Geographic distribution and determinants of mental health stigma in central Mozambique

Abstract

Purpose: This study describes patterns of community-level stigmatizing attitudes towards mental illness (MI) in central Mozambique.

Methods: Data for this study come from a representative community household survey of 2933 respondents ≥ 18 years old in Manica and Sofala Provinces, Mozambique. Six MI stigma questions represented primary research outcomes. Bivariate and multivariable analyses examined the relationship between key explanatory factors and each stigma question. Spatial analyses analyzed the smoothed geographic distribution of responses to each question and explored the association between geographic location and MI stigma controlling for individual-level socio-demographic factors.

Results: Stigmatizing attitudes towards MI are prevalent in central Mozambique. Analyses showed that males, people who live in urban places, divorced and widowed individuals, people aged 18-24, people with lower education, people endorsing no religion, and people in lower wealth quintiles tended to have significantly higher levels of stigmatizing attitudes towards MI. Individuals reporting depressive symptoms scored significantly higher on stigmatizing questions, potentially indicating internalized stigma. Geographic location is significantly associated with people's response to five of the stigma questions even after adjusting for individual-level factors.

Conclusion: Stigmatizing attitudes towards MI are common in central Mozambique and concentrated amongst specific socio-demographic groups. However, geographic analyses suggest that structural factors within communities and across regions may bear a greater influence on MI stigma than individual-level factors alone. Further implementation science should consider focusing on identifying the most significant modifiable structural factors associated with MI stigma in LMICs to inform the development, testing, and optimization of multi-level stigma prevention interventions.

Keywords: Determinants of stigma; Geographic information systems; MI stigma; Mozambique; Spatial analysis.

Fig. 1

Smoothed geographic maps of response to the question “are you willing to have a friend with MI”. Left: unadjusted map, deviance explained = 1.59%, p< 0.01; right: adjusted map, deviance explained = 5.25%, p < 0.01. Notes. Smoothed maps of predicted prevalence were produced using a locally weighted regression smoother in a generalized additive model (GAM): Logit (P) = α0 + S(lat, lon); Logit(P_a) = α0 + β1(x1) + S(lat, lon). P is the probability of reporting any stigma versus reporting no stigma; × 1 is the linear combination of demographic variables (age, gender, marital status, education, wealth quintile); and S is the spatial smoothing term of the log odds of reporting any stigma relative to reporting no stigma over the geographic extent of the study area. We adjusted for individual-level factors in the model to estimate the residual spatial surface and test whether it was significantly different from a flat surface. We then plotted the ‘residual’ surface to explore the spatial clustering of individuals reporting stigma relative to individuals not reporting stigma beyond that explained by individual-level covariates

Fig. 2

Smoothed geographic maps of response to the question “are you willing to let someone with MI to take care of your children”. Left: unadjusted map, deviance explained = 1.01%, p < 0.01, right: adjusted map, deviance explained = 3.05%, p < 0.01

Fig. 3

Smoothed geographic maps of response to the question “are you willing to assist someone with MI”. Left: unadjusted map, deviance explained = 2.16%, p < 0.01; right: adjusted map, deviance explained = 4.19%, p < 0.01

Fig. 4

Smoothed geographic maps of response to the question “do you agree that people with MI should be chained, tied and locked in their homes”. Left: unadjusted map, deviance explained = 0.887%, p < 0.01; right: adjusted map, deviance explained = 2.11%, p = 0.45

Fig. 5

Smoothed geographic maps of response to the question “do you agree that it is possible to catch MI from treating or helping someone with MI”. Left: unadjusted map, deviance explained = 0.84%, p < 0.01; right: adjusted map, deviance explained = 2.58%, p < 0.01

Fig. 6

Smoothed geographic maps of response to the question “do you agree that mental health problems are caused by witchcraft or a curse placed on a person by someone else”. Left: unadjusted map, deviance explained = 4.83%, p < 0.01; right: adjusted map, deviance explained = 6.69%, p < 0.01