CARDIOVASCULAR JOURNAL OF AFRICA • Volume 34, No 5, November/December 2023 AFRICA 309 The quality of reporting was assessed for each included study according to the STROBE statement.14 Each item was coded as completely reported, not completely reported or not reported. These assessments (risk of bias and quality of reporting) were performed independently by PC and CL with the help of MAN in the case of disagreement. Statistical analysis Agreement between the two investigators for study eligibility was assessed by the Cohen kappa coefficient: a kappa coefficient = 0.60–0.74 was considered good, and ≥ 0.75 was very good. Data were expressed as mean or percentage when available. Median and interquartile range were calculated. Analyses were performed through script developed in the R software [version 3.5.1 (2018–07–02)]. Results We identified 4 613 references, resulting in 3 447 unique citations, after removing duplicates. A flow diagram summarising the identification and selection process for included studies is presented in Fig. 1. We screened 3 447 studies on the basis of the titles and abstracts leading to the exclusion of 3 399 irrelevant citations. A total of 48 publications were reviewed for eligibility on the basis of full-text reading. Thirty studies10,15–43 met the eligibility criteria. Agreement for full-text eligibility between the two investigators was very good [Cohen kappa coefficient 0.75 (0.55–0.96)]. The study characteristics are presented in Table 2. The 30 studies from 11 countries were published between 2003 and 2018. Among the 11 countries in the review, five countries were low income (Burkina Faso, Democratic Republic of the Congo, Uganda, Ethiopia, Eritrea), five were lower middle income (Côte d’Ivoire, Ghana, Nigeria, Congo, Kenya) and one country was upper middle income (South Africa) (Fig. 2). Most of the studies were cross-sectional studies (21 articles). Three studies were multicentre19,20,40 and none was multinational. Almost half of the studies were conducted in Nigeria (14 articles) (Fig. 3), four were conducted in Ethiopia, two in Côte d’Ivoire and two in South Africa. The other eight studies were conducted in various countries of SSA. Almost all studies unfolded in tertiary-care Screening Records identified through database searching n = 4613 MEDLINE = 1 817 EMBASE = 2 352 Cochrane = 444 Records after duplicates removed n = 3 447 Records screened on titles and abstracts n = 3 447 Records excluded n = 3 399 Full-text articles assessed for eligibility n = 48 Full-text articles excluded, with reasons n = 18 Irrelevant, n = 12 Study population, n = 6 Studies included in qualitative synthesis n = 30 10 956 patients 11 countries Identification Eligibility Included Fig. 1. PRISMA flow diagram summarising the identification and selection process for inclusions in the systematic review. Upper-middleincome countries Lower-middleincome countries Low-income countries Fig. 2. Income level of countries in the studies included in the review. 1 study included 2 studies included 4 studies included 14 studies included Fig. 3. Numbers of studies included by countries in the review.
RkJQdWJsaXNoZXIy NDIzNzc=