Validity and Reliability of the Self-administered Psycho-TherApy-SystemS (SELFPASS) Item Pool for the Daily Mood Tracking of Depressive Symptoms: Cross-sectional Web-Based Survey (Preprint)

Abstract

BACKGROUND E-mental-health applications targeting at depression and anxiety have gained increased attention in mental health care. Daily self-assessment is an essential part of e-mental-health apps. The app SELFPASS (Self-administered-Psycho-TherApy-SystemS) is a self-management app to manage depressive and anxious symptoms. A self-developed item pool with 40 depression items and 12 anxiety items is included to provide symptom specific suggestions for interventions. However, the psychometric properties of the item pool have not yet been tested. OBJECTIVE The aim of this study is to investigate the validity and reliability of the SELFPASS item pool that has been developed for an internet-based daily self-assessment of depressive and anxious symptoms. METHODS An online link with the SELFPASS item pool and validated mood assessment scales were distributed to healthy subjects and patients who had received a diagnosis of a depressive disorder within the last year. Two scores were derived from the SELFPASS item pool: SELFPASS depression (SP-D) and SELFPASS anxiety (SP-A). The reliability was examined using Cronbach’s α. The construct validity was assessed via Pearson correlations with the Patients Health Questionnaire-9 (PHQ-9), the General Anxiety Disorder Scale-7 (GAD-7) and the WHO-5-Wellbeing-Scale (WHO-5). A logistic regression was performed as an indicator for concurrent criterion validity of SP-D and SP-A. A factor analysis provides information about the underlying factor structure of the item pool. Item-scale-correlations were calculated in order to determine item quality. RESULTS A total of n=284 participants were included, with n=192 (67.6%) healthy subjects and n=92 (32.4%) patients. Cronbach’s α was α=0.94 for SP-D and α=0.88 for SP-A. We found significant positive correlations of SP-D and PHQ-9 (r=0.87, P<.001), SP-A and GAD-7 (r=0.80, P<.001), and negative correlations of SP-D and WHO-5 (r=-0.80, P<.001) and SP A and WHO-5 (r=-.69, P<.001). Increasing scores of SP-D and SP-A led to increased odds of belonging to the patient group (SP-D: OR=1.03 (1.01 – 1.05), P<.001; SP-A: 1.05 (1.05 – 1.01), P=.01). The item pool showed two factors with one that consisted of mood-related items and another factor with somatic-related items. CONCLUSIONS The SELFPASS item pool showed good psychometric properties in terms of reliability, construct and criterion validity. However, the underlying factor structure could not be reduced to the two diagnostic categories depression and anxiety, but to a more mood related and a rather somatic factor. Few items should be replaced for future use.