The Influence of Between and Within-group Variance on Recurrent Miscarriage in Ghana

Abstract Background : Recurrent miscarriage is a spontaneous loss of two or more pregnancies. It is a major global reproductive public health problem and the least researched in Ghana. There is an urgent need to move from more individualized and genetic risk factors that may be culturally inappropriate to assess and explore the influence of socio-ecological or contextual factors on recurrent miscarriage in Ghana. This study explored the between and within-group variance of zonal, regional, cluster, and household as well as the influence of selected socio-demographic characteristics on recurrent miscarriage. Methods: We conducted a secondary analysis of the 2017 cross-sectional demographic health survey data (25,062 women) in their reproductive ages (15-49 years) in Ghana. Data were collected from zones (3), regions (10), clusters (900), and households (27000). A logistic regression: Mixed effect model was used. Random effects were used to estimate the between and within groups differences in recurrent miscarriage. The fixed effects of selected socio-demographic variables were also assessed. Results: We found that cluster differences in recurrent miscarriage risk were more prominent compared to zonal, regional, and household. There was higher variability in recurrent miscarriage among women in the same household, cluster, and region compared to those from different households in the same cluster or those from different clusters in the same region and household. In the fixed-effects model, the odds of recurrent miscarriage were 1.54 times (95% CI=1.16-2.03) higher among urban dwellers than rural dwellers. Increasing maternal age was significantly associated with recurrent miscarriage (p=0.0001). There was a direction association between educational level and recurrent miscarriage. Insignificant protective factors of partner support and access to first-trimester antenatal care visits were also found. Conclusion : Differences in recurrent miscarriage risk exist at the contextual level and should be the focus of public health prevention efforts. Our key finding highlights the need to contextualize and localize recurrent miscarriage risk reduction efforts if interventions to reduce recurrent miscarriage in women with a history of unexplained pregnancy loss are to be successful in Ghana.