To assess associations between heat exposure and suicide risk in Bangladesh and examine whether sleep quality or emotional dysregulation statistically mediated those associations.
Approach:
Design and participants: Two-round longitudinal panel study using nationally representative Bangladesh survey data collected in winter and summer 2024. A balanced panel included 7,042 participants aged 15 years or older.
Measures and analysis: Researchers linked survey responses to weather-station estimates of days above several temperature thresholds in the prior 2 weeks. Adjusted Tobit models assessed suicide risk scores; structural equation models examined statistical mediation by sleep quality a…
Key Findings:
Each additional day with a maximum temperature of at least 35°C in the prior 2 weeks was associated with an approximately 1-point higher adjusted suicide risk score. Associations were smaller at 28°C to 32°C; the 25°C threshold was not significantly associated…
In the mediation model, poor sleep quality accounted for 29.5% of the modeled total association (indirect association, 0.08 points), compared with 1.8% for emotional dysregulation (0.01 points). The modeled total association was 0.29 points.
Population-weighted mean suicide risk scores rose from 11 in winter to 13 in summer; sleep quality and emotional dysregulation scores also worsened.
Sensitivity analyses found the association at the 35°C threshold remained consistent. Heat exposure 15 to 28 days before the survey was not significantly associated with suicide risk.
Interpretation:
The observational findings show an association between recent extreme heat and higher suicide risk scores, with poor sleep accounting for a larger share of the modeled indirect association than emotional dysregulation. The study could not establish causation.
Limitations:
Sleep, emotional regulation, and suicide risk were measured concurrently, preventing confirmation of causal mediation.
Suicide risk was measured by questionnaire, not observed attempts or deaths.
Outdoor weather-station estimates may not reflect individual exposure, indoor conditions, or nighttime temperatures.
Data were unavailable for several potential confounders, including humidity, air pollution, psychiatric disorders, substance use, stressful events, prior attempts, and access to mental health care.
Conclusion:
The investigators called for further research with more frequent or temporally separated measurements to assess whether the proposed pathways are causal. They suggested considering cooling infrastructure and psychological interventions in climate adaptation strategies.
