Roger!

I interpreted Table 4 by looking at the confidence intervals, which crossed 1 (the comparator), meaning they failed to achieve statistical significance. The little superscript "C" denoting p <.05 could only be applied to one measure.

Your point about the baseline, the measures that were chosen and other dimensions "on the ground" is a good point.

That being said, with our national mania for "evidence-based medicine" and "comparative effectiveness," you'd think that our headlong rush into electronic records was built more on science and less on faith.