JMIR Publications

ABSTRACT

Background:

New technologies are changing the access to medical records and the relationship between physicians and patients. Professionals can now use e-mental health tools to provide prompt and personalized responses to patients with mental illness. However, there is a lack of knowledge about the digital phenotype of patients that use e-mental health.

Objective:

To reveal the profiles of users of a mental health app through machine learning techniques.

Methods:

We have applied a non-parametric model, the Sparse Poisson Factorization Model (SPFM), to discover latent features in the response patterns made by 2254 psychiatric outpatients to a short self-assessment on general health. The assessment was completed through a mental health app after the first login.

Results:

The results show four different profiles of patients: 1) the common pattern for all patients included feelings of worthlessness, aggressiveness and suicidal ideas; 2) one in four reported also low energy and difficulties to cope with problems; 3) less than a quarter of the sample described depressive symptoms with extremely high scores in suicidal thoughts and aggressiveness, and 4) a small number, possibly with the most severe conditions, reported a combination of all these features.

Conclusions:

E-mental-health user profiles do not overlap with clinician-made diagnoses. Since each profile seems to be associated with a different level of severity, they could be useful for the prediction of behavioural risks among users of e-mental health apps.