Mining EMRs speeds up genetic research

Mining electronic health records (EHRs) produced by different companies for genetic studies on specific chronic conditions still can be far faster and cheaper than trying to recruit patients to collect data, according to researchers from Northwestern Medicine in Chicago. However, EHRs still need to improve parameters that can be researched, they observed.

In the study, which appears this month in Science Translational Medicine, the researchers were able to take patient information in EHRs from routine physician visits at five national sites that used different brands of medical record software.

The researchers then used the data to identify patients with five types of diseases or health conditions--type 2 diabetes, dementia, peripheral arterial disease, cataracts and cardiac conduction. Five institutions participated in the study, and patients in advance agreed to the use of their records for studies.

To identify the diseases, the researchers reviewed the EMRs using criteria such as medications, diagnoses, and laboratory test results. They then tested their findings against reviews by the patients' physicians (who would confirm the results). The EHRs allowed the researchers to identify patients' diseases with 73 to 98 percent accuracy.

The study, however, did show across-the-board weaknesses among the institutions' EHRs: In particular, the records often fell short in capturing the patients' race and ethnicity, smoking status, and family history.

"It shows we need to focus our efforts to use electronic medical records more meaningfully," said lead investigator Abel Kho, MD, an assistant professor of medicine at Northwestern University Feinberg School of Medicine and a physician at Northwestern Memorial Hospital.

For more details:
- see the study abstract
- view the UPI article

Suggested Articles

Roche, which already owned a 12.6% stake in Flatiron Health, has agreed to buy the health IT company for $1.9 billion.

Allscripts managed to acquire two EHR platforms for just $50 million by selling off a portion of McKesson's portfolio for as much as $235 million.

Artificial intelligence could help physicians predict a patient's risk of developing a deadly infection.