Clinical, socioeconomic and health system factors associated with caesarean section among Indonesian National Health Insurance members: a retrospective pooled cross-sectional analysis of administrative claims, 2015–2024

Abstract Background Caesarean section rates in Indonesia have risen substantially alongside the expansion of the National Health Insurance (NHI) programme, while socioeconomic and geographic disparities persist. This study examined clinical, socioeconomic and health system factors associated with caesarean delivery among NHI members using nationwide administrative claims data. Methods We conducted a retrospective pooled cross-sectional analysis of 430,420 hospital delivery claims from the Social Security Administrative Body for Health Sample Data, 2015–2024. Adjusted associations were estimated using survey-weighted cross-classified multilevel logistic regression with random intercepts for women and districts/cities (429,999 episodes; 346,913 women; 494 districts/cities). Antenatal care (ANC) was reconstructed within a pregnancy-specific window but, being unrecorded for 56.5% of records, was not included in the primary model. Four sensitivity analyses restricted to first observed deliveries, added prior observed caesarean, excluded same-month records or added ANC (documented-ANC episodes only). Results The weighted caesarean proportion was 63.9%, rising from 54.6% in 2015 to 68.1% in 2024. Among records with documented ANC, the median number of recorded visits was 2 (interquartile range 1–3). After adjustment, 17.7% of the residual latent variation lay between districts/cities. Fetal distress (adjusted odds ratio [aOR] 61.64; 95% confidence interval [CI] 46.01–82.58), oligohydramnios (aOR 21.82; 95% CI 18.21–26.14) and high-risk pregnancy (aOR 5.84; 95% CI 4.36–7.82) were most strongly associated with caesarean delivery. Self-employed members (aOR 1.96; 95% CI 1.87–2.05 vs. national-budget subsidised members) and Class I care (aOR 1.43; 95% CI 1.35–1.52 vs. Class III) had higher adjusted odds, as did private (aOR 4.70; 95% CI 4.47–4.94), state-owned enterprise and police/army hospitals relative to government hospitals. Adjusted odds rose over time (2024 vs. 2015: aOR 5.97; 95% CI 5.10–6.98). Prior observed caesarean was strongly associated with caesarean delivery (aOR 7.95; 95% CI 7.52–8.42) and halved the temporal trend (2024 vs. 2015: aOR 2.65), but no significant association reversed direction in any sensitivity analysis. Conclusions Within this universal-coverage context, caesarean utilisation was associated with clinical conditions, repeat caesarean and non-clinical factors, including insurance segment and facility ownership, and varied substantially between districts. As observational claims data, the findings describe associations rather than causal effects and cannot establish whether individual procedures were clinically appropriate.