Monthly Archives: August 2007

EHRs, Media and Statistics: Misinterpreted Results Skew Understanding

by Jane Sarasohn-Kahn


“Electronic Health Records Didn’t Improve Quality of Outpatient Care”

“Electronic Health Records Don’t Lift Care”

“Electronic Records Don’t Always Improve Care”

“No Quality Benefits Seen with Electronic Health Records”

“Electronic Medical Records May Not Live Up to Hype”

So said some of the newspaper headlines about the July 9 Archives of Internal Medicine paper, “Electronic Health Record Use and the Quality of Ambulatory Care in the United States.”

When I read the news coverage emanating from the study, it caught me — and I suppose many of you readers — off guard. I’m not one to bash the mass media, but reporters got this latest study on electronic health records and outcomes wrong. Journalists need a quick course in statistics, and perhaps simple reading mastery, to know the difference between causality and simple association.

A highly credible and switched-on team from Harvard and Stanford universities wrote the study, which the Agency for Healthcare Research and Quality funded. For the study, researchers studied data from the 2003 and 2004 National Ambulatory Medical Care Survey published by CDC. The data set detailed EHR use coupled with 17 ambulatory care quality indicators. These indicators covered medical management of common diseases, antibiotic prescribing, preventive counseling, screening tests and other services. According to the analysis, physicians’ performance on these quality indicators was not associated with the “use” of an EHR system.

All you have to do is read the second sentence in the paper abstract’s background paragraph to realize that the researchers were assessing “the association between EHR use, as implemented, and the quality of ambulatory care in a nationally representative survey.” Herein lies the nuance of the study: the authors did not seek to address whether the installation of an EHR would result in better outcomes, as newspapers incorrectly interpreted. They simply sought an association between EHRs and quality of care — and that they did not find.

It’s also important to closely look at the second half of that introductory sentence: the simple phrase, “EHR use, as implemented” (emphasis added). That is the point.

So, before you swallow the mass media line of reasoning that “EHRs don’t work,” take a few minutes to understand what’s really in the study.

Continue article:


Rhode Island Awards Bid To Build First-Ever Statewide EHR

July 31, 2007

The Rhode Island Department of Health has awarded a three-year, $1.7 million contract to EDS to design, implement and manage the country’s first statewide electronic health record network, Healthcare IT News reports (Pizzi, Healthcare IT News, 7/30).

The contract could last up to seven years if the state uses all four of the optional one-year extensions, Government Health IT reports.

EDS will use InterSystems’ health care software to build and integrate the system.

The network, called the Rhode Island Health Information Exchange, will consolidate state residents’ health data and provide authorized hospitals, physicians, pharmacists and other health care providers with access to the health records (Wakeman, Government Health IT, 730).

The EHR network will be “developed with strict adherence to patient-consent policies and in conjunction with industry best practices with regard to security and privacy standards,” according to a press release (Providence Business News, 7/30). In addition, residents must give permission before their records are stored on the network.

The health data exchange is expected to go live in summer 2008 (Government Health IT, 7/30).