Background: Sub-Saharan Africa (SSA) has the biggest maternal mortality burden, but the region lacks accurate data.
Objective: We reviewed methods historically used to measure maternal mortality in SSA to inform future study methods.
Search strategy: We searched databases: PubMed, Medline, WorldCat and CINHAL, using keywords "maternal mortality", "pregnancy-related death", "reproductive age mortality", "ratio", "rate" and "risk", using Boolean operators "OR" and "AND" to combine the search terms.
Selection criteria: We searched for empirical and analytical studies that: 1) measured maternal mortality levels, 2) in SSA, 3) reporting original results, 4) de-duplicating the studies. We included studies published in English since 1980.
Data collection and analysis: We screened the studies using titles and abstracts, reading full articles of selected studies. We analysed the estimates and strengths, and limitations of the methods.
Main results: We identified 96 studies that used nine methods: demographic surveillance (n=4), health record reviews (n=18), confidential enquiries and maternal death surveillance and response (n=7), prospective cohort (n=9), reproductive age mortality survey (RAMOS) (n=6), sisterhood method (n=35), mixed methods (n=4), and mathematical modelling (n=13).
Conclusion: Sisterhood method studies and RAMOS studies that combined institutional records and community data produced MMRs more comparable with WHO estimates.
Keywords: Sub-Saharan Africa; maternal death; maternal mortality; maternal mortality measurement; pregnancy-related deaths; systematic literature review.