Community Mental Health Service Transport Voucher Eligibility and Missed Antipsychotic Injections Among Rural Namibian Adults: A Regression-Discontinuity Estimation of Travel-Cost Exposure

Travel distance to community mental health services is widely assumed to shape antipsychotic adherence in rural Africa, yet the causal status of travel-cost exposure remains poorly identified because distance and need are jointly determined. The design treats the voucher threshold as a source of quasi-random variation in out-of-pocket travel burden, defines a sharp discontinuity in exposure at the eligibility cut-off, and specifies the assumptions under which a local average treatment effect is identified. Using the DHS Program indicator 'Knowledge of any method of contraception among married women' (FP_KMET_W_ANY) as a calibration benchmark for the survey infrastructure that would generate the outcome data, the paper reports the indicator's distribution across 162 country-survey observations in 38 African countries between 1986 and 2025 (mean 91.776, standard deviation 11.342, minimum Chad 57.80, maximum Eswatini 99.90). The contribution is methodological: it specifies the identification conditions, the bandwidth and polynomial choices, the manipulation and covariate-balance diagnostics, and the interpretation of the estimand for a policy-relevant travel-cost exposure, and it argues that the voucher threshold is a defensible instrument only when distance is measured administratively rather than self-reported. The paper concludes that the design can support credible inference on travel-cost exposure where the threshold is sharp, the running variable is not manipulable, and the outcome is ascertained through clinic records rather than patient recall.