13. Recommendations for future research.

Further research should focus on the following areas
Population	Pregnant women who may be at risk of delivering by caesarean section without a medical indication or need Low‐risk group of women (Robson Groups 1 to 4; Robson 2001) Women with a previous caesarean section (Robson Group 5)
Settings	All areas with high or increasing caesarean section rates All settings where women receive maternity or delivery care (community, home, clinics, hospitals, birth centres)
Study designs	Pragmatic randomised trials or cluster‐randomised trials (involving clusters of practices, hospitals, birth centres, labour units). Where these are not feasible, interrupted time series designs should be used Studies should be sufficiently powered (include adequate sample sizes) for primary and secondary outcomes Include sufficient sample sizes to allow assessment of intervention effect by factors such as parity, socioeconomic status, staffing patterns, practice setting (private versus public), geographical region (urban versus rural), among others. Multisite studies are encouraged to increase sample size and generalisability Studies should be preceded with formative research to define main determinants of caesarean births
Interventions	Multifaceted (rather than single‐component) interventions tailored to local determinants (facilitators) of caesarean section practices are recommended The certainty of evidence for caesarean section rate was low to very low for the following interventions. Further studies are needed to address the uncertainty in the effect of these interventions Educational interventions targeted at women or families Education, birth preparation classes and support programmes Psychoeducation by telephone Prenatal education for husbands of pregnant women Different formats of educational interventions (decision support tools) Interventions targeted at healthcare professionals Audit and feedback using the Robson classification (Robson 2001) Education of public health nurses on childbirth classes (Hemminki 2008). Interventions targeted at healthcare organisations or facilities Insurance reforms equalising physician fees for vaginal and caesarean deliveries Collaborative midwifery‐labourist model of care Although not specifically designed to reduce caesarean births, the following interventions examined in related reviews showed benefits in reducing caesarean births and improving other birth outcomes (further studies are required to confirm observed benefits in areas with high caesarean section rates) Continuous one‐to‐one intrapartum support (by nurse‐midwives, lay companion and doulas) Midwifery care versus other care models (such as obstetric care) We did not identify any eligible studies on the following prespecified interventions (outlined in Table 1); studies evaluating the effects of these interventions are needed. Use of opinion leaders Dissemination of information or advocacy with support or campaigns from local or international opinion leaders (role models, leadership persons, public celebrities) Public dissemination of caesarean section rates Informing the public about caesarean section rates by releasing performance data (e.g. for individual physicians or hospitals) in written or electronic form Financial strategies for healthcare professionals or organisations Pay for performance (target payments) Payment for 24‐hour shifts (not for number of procedures) Additional payment if caesarean section rate during shifts is maintained below a predefined threshold Goal setting for caesarean section rates Setting specific predetermined goal for caesarean rate Policies that limit financial/legal liability in case of litigation of healthcare professionals or organisations (tort reforms) Changing the physical or sensory environment of labour and delivery Adding or altering equipment or layout Place of birth (planned home versus hospital births) Strategies to change the organisational culture Strategies include various components of organisational culture, e.g. shared values, behaviours, norms, traditions, sense‐making, which may shape or contribute, or both, to the overall environment of an organisation
Outcomes	Limited data were available from the included studies on maternal mortality and morbidity, neonatal mortality and morbidity, resource use and costs. Future studies should address these outcomes to aid assessment of the desirable and undesirable effects of unnecessary caesarean sections. Studies should address both short‐term and long‐term maternal and neonatal outcomes.
Methodological considerations	Classification of caesarean section The included studies measured and reported caesarean sections in different ways (overall, elective, emergency, intrapartum). This made synthesis and interpretation of findings across studies difficult. A unified system for classifying and reporting caesarean sections would be useful. Taxonomy of caesarean section interventions Given the broad range of interventions intended to reduce caesarean sections (targeting women, community, public, healthcare professionals, healthcare organisations, facilities and systems), there is a need to develop a comprehensive typology of these interventions. This would aid identification, categorisation, comparison and synthesis in systematic reviews and related research. Reporting interventions Studies should fully describe components of interventions (including standard care) to help implementation and replication. Use of the Template for Intervention Description and Replication (TIDieR) checklist is recommended (Hoffmann 2014).