ATLANTA -- A mobile app that assessed cellphone activity and usage correlated well with clinically assessed depression scores, researchers reported here.
In a pilot study of the SOLVD app, there was significant correlation between PHQ-9 scores and user behavior, particularly for the more severe cases of depression, according to , of Baylor College of Medicine in Houston, and colleagues.
"We did find correlations between PHQ-9 results and activity -- so how much walking did they do, how much did they text," Moukaddam said during a press briefing at the American Psychiatric Association (APA) annual meeting. "When depression got worse, they didn't text message as much, for instance. There is that psychomotor impairment that came through."
Action Points
- Note that this study was published as an abstract and presented at a conference. These data and conclusions should be considered to be preliminary until published in a peer-reviewed journal.
Development of apps for health conditions is an "exploding area," said APA president , of the University of California San Francisco. Binder helped develop an "innovation lab" at this year's meeting, where participants can pitch their ideas for smartphone apps relevant to psychiatry.
That's been fostered by FDA's decision to exercise "" for medical apps, by which it doesn't intend to enforce requirements on lower-risk apps, such as those that help patients self-manage their disease, track health information, or document and communicate information to their physicians.
Moukaddam said her group is trying to develop an app that can help better assess depression.
"When we treat people, we are relying on what they are telling us, or we can interview the family, but it's a complicated process and it's a not black-and-white result," she said.
The Smartphone and Online Usage Based Evaluation for Depression (SOLVD) app regularly asks how patients are feeling via a self-rated visual scale. It also monitors smartphone usage data, in order to get a better idea of how a patient experiences their disease on a day-to-day basis, which was something patients seem comfortable with, she said.
"They have a Fitbit or some other technology and it has become acceptable that something can be monitoring your activity," she explained.
The usage monitoring data is intended to look at how often and how much time patients spend on their calls; how often and how frequently they are sending text messages; and how much they use their Internet browser and other apps, in order to better understand their mood. "'Are you calling more people? Are you clicking on your phone at 3 a.m.?'... things like that," Moukaddam said.
The current study was designed to take that information and see how well it correlates with gold standard measures of depression, such as the PHQ-9, the Hamilton Rating Scale for Depression (HAM-D), and the Hamilton Anxiety Rating Scale (HAM-A).
The researchers enrolled a total of 25 patients, mean age 51, who came to the clinic every 2 weeks for a basic psychiatric visit in which all of those ratings scales were deployed.
The adherence rate to the daily self-reported mood input was about 82%, and they achieved a 95% attendance rate for clinic visits.
Moukaddam said patients were giving the clinical team feedback about the app, and that they seemed to be more comfortable with the data monitoring component if their doctors saw the data.
"I think we see that in clinical practice, when somebody comes in with their band data and says look, doc, I've been sleeping 7 hours this week," she said. "We know patients were more comfortable when they did that."
Overall, they found a correlation between self-reported mood data and clinician-rated depression, and they also saw relationships between the PHQ-9 and specific aspects of cellphone usage monitoring, including daily steps taken, text message frequency, and time spent text messaging.
She noted, however, that correlation was better when patients were severely depressed, and that it wasn't seen in milder cases of depression.
That may suggest that the app may be particularly useful in those with more severe depression -- a PHQ-9 score of 14 and up -- and for being able to predict depression severity "even if the patient is not telling you everything, because that activity element does change," she said.
"It may serve as an educational tool for patients and as an adjunct for information gathering for the physician, and as a tool to improve the physician-patient relationship," she said, adding that her group is in the process of expanding the study the app's role in adolescent and perinatal depression.
, of Harvard, has been working on a similar app that's designed specifically for research, not for a commercial setting. It also tracks mobile user data in order to figure out what new lessons psychiatrists can learn from that information.
"It's like having an entirely new lens on psychiatric illnesses," Torous told 51˶. "We have a whole new way to measure things, and we could potentially learn a lot of new things. But we need to approach it scientifically and there are a lot of ethical concerns."
Chief among those are how companies may be using or selling patient data, he said: "You have to consider your data and who you are giving it away to and what they are doing with it."
Disclosures
Moukaddam and co-authors disclosed no relevant relationships with industry.
Primary Source
American Psychiatric Association
Truong A, et al "Concordance of clinical & electronic data in assessment of depression: Findings from the smartphone and online usage based evaluation for depression (SOLVD) study" APA 2016; Abstract 81.