EHRs Detect Depression When Many PCPs Can't
By Katie Wike, contributing writer
Depression is one of the hardest disorders to diagnose, yet it affects 14 percent of the world’s population. Researchers have found factors in EHRs may be key to predicting a diagnosis of depression.
While depression comes at a high cost to those who suffer from it, the actual price tag in the United States reaches over $44 billion annually. This takes into account, among other things, lost productivity and direct expenses. Depression is a diagnosis that is often missed by primary care physicians, despite the fact that it is the second most common chronic disorder they treat.
According to EHR Intelligence, researchers from Stanford University have worked to use EHR systems as a tool to help predict depression diagnoses. In the study, published by the Journal of the American Medical Informatics Association, researchers say valuable information already stored in the EHR can be used to predict depression up to a year in advance.
“Depression is a prevalent disorder difficult to diagnose and treat. In particular, depressed patients exhibit largely unpredictable responses to treatment,” explain researchers. “Many depressed patients are not even diagnosed … primary care physicians, who deliver the majority of care for depression, only identify about 50 percent of true depression cases.”
The Stanford team used EHR data including demographic data, ICD-9, RxNorm, CPT codes, progress notes, and pathology, radiology, and transcription reports. From these, they used a model which factored in three criteria: the ICD-9 code, the presence of a depression disorder term in the clinical text, and the presence of an anti-depressive drug ingredient term in the clinical text.
These factors were then compared to predict a diagnosis of depression, response to treatment, and determine the severity of the condition.
“It is possible to use EHR data to predict a diagnosis of depression up to 12 months in advance and to differentiate between extreme baseline levels of depression. The models use commonly available data on diagnosis, medication, and clinical progress notes, making them easily portable,” concluded researchers. “The ability to automatically determine severity can facilitate assembly of large patient cohorts with similar severity from multiple sites, which may enable elucidation of the moderators of treatment response in the future.”