According to the Centers for Disease Control (CDC), an estimated 50% of Americans are diagnosed with a mental health disorder at some point in their lifetime, and these disorders are the third most common cause of hospitalization.
Diagnosis and treatment for mental health disorders can be a time and resource-intensive process for patients and providers alike. Could data from digital fitness trackers help?
In a recent study published in npj Digital Medicine, a research team from the Department of Psychiatry at Massachusetts General Hospital details a promising new approach for using fitness tracking data to detect changes in symptom severity for individuals with bipolar disorder and major depressive disorder (MDD).
Researchers at Mass General and elsewhere have been interested in translating fitness tracker data into mental health biomarkers because the data collection method is scalable, unobtrusive and cost-effective. Data on activity levels could be particularly helpful for MDD and bipolar disorder, as both disorders are characterized by changes in energy and movement, as well as erratic sleeping patterns.
In their study, investigators Nicholas C. Jacobsen, Hilary Weingarden, PhD, and Sabine Wilhelm, PhD, used a machine-learning algorithm they developed to re-analyze publicly available fitness tracker data from 23 participants with primary MDD or bipolar disorder, as well as data from 32 healthy controls. The participants wore fitness trackers at all times for two weeks, except for when bathing.
By looking at changes in activity and sleeping patterns over the course of two weeks, the algorithm correctly predicted the patient’s diagnosis 89% of the time.
The algorithm also had success predicting changes in symptom severity over the course of two weeks by identifying telltale changes in activity patterns. The predictions were cross-checked against participants’ self-reported changes in symptom severity.
While more research needs to be done to confirm the results in larger data sets, it’s an encouraging start. Combining digital data with more traditional approaches to mental health services could help clinicians make quicker, more accurate diagnoses and provide patients with more timely treatment when symptoms change.
For example, if a patient’s fitness tracker detects a change in activity that indicates the worsening of depressive symptoms, the device could alert the patient and suggest coping strategies. The device could also alert the clinician, who could follow up with the patient to see if additional treatment is needed.