Community Mental Health Service Closure-Induced Caseload Spillovers and Crisis Readmission in South African Municipalities: A computed Control Estimation

The closure of community mental health clinics in South African municipalities is frequently justified by fiscal consolidation, yet the resulting redistribution of caseloads across the surviving service network remains poorly characterised. Because no municipality-level panel of clinic closures and readmission is presently available in the public domain, the contribution is methodological: the paper specifies the donor pool, the predictor set, the placebo inference procedure and the spillover decomposition that a credible computed control estimation of this question would require, and it calibrates the expected scale of the problem against the DHS Program indicator 'Family planning messages at community meeting or event' (FP_EFPM_M_COM), which was measured across 15 African countries between 2021 and 2025 (N = 15, mean = 29.733, SD = 17.495, range 6.60 in Mali to 73.40 in Malawi). Closure effects are not confined to the treated municipality but propagate through inter-municipal referral and transport networks, so that the appropriate estimand is a spillover-adjusted treatment effect rather than a conventional average treatment effect on the treated. It further argues that crisis readmission is the correct outcome for this design because it responds to closure on a shorter horizon than community follow-up contact and is recorded in routine district health information systems. The framework is presented as a protocol for prospective evaluation, not as a report of completed estimation, and its principal implication is that closure decisions taken without a spillover-adjusted counterfactual will systematically understate the readmission cost borne by neighbouring municipalities.