Physicians Using Data Exchanges Order Fewer Lab Tests, Study Finds

Physicians who can access patients' prior test results through a health information exchange order fewer laboratory tests than doctors without such access, according to a study published in the Archives of Internal Medicine, Reuters reports.

Study Details

For the study, researchers examined care provided to 117,606 outpatients at Brigham and Women's Hospital and Massachusetts General Hospital between January 1, 1999 and December 31, 2004. The two hospitals established a health information exchange in 2000.

Among the patients studied, 346 had received recent tests at the other hospital and 44 of those patients had received the tests before the data exchange was implemented.

Key Findings

In 1999 -- before the data exchange was implemented -- physicians ordered an average of seven lab tests per patient, according to Alexander Turchin, one of the study's lead authors and an assistant professor at Harvard Medical School. In 2004, that number decreased to four lab tests per patient, Turchin noted.

When researchers looked only at patients who did not have prior lab tests available, they found that the amount of tests ordered increased slightly from five tests per patient in 1999 to six tests per patient in 2004.

When researchers looked only at patients who had prior lab tests available, the number of tests ordered decreased by about 49% after the implementation of the exchange. After accounting for factors such as age and gender, the number of tests ordered per patient with prior lab tests available decreased by about 53% after the exchange's implementation.

Findings Appear To Conflict With Recent Study

The findings appear to conflict with a recent study published in the journal Health Affairs (Seaman, Reuters, 3/29).

The Health Affairs study -- by the Cambridge Health Alliance -- found that physicians with electronic access to patients' previous imaging results ordered tests 40% more frequently than their peers using paper-based records (iHealthBeat, 3/13).

Danny McCormick -- author of the Health Affairs study -- said the results of the two studies could be different for several reasons, including the fact that researchers looked at different populations.

Turchin noted that the two studies had different methodology and that his study did not include imaging results like McCormick's study (Reuters, 3/29).

Peter Basch
As someone who questioned the validity of the results of the McCormick study, I would ask Hebel et al. the same questions re maturity of software, training of providers, workflow design, and economic incentives for over- or under- ordering of tests. I am also curious if Hebel et al. looked at actual use of the internal HIE - and did not simply assume that because it was available that it was used. As I articulated in my critique of McCormick, supporters of health IT should not be unduly distressed by the seeming negative results of retrospective correlative studies; nor should they be overly excited by similar studies which appear to show benefit. What would make this study more interesting; redo the analysis using the lens of appropriateness. Thus, did providers reduce unnecessary testing because within the workflow of ordering new tests, they saw recent results that made additional appropriate testing unnecessary.

to share your thoughts on this article.