A meta-analysis of depressive symptoms among Ethiopian prisoners and a narrative description of its associated factors: a country based systematic review and meta-analysis study.

Background: The burden of depression in prisoners is increasing and factors such as co-existence of medical illness, lack of social support and longer duration of sentences are contributing to it. However, no pooled evidence on the magnitude and factors of depression in prisoners existed in Ethiopia. The current meta-analysis was therefore aimed to have aggregate evidence on the magnitude and factors of depression in prisoners of Ethiopia.

Methods: A search of databases on PubMed, Scopus, and EMBASE was carried out systematically. Besides, grey literature sources were extensively investigated. Moreover, the reference lists of the articles selected were searched. Random effects and quality-effects models were used to describe the pooled prevalence of depressive symptoms with 95% CI. We also detect heterogeneity between studies using Cochran's Q- statistic and the Higgs I test. A sensitivity analysis was also implemented. Publication bias was checked with Egger's test and funnel plots visually.

Results: Among 232 papers identified through the specified database searches only 17 full-text articles were assessed for eligibility and only nine (9) studies fulfilled the prespecified criteria and incorporated in the final meta-analysis. The pooled prevalence of Depressive symptoms among prisoners was 53.40%(95% CI: 41.33, 65.46). The pooled prevalence of Depressive symptoms in prisoners was 41.9% in Southwest Ethiopia, 44.43% in North West of Ethiopia, 59.05% in Addis Ababa, and 72.7% Southern Ethiopia. Besides, the pooled Depression symptoms prevalence among prisoners was 51.24% as measured with PHQ-9 and 56.15% with BDI-II. Besides, studies that utilized a relatively large sample size (≥350) yields a smaller pooled prevalence of Depression symptoms, 51.93% than those which utilized smaller sample sizes (< 350); 54.13%.

Conclusion: The pooled magnitude of depression in prisoner's population is very high, 53.40%. This pooled effect size for the Depression symptoms was significantly higher in the southern region of the country than in the southwest region. Besides, the pooled prevalence was significantly higher as measured by the BDI-II tool than by PHQ-9. Also, studies that utilized a larger sample size provided a significantly lower pooled magnitude of symptoms of depression than studies that utilized a smaller sample size.