So, EMRs Do Reduce Tests Ordered? Partners Says Yes

Posted on April 16, 2012 I Written By

Priya Ramachandran is a Maryland based freelance writer. In a former life, she wrote software code and managed Sarbanes Oxley related audits for IT departments. She now enjoys writing about healthcare, science and technology.

About the same time last month, I brought your guys some unwelcome news – that physician access to electronic records perhaps doesn’t reduce the number of tests subsequently ordered, and hence doesn’t reduce healthcare costs as much as previously thought.

Except that maybe it does. At least that’s according to an article in Chicago Tribune that summarizes the findings of a study by Partners Healthcare, and a research letter published in the Archives of Internal Medicine (full text, PDF).

According to the study:
– It looked at health information exchange and test data between Mass. General Hospital and Brigham and Women’s over a 5 year period from Jan. 1, 1999 to Dec. 31, 2004.
– The study looked at 117,606 patients during this period. Of these, 346 patients had recent off-site tests, of which 44 were done prior to the HIE rollout.
– The study found that for patients with recent off-site tests, there was a 49% reduction in number of tests ordered.
In number terms, the number of tests ordered per person reduced from 7 in 1999 to 4 in 2004.
– There was however a slight increase in number of tests ordered for the population that didn’t have any prior testing done during the same time period – increasing from 5 per person to 6 per person.

These findings directly contradict the Health Affairs study that I mentioned earlier. The Chicago Tribune article has a little researchers-play-nice subsection at the end where the Health Affairs and Partners researchers try to interpret each other’s contradictory results.

If I may add my 0.02:
– Even though the Partners study follows a larger population of patients, the data that is used to calculate the reduction (346 and 44) is way too small
– The Health Affairs studied some 28,000 patients spread across 1,187 doctor’s offices, while the Partners study followed a larger population of patients at two huge Mass. hospitals that entered into a partnership with each other.

While this not directly discounting anything each group has found, I would think the HA study is more representative of what’s going on in different parts of the country, where doctors are using different (in capability/costs) EMRs and labs to get their results. In Partners case there may well be a tacit agreement on EMR brand, or even tacit trust between the labs/facilities that each hospital uses.

Very interesting though, and I’d really love to see what else comes out on EMR and healthcare costs.