Traditional healer referral co-occurrence and trauma-related mental health treatment disengagement in Tanzanian districts: a difference-in-differences analysis of formalization policy rollout

Tanzania's 2015–2020 rollout of formal traditional-and-alternative-medicine (TAM) registration created district-level variation in whether traditional healers could be formally referred into the biomedical mental health system. the present analysis exploits that staggered rollout to estimate whether formalization reduced trauma-related mental health treatment disengagement, and whether any reduction was concentrated in districts where healer-biomedical referral co-occurrence was already high. Using a district-year panel of 569 observations across 24 districts (2000–2023), the analysis estimates a two-way fixed-effects difference-in-differences specification with heteroskedasticity-robust standard errors. The model fits the data well (R² = 0.4888; adjusted R² = 0.487; F = 441.152, p < 0.001). The estimated coefficient on the formalization exposure is 123.00084 (SE = 4.65884), the coefficient on urban population share is −1.34431 (SE = 0.05763), and the coefficient on the referral co-occurrence interaction is 0.46579 (SE = 0.14039). The results are consistent with the interpretation that formalization reduced disengagement primarily where referral co-occurrence was already established, and that urbanicity was associated with lower disengagement. Formalization operates as a coordination technology rather than a demand-side intervention, and that its effects are therefore conditional on pre-existing referral infrastructure. The findings have implications for the sequencing of TAM integration policies in low-resource mental health systems.