Validity of Mobile Mood Thermometer App (mMTA) to Identify Psychiatric Morbidity and Suicide Ideation: A Population-Based Study in Taiwan

Jia-In Lee ； Chia-Yi Wu ； Ming-Been Lee ； Chia-Ta Chan ； Shih-Cheng Liao ； Chun-Ying Chen

《Journal of Suicidology》 17卷1期 Vol.17 No.1 (2022/03) Pp. 72-79

https://doi.org/10.30126/JoS.202203_(1).0012

摘要

Objective: Suicide is a major health issue worldwide. A majority of suicidal deaths occur at their first attempt. Thus, early identification of suicidal ideation (SI) and its associated risks is crucial to prevent suicide. The five-item Brief Symptom Rating Scale (BSRS-5) was proven to be a satisfactory instrument to screen psychiatric morbidity or SI in Taiwan. A mobile BSRS-5 app (nick-named mood thermometer, mMTA) has been developed and launched in 2018. The present study aimed to assess the validity of the mMTA in identifying psychiatric morbidity (PM) and SI in the general population. Methods: A nationwide data of the BSRS-5 extracted from the mMTA records in a three years were used for analysis. In total, 73,728 respondents downloaded the mMTA and 353,135 records of BSRS-5 ratings were collected during the study period. Each record, comprising 5 items of psychopathology from the BSRS-5, an additional item of suicidal ideation and demographic information, was regarded as a study subject. The regression analysis and ROC curve were performed to test the validity of the mMTA to identify the PM and SI. Results: Among the 353,135 subjects, there were 97,797 females (27.69%) and 255,338 males (72.31%). The profile of age distribution by gender was similar with a great majority in ages of 15-44 (male= 87.1% and female=96.8%, respectively). The females presented significantly higher rates of PM (37.1%) and SI (4.19%) than the males (13.39% and 1.79%, respectively). The subjects with ages under 15 or above 65 presented a significantly higher prevalence of PM and SI than others. The logistic regression analysis on SI revealed that all BSRS-5 items significantly predicted the presence of SI with odds ratios as follows: inferiority (6.71), depression (6.30), hostility (3.44), anxiety (2.00) and insomnia (1.81). Further, the stepwise multiple regression on SI indicated that all five items significantly explained 47.5% of the SI variance. Using 5/6 of the BSRS-5 score as a cut-off determined by ROC analysis to predict SI, the accurate classification rate of SI was 87.0% (with sensitivity=0.86 and specificity=0.88). Conclusion: The BSRS-5 in mMTA was valid to determine the PM and SI among the general population users. The profile of psychopathology (psychiatric morbidity and SI) by age reflected the recent trend of suicidal mortality in Taiwan.

關鍵字

suicidal ideation ； psychopathology ； risk factors ； mood thermometer app ； psychometric evaluation ； BSRS-5