Out of sight, out of mind? Evidence from cross-sectional surveys on hidden caesarean sections among women with stillbirths in Ghana, 2007 and 2017.

Background: Caesarean section (CS) rates in women experiencing stillbirth have not been studied with nationally representative data. Two Ghana Maternal Health Surveys (GMHS) have captured pregnancy and mode of birth data for all women including those with stillbirths. We compared CS rates between women with live births and stillbirths, and identified socio-economic and pregnancy-related factors associated with CS in stillbirths.

Methods: A population-based cross-sectional study was conducted in a pooled sample of 17 138 women who had given birth within 5 years preceding the 2007 and 2017 GMHS. CS rates were compared between women with stillbirths and very early neonatal deaths (SBVENDs) and women with live births who survived the first day. Bivariate and multivariable logistic regressions explored variables associated with CS. Effect modification of household's wealth and maternal educational level by birth outcome was assessed using multivariable logistic regression with interaction terms.

Results: CS rate in women with SBVEND was 19.3% compared with 9.6% in women with live births who survived the first day (rate ratio 2.2; 95% CI 1.6 to 2.9). In multivariable analysis, attaining middle school compared with no formal education (adjusted OR, aOR 2.8; 95% CI 1.1 to 7.1), having had five or more births compared with nulliparity (aOR 3.7; 95% CI 1.3 to 10.7) and reporting prolonged or obstructed labour (aOR 3.3; 95% CI 1.3 to 8.3) were associated with CS in women with SBVEND. Higher household wealth and educational levels were associated with an increased risk of CS in both study groups, with no statistically significant difference in effect.

Conclusion: Disaggregating CS rates by birth outcome revealed a high rate among women with SBVEND, twice the overall rate compared with live births. Exclusion of these 'hidden' CSs from rate calculations may lead to underestimation of (inter)national CS rates and potentially conceals CS overuse or misuse.