Incorporation of Mobile Application (App) Measures Into the Diagnosis of Smartphone Addiction

Abstract

Objective: Global smartphone expansion has brought about unprecedented addictive behaviors. The current diagnosis of smartphone addiction is based solely on information from clinical interview. This study aimed to incorporate application (app)-recorded data into psychiatric criteria for the diagnosis of smartphone addiction and to examine the predictive ability of the app-recorded data for the diagnosis of smartphone addiction.

Methods: Smartphone use data of 79 college students were recorded by a newly developed app for 1 month between December 1, 2013, and May 31, 2014. For each participant, psychiatrists made a diagnosis for smartphone addiction based on 2 approaches: (1) only diagnostic interview (standard diagnosis) and (2) both diagnostic interview and app-recorded data (app-incorporated diagnosis). The app-incorporated diagnosis was further used to build app-incorporated diagnostic criteria. In addition, the app-recorded data were pooled as a score to predict smartphone addiction diagnosis.

Results: When app-incorporated diagnosis was used as a gold standard for 12 candidate criteria, 7 criteria showed significant accuracy (area under receiver operating characteristic curve [AUC] > 0.7) and were constructed as app-incorporated diagnostic criteria, which demonstrated remarkable accuracy (92.4%) for app-incorporated diagnosis. In addition, both frequency and duration of daily smartphone use significantly predicted app-incorporated diagnosis (AUC = 0.70 for frequency; AUC = 0.72 for duration). The combination of duration, frequency, and frequency trend for 1 month can accurately predict smartphone addiction diagnosis (AUC = 0.79 for app-incorporated diagnosis; AUC = 0.71 for standard diagnosis).

Conclusions: The app-incorporated diagnosis, combining both psychiatric interview and app-recorded data, demonstrated substantial accuracy for smartphone addiction diagnosis. In addition, the app-recorded data performed as an accurate screening tool for app-incorporated diagnosis.