Mental health/illness and its social determinants in Uganda

Background

The overall aim of this thesis was to investigate mental health and illness in Uganda. More specifically, to explore understandings of mental health and to measure anxiety and hazardous alcohol use. Mental illness exists worldwide, but conceptualisations vary across contexts and population-level prevalence data are scarce from many countries. Information on conceptualisations and prevalences is important for evidence-based mental health promotion.

Methods

Data were collected through an open population-based cohort study in Uganda, the Africa Medical and Behavioral Sciences Organization (AMBSO) Population Health Surveillance (APHS). APHS includes urban, semi-urban, and rural sites in two Ugandan districts, Wakiso and Hoima. The backbone of APHS is an annual household census of the study communities and subsequent individual surveys of individuals >13 years covering multiple health and behavioural outcomes. Paper I provided a detailed methods description as a basis for the subsequent papers.

Paper II used data from a qualitative study nested in the first round of APHS. Three focus group discussions and 31 in-depth interviews were conducted with men and women aged 14 to 62 years recruited purposively from APHS sites to explore perceptions and knowledge of mental health. Coding followed a deductive approach. Thematic analysis was undertaken and results were organised according to the domains of the mental health literacy framework.

For paper III and IV, cross-sectional data from the 2021-22 APHS survey were used. For paper III, the Generalized Anxiety Disorder-7 scale (GAD-7) was translated into Runyoro and Luganda and certain psychometric properties (construct validity, internal consistency and concurrent validity) of the translated scales were examined. Additionally, the prevalence of anxiety symptoms and likely anxiety disorders measured with GAD-7 was assessed. In paper IV, hazardous alcohol use was measured with the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) and covariates selected according to the World Health Organization's social determinants of health framework. Analysis was restricted to individuals aged ≥18 years, using Poisson regression with robust variance to estimate adjusted prevalence ratios.

Results

Paper I described baseline findings of APHS, including that HIV prevalence was highest in semi-urban areas. In paper II, participants recognised that mental illness exists in their communities and described more introvert (e.g. social withdrawal) or erratic (e.g. using vulgar language) behaviours as typical signs of it. Financial and other social stressors were identified as primary causes of mental illness, supernatural explanatory models were not mentioned. Poor social relationships were described as barriers for coping and care-seeking.

In paper III, a total of 4107 individuals participated, of whom 2206 spoke Runyoro and 1901 spoke Luganda. Both the Runyoro and Luganda versions of GAD-7 fit the designed one-factor structure, had good (Runyoro) and excellent (Luganda) internal consistency and exhibited a moderate (Runyoro) and strong (Luganda) concurrent validity with the depression screening instrument Patient Health Questionnaire-9. Using GAD-7 >10 points as a binary cut-off, the total prevalence of likely anxiety disorders was 1.5%, 0.5% among males and 2.2% among females.

In paper IV, among the 3459 participants aged >18 years, hazardous alcohol use in the past year was prevalent among 5% of women and 18% of men. Among the social determinants found to be associated with hazardous alcohol, for both women and men, Muslim and Pentecostal religion had the strongest negative associations, female aPR 0.4 (95%CI 0.2-0.6) and male aPR 0.3 (0.2-0.5) compared with Catholic or Protestant religion. Negative associations were also found for high educational attainment and for women living in a marriage or union. Smoking and living in semi-urban areas had the strongest positive associations for both women and men, female aPR 2.1 (95% CI 1.3-3.3) and male aPR 1.8 (95% CI 1.4-2.5) comparing semi-urban with urban residence. Positive associations were also found for past-year perpetration of intimate partner violence, men in middle age and women working in the hospitality sector. HIV status was not substantially associated with hazardous alcohol use in the adjusted models.

Conclusions

For paper I, we concluded that a more nuanced distinction in epidemiological research of different types of urbanity beyond rural and urban, such as semi- urban, is important to capture specific health effects of such environments.

For paper II, we concluded that symptoms of mental illness may stem from economic and social factors that can be changed by political action. Further research should explore how economic and supernatural explanatory models of mental health and illness are related in rapidly changing African societies.

In paper III, likely anxiety disorders were less prevalent than previously reported in similar settings. Possible explanations for this discrepancy include individuals with more severe anxiety symptoms avoiding participation or GAD-7 not considering local idioms of anxiety. Given robust psychometrics but highly varying prevalence estimates, use of GAD-7 in healthcare settings should be preceded by clinical validation. Further research should also investigate whether people with more severe symptoms of mental illness are systematically missed in population cohort studies.

In paper IV, hazardous alcohol use was prevalent and specific target groups for public health and clinical interventions were identified. Residents of semi-urban areas may be exposed to unique health risks, including for hazardous alcohol use, which warrant further longitudinal investigation and mitigation through public health policy.

List of scientific papers

I. Cohort Profile: The Africa Medical and Behavioral Sciences Organization (AMBSO) Population Health Surveillance (APHS) in rural, semi-urban and urban Uganda. Mugamba S, Ziegel L, Bulamba R, Kyasanku E, Johansson Århem K, Sjöland CF, Miller AP, Nakigozi G, Nalwoga GK, Watya S, Kiwanuka N, Kagaayi J, Kiwanuka D, Ddaaki W, Wagman JA, Kigozi Go, Ekström AM, Nalugoda F International Journal of Epidemiology. 2023 Apr 19;52(2):e116-e124. https://doi.org/10.1093/ije/dyac164

II. Not Enough Money and Too Many Thoughts: Exploring Perceptions of Mental Health in Two Ugandan Districts Through the Mental Health Literacy Framework. Miller AP, Ziegel L, Mugamba S, Kyasanku E, Wagman JA, Nkwanzi-Lubega V, Nakigozi G, Kigozi Go, Nalugoda F, Kigozi Gr, Nkale J, Watya S, Ddaaki W Qualitative Health Research. 2021 Apr;31(5):967-982. https://doi.org/10.1177/1049732320986164

III. Prevalence of anxiety symptoms in a Ugandan population sample and psychometric properties of the Generalized Anxiety Disorder-7 scale (GAD-7) in Luganda and Runyoro Ziegel L, Espinosa da Silva C, Bulamba R, Daama A, Kigozi Gr, Miller AP, Kigozi Go, Kyasanku E, Mugamba S, Hammarberg A, Ekström AM, Nalugoda F, Hollander AC [Manuscript]

IV. Social determinants of hazardous alcohol use in a Ugandan population cohort. Ziegel L, Sjöland CF, Nabunya E, Bulamba R, Kyasanku E, Mugamba S, Kigozi Go, Daama A, Kigozi Gr, Miller AP, Hollander AC, Hammarberg A, Nalugoda F, Ekström AM [Manuscript]