EMR Data Often “Inaccurate” Or “Missing”, Study Says

Posted on September 17, 2012 I Written By

Anne Zieger is veteran healthcare consultant and analyst with 20 years of industry experience. Zieger formerly served as editor-in-chief of FierceHealthcare.com and her commentaries have appeared in dozens of international business publications, including Forbes, Business Week and Information Week. She has also contributed content to hundreds of healthcare and health IT organizations, including several Fortune 500 companies. Contact her at @ziegerhealth on Twitter or visit her site at Zieger Healthcare.

EMR adoption continues to march forward, spurred not only by Meaningful Use requirements but also the need for doctors to access data remotely and the rise of cloud infrastructure to support such initiatives.  According to research firm IDC, 80 percent of healthcare organizations should adopt EMRs by 2016.  Pretty much what you might expect.

Hopefully, this will have a positive impact on clinical care. However, EMRs may be less useful than they should be for population health research, as data is often inaccurate or missing, according to a new report published in the Journal of  The Medical Informatics Association.

Researchers behind the report said that while data from EMRs can be useful, it’s prone to certain types of errors which undermine its value.  For example, EMR data accuracy varies depending on whether the patient was treated during the day or during the night, in part because patients at night are often sicker, according to Dr. George Hripcsak, a professor of biomedical informatics at Columbia, who recently spoke with eWeek magazine.

Another issue of concern is that patient symptoms are often poorly documented in EMRs before death. For example, patients with community-acquired pneumonia who enter the ED and die quickly don’t have symptoms entered into the EMR before they die. Later on, their medical records make it look as though a healthy patient died, the researchers note.

Dr. Hripcsak told the magazine that researchers in informatics, computer science, statistics, physics, mathematics, epidemiology and philosophy will need to work together to get an accurate read on EMR data and avoid biases. (Whew!)

Clearly, the kind of teamwork Dr. Hripcsak has in mind will take a great deal of resources. They’re on their way, it seems. For example, I’m betting that the new Johns Hopkins center for population health IT will serve as a model for the kind of interdisciplinary efforts he’s describing. But that’s just one effort. It will be interesting to see whether other universities follow in Johns Hopkins’ footsteps.