The national caesarean rate in Bangladesh climbed from 19.5% (95% CI 18.1 to 20.9) in 2012-13 to 52.0% (50.7 to 53.2) in 2025. That represents a 32.5 percentage-point increase over the period. The authors analysed data from three UNICEF Multiple Indicator Cluster Survey waves (2012-13, 2019 and 2025) comprising 26,539 recent mothers. They reported absolute and relative changes with confidence intervals and documented the temporal trend using repeated cross-sectional survey data.
The descriptive trajectory was steep. Births moved into facilities in large numbers. Private facilities accounted for a growing share of those facility births. Both trends coincided with rising overall caesarean prevalence, a pattern the investigators explored with decomposition and spatial analyses.
To separate drivers of the national rise, the study team used a Kitagawa decomposition. This split the 32.5-point increase into contributions from changing place of birth and changing caesarean rates within sectors. The decomposition found that increased facility births explained 82.8% (78.9 to 87.7) of the rise. A shift of births toward the private sector explained 12.6% (10.1 to 15.0).
The authors also ran a one-step decomposition that combined effects differently; in that specification place of birth accounted for 94.7% (90.7 to 98.9) of the increase. Those results indicate most of the national rise can be attributed to more women delivering in facilities—especially private facilities—rather than changes in clinical thresholds alone.
Sector rates moved in opposite directions. The private sector caesarean rate increased from 79.1% to 86.1% between 2012-13 and 2025. By contrast, the public sector rate declined from 38.7% to 33.5% across the same interval. The study team examined subgroup patterns: private-sector rates in 2025 were consistently high across wealth strata, with rates of 85% to 87% reported in every wealth group.
A focused comparison looked at first births in private facilities, a subgroup where prior caesarean is uncommon. For private first births, the observed 2025 rate exceeded the 2012-13 model prediction by 4.8 points (0.3 to 10.2). The authors presented this as evidence that private-sector caesarean use rose even after accounting for expected compositional changes.
The investigators mapped caesarean use among facility births across 2,720 clusters in the 2025 survey. They modelled travel time to facilities and tested associations with caesarean use and with the decision to deliver in a facility. Longer travel time correlated with fewer caesareans, but the effect operated mainly by reducing the likelihood of delivering in a facility rather than lowering the procedure rate once in a facility.
When analyses were restricted to private-sector births, travel time showed no clear link to caesarean prevalence (reported prevalence ratio 0.99 per 10 minutes, 0.97 to 1.01). In other words, after a woman reached a private facility, distance-related barriers did not explain variation in caesarean use.
Spatial clustering was apparent. Caesarean use among facility births clustered geographically, and the private share of births in nearby clusters—rather than measured household-level covariates—accounted for much of this clustering. The authors thus point to local supply-side patterns and neighbourhood-level sector mix as strong correlates of procedure frequency.
The source data are survey-based and anonymised; the MICS microdata and 2025 cluster coordinates are available to registered users under MICS terms. The authors noted that the data cannot determine clinical necessity. The study did not report clinical indications for individual caesareans, nor did it provide facility-level clinical practice details that would allow judgement on appropriateness. The 2012-13 report did not describe its ethical review; the 2019 and 2025 surveys reported ethical approvals and verbal informed consent procedures.
Other limitations inherent in repeated cross-sectional survey analysis apply: the design captures population-level shifts in place of birth and reported mode of delivery but cannot establish causal pathways linking policy, provider incentives, or clinical decision-making to the observed changes. The mapping and travel-time models illuminate spatial associations but do not identify facility ownership, clinician behaviour, or non-measured confounders that might drive high caesarean use in specific localities.
The study ends on concrete unresolved questions that the data cannot answer: how many of the additional caesareans were medically indicated, what role provider incentives or referral patterns played, and how facility-level practices differed by sector. Those questions remain for future work that combines administrative, clinical and qualitative data sources.