Written by: Katherine Rourke
A group of researchers have completed a study which found new links between patients’ genetic profile and specific diseases by mining EMR data, reports a story in iHealthBeat.
The research, which was conducted by the Electronic Medical Records and Genomics Network, a consortium of medical research institutions including the Mayo Clinic and Vanderbilt University School of Medicine, analyzed data from about 13,000 of EMRs.
The participants then grouped about 15,000 billing codes contained in the EMRs into 1,600 disease categories. Next, they looked for links to diseases in EMRs which contained DNA data.
The researchers, whose study was published in the journal Nature Biotechnology, found 63 new genetic links to diseases, ranging from skin cancer to anemia, iHealthBeat said.
The EMR study method, which is known as a phenome-wide association study, is a departure from the 13-year old genome-wide association model, which has been used to search for common mutations in the DNA of patients of people with the same diseases.
Co-author Joshua Denny, a biomedical informatics researcher at Vanderbilt, says that the newer method can help link seemingly unrelated symptoms, detect potentially harmful side effects of a drug, and help find new uses for drugs.
This is just the tip of the iceberg where translation medicine and EMRs are concerned. Using EMRs to conduct genomic research is becoming an increasingly popular exercise, cutting across a wide range of clinical disciplines.
And it’s not just institutional academic research houses getting into the act. For example, this summer a large northern Virginia hospital announced that it had struck a deal with a Massachusetts analytics firm to see if data mined from EMRs can better predict the risk of preterm live birth.
Now, genomics research is not for just any hospital — it’s obviously a major undertaking — but I think it’s likely more hospitals will get into the game. By this time next year I think there will be a crop of interesting new genomics projects mining EMRs. Although, it will be interesting to see how the 23andMe FDA battle impacts this as well.