A scoping review of stroke registers in Sub-Saharan Africa

Stroke registers are recommended as a key priority by the Lancet Neurology World Stroke Organization Commission for Stroke, 2023, and the African Stroke Leaders' Summit, 2022.

This scoping review aims to map where stroke registers have been implemented in Sub-Saharan Africa (SSA). The article then compares and critiques the methods and definitions used and summarizes key results from the registers. The scoping review searched EMBASE, MEDLINE, and CABI Global Health databases and included all studies with a prospective longitudinal design in SSA, where adult acute stroke was the primary condition studied. Articles were screened against inclusion and exclusion criteria independently by two authors.

We identified 42 unique stroke registers from 48 individual studies. The registers were located in 19 countries, with 19 from East Africa, 15 West Africa, 6 Central Africa, and 2 from Southern Africa. Cumulatively, the registers recruited 12,345 participants with stroke, the median number of participants was 183 (interquartile range (IQR): 121-312), and the range was 50-1018. Only one study was a population-based register, and 41 were hospital-based registers. Of the hospital-based registers, 29 were single site, 10 were conducted at two sites, and 2 at three sites. Twenty-three (54.7%) of the registers were located in the capital city of their respective country, and only one of the hospital-based registers was in a self-described rural area. Length of recruitment ranged from 4 months to 6 years; the median length of recruitment was 12 months. Methodology and definitions were heterogenous between the registers. Only seven (19.4%) registers referenced the WHO STEPwise approach to implementing stroke registers. Twenty-seven (64.3%) registers used the WHO definition of stroke. The mean neuroimaging rate was 84%, and ranged from 0% to 100%. Stroke severity was measured using the National Institute of Health Stroke Scale (NIHSS) in 22 (52.4%) registers, four registers used the Glasgow Coma Scale (GCS), two registers used the miniNIHSS, one used the Scandinavian Stroke Scale, one modified Rankin Scale (mRS), and 11 registers did not report a stroke severity measure. Seventeen (40.5%) registers used the mRS to measure function, six registers used Barthel Index alone, and three registers used both mRS and Barthel Index. Only two registers included a quality-of-life measure, the EQ-5D. Eight registers included a quality-of-care measure, and 26 (61.9%) registers recorded socioeconomic status or a socioeconomic status proxy, most frequently educational attainment.

This scoping review found high heterogeneity of methods and definitions used by stroke registers, with low uptake of the WHO stepwise method of stroke surveillance. A drive to standardize methodology would improve the comparability of stroke data in SSA. The shared use of educational attainment by registers in our review may enable future meta-analyses of inequities in stroke in SSA. Incorporating health-related quality-of-life measures, such as EQ-5D, into stroke registers should be encouraged, bringing a patient perspective, and allow the estimation of quality-adjusted life years lost to stroke. Agreement on a standardized register methodology or further promotion and uptake of the WHO stepwise method is essential to produce comparable data to improve stroke prevention and care.