Prevalence and social ecological factors associated with mental health issues among teenage mothers in Rwamagana district of Rwanda: a community based cross sectional study

Abstract Background Teenage motherhood can be accompanied by substantial psychosocial stress, but evidence on psychological symptoms among teenage mothers in rural Rwanda is limited. This study estimated the proportion of teenage mothers in Rwamagana District who met a study-defined threshold for mental health issues and explored associated factors across social-ecological levels. Methods A cross-sectional study included 96 teenage mothers aged 18 to 19 years with at least one living child, sampled from three sectors. Structured interviews were conducted in Kinyarwanda. The outcome was defined as at least two of nine symptom domains occurring “frequently” or “almost every day” during the previous two weeks. This study-developed threshold was not validated against a clinical diagnostic interview. Separate Firth penalized logistic regression models were fitted for each social-ecological level. Results Twenty-one participants (21.9%; 95% CI: 14.3–31.7) met the study-defined threshold. In separate ecological-level models, lack of family support (OR 5.48, 95% CI: 1.90–17.10; p = 0.002) and fear of disclosing the partner because of possible legal penalties (OR 5.23, 95% CI: 1.38–29.68; p = 0.013) were associated with higher odds of meeting the threshold. Financial support, lack of knowledge about legal services, and financial independence were also associated with the outcome, but estimates were imprecise, and the directions of the two financial associations were unexpected. Conclusions Approximately one in five participants met the study-defined threshold for mental health issues. The findings suggest that family support and concerns about partner disclosure warrant further investigation, but the small cross-sectional study and unvalidated symptom threshold do not support diagnostic or causal conclusions. Larger studies using validated measures and longitudinal designs are needed.