BACKGROUND
Use of behavioral, emotional, and physiologic data from personal digital devices has been offered as a strategy to identify and monitor symptoms associated with mental health conditions. To provide an analyses of the current state of the science, Mendes and colleagues (1) conducted a systematic review of scientific literature and publicly available datasets to identify and characterize existing work regarding digital phenotyping of mental health (DPMH). In doing so, they focused on technical features of the apps (e.g., sensors used to collect data, types of context data).
OBJECTIVE
This is a secondary analysis of articles reviewed by Mendes et al. (1) and was conducted using a clinical research focus to characterize efforts related to suicide risk identification among military and Veteran cohorts, as well as rate studies identified in terms of risk of bias.
METHODS
Twenty-six studies that used active or passive assessments to phenotype mental health conditions were reviewed and included in this secondary analysis. Data from studies were extracted into evidence tables. Risk of bias rating pertaining to studies of interest was conducted.
RESULTS
Of the included articles, only three studies explicitly investigated DMPH in relation to suicide risk and only one reported military/Veteran status of the participants. One study examined the acceptability and feasibility of using a smartphone app to evaluate suicide risk via active and passive monitoring. In another study, active and passive monitoring of data from patients receiving treatment on a psychiatric inpatient unit was used to develop a suicide risk prediction algorithm via machine learning. In the third study, actively and passively collected data obtained from smartphones was retrospectively evaluated to assess suicide risk in military Veterans. All three DPMH articles related to suicide risk were noted to have a high risk of bias and the overall quality of evidence related to suicide risk was determined to be low.
CONCLUSIONS
The body of literature related to DPMH for suicide risk among cohorts of interest was highly limited, with only one study reviewed by Mendes et al. (1) focused on military/Veterans. Mendes et al. (1) suggested that the data presented within the review provided preliminary evidence that active and passive monitoring may be used to identify suicide risk. However, from a clinical research perspective evidence remains limited. To advance DPMH-related suicide prevention efforts, future research should incorporate subject matter experts, including those living with mental health conditions and care providers, at the design stage of application development, as well as move away from relying on convenience sampling. More complete reporting of sample characteristics is also indicated. Such efforts could be used to develop tools aimed at identifying and monitoring suicide risk among high-risk cohorts, including those with a history of military service.
CLINICALTRIAL
Not applicable.