Logo Lanfrica

<p><span>Estimating The Prevalence of Mycobacterium Tuberculosis in South Africa: A Scoping Review and Meta-Analysis</span></p>

Domaine:

healthcare

Type de record:

paper
Créateur:
AdaMatLauLei
Éditeur:
Elsevier BV
Hôte:
Background. Mycobacterium Tuberculosis (MTB) is the organism responsible for tuberculosis, one of South Africa (SA)'s most pressing public health challenges. Despite its importance in informing public health decisions, the understanding of MTB prevalence in SA remains limited. Objective. To analyse the risk factors associated with MTB in SA, identify high risk subgroups and provide a consolidated national prevalence estimate. Method. A scoping review was conducted, using three electronic databases, to identify studies reporting the prevalence of MTB in SA, as determined by a Tuberculin Skin Test or Interferon-Gamma Release Assay. MTB prevalence estimates were adjusted to reflect differences in the sensitivity and specificity of the tests before fitting a meta-regression model to assess the effect of different risk factors on prevalence. The model estimates were then weighted using census data to derive national and provincial prevalence estimates. Results. From 544 studies screened, we included 38 studies, with the majority (25) focused on the Western Cape. The studies were stratified into 175 independent datapoints representing 58 143 individuals. Transformed median age, province, and race were significant predictors of MTB prevalence. The national MTB prevalence estimate was 38.6% (95% Confidence interval: 27.0%; 49.3%), with substantial age variation (15.2% in 0-14 year olds, 46.3% in 15-49 year olds, and 48.9% in adults age 50 or older). Provincial prevalence estimates varied between 14.6% in Limpopo and 67.4% in Eastern Cape. Conclusion. We found higher MTB prevalence in people living in the Western Cape and Eastern Cape, and people over the age of 20. However, study heterogeneity and the small number of studies outside of the Western Cape together imply wide uncertainty ranges around prevalence estimates.

Similaires