Algorithm for identifying PTSD in electronic health records
Researchers with the VA’s Million Veteran Program have designed a computer algorithm to identify post-traumatic stress disorder (PTSD) in electronic health records. Although PTSD may be noted in a patient’s electronic health record, researchers would need to read individual records to pick out patients with the condition for studies.
The algorithm searches records for diagnoses and symptoms to classify patients as likely PTSD, possible PTSD, and likely not PTSD, and had high accuracy when compared with manual chart review, according to the agency. The algorithm could be useful for research and quality improvement endeavors within the VA, say the researchers. A quick way to identify thousands of cases using electronic health records would help to recruit for large genome-wide association studies of PTSD.
An article describing the research recently appeared in the Journal of Traumatic Stress.