<![CDATA[Self-harm is a significant public mental health concern influenced by complex interactions between psychological, social, and environmental factors. Understanding long-term patterns and identifying potential determinants of self-harm are essential for developing effective prevention strategies and counseling interventions in Indonesia. This study analyzed annual self-harm attempt rates in Indonesia from 1991 to 2025 using an autoregressive integrated moving average with exogenous variables (ARIMAX) approach. The explanatory variables included unemployment rate, inflation rate, average working hours, and a COVID-19 pandemic indicator. Several ARIMA specifications were evaluated based on model performance, parameter significance, and information criteria. Residual diagnostic tests were conducted to assess model adequacy, including examinations for autocorrelation and heteroscedasticity. The ARIMAX(1,2,1) model was selected as the preferred specification based on its statistical performance and the complete estimation of its parameters. The findings showed that average working hours (? = 0.1908, p = 0.024) and the COVID-19 pandemic period (? = 0.2275, p < 0.001) were significant predictors of self-harm rates, whereas unemployment and inflation were not statistically significant. Forecasting results indicated a gradual decline in self-harm rates, from 13.06 in 2026 to 12.34 in 2028, with increasing uncertainty over longer forecasting horizons. The findings suggest that self-harm trends in Indonesia are closely associated with psychosocial stressors, particularly occupational burden and major social disruptions. Integrating time-series evidence with counseling and public mental health strategies may support more targeted prevention efforts, early identification, and improved psychological support systems.]]>
