Antenatal care trial interventions: a systematic scoping review and taxonomy development of care models

Andrew Symon, Jan Pringle, Soo Downe, Vanora Hundley, Elaine Lee, Fiona Lynn, Alison McFadden, Jenny McNeill, Mary J Renfrew, Mary Ross-Davie, Edwin van Teijlingen, Heather Whitford, Fiona Alderdice

Antenatal care models vary widely around the world, reflecting local contexts, drivers and resources. Randomised controlled trials (RCTs) have tested the impact of multi-component antenatal care interventions on service delivery and outcomes in many countries since the 1980s. Some have applied entirely new schemes, while others have modified existing care delivery approaches. Systematic reviews (SRs) indicate that some specific antenatal interventions are more effective than others; however the causal mechanisms leading to better outcomes are poorly understood, limiting implementation and future research. As a first step in identifying what might be making the difference we conducted a scoping review of interventions tested in RCTs in order to establish a taxonomy of antenatal care models. A protocol-driven systematic search was undertaken of databases for RCTs and SRs reporting antenatal care interventions. Results were unrestricted by time or locality, but limited to English language. Key characteristics of both experimental and control interventions in the included trials were mapped using SPIO (Study design; Population; Intervention; Outcomes) criteria and the intervention and principal outcome measures were described. Commonalities and differences between the components that were being tested in each study were identified by consensus, resulting in a comprehensive description of emergent models for antenatal care interventions. Of 13,050 articles retrieved, we identified 153 eligible articles including 130 RCTs in 34 countries. The interventions tested in these trials varied from the number of visits to the location of care provision, and from the content of care to the professional/lay group providing that care. In most studies neither intervention nor control arm was well described. Our analysis of the identified trials of antenatal care interventions produced the following taxonomy: Universal provision model (for all women irrespective of health state or complications); Restricted 'lower-risk'-based provision model (midwifery-led or reduced/flexible visit approach for healthy women); Augmented provision model (antenatal care as in Universal provision above but augmented by clinical, educational or behavioural intervention); Targeted 'higher-risk'-based provision model (for woman with defined clinical or socio-demographic risk factors). The first category was most commonly tested in low-income countries (i.e. resource-poor settings), particularly in Asia. The other categories were tested around the world. The trials included a range of care providers, including midwives, nurses, doctors, and lay workers. Interventions can be defined and described in many ways. The intended antenatal care population group proved the simplest and most clinically relevant way of distinguishing trials which might otherwise be categorised together. Since our review excluded non-trial interventions, the taxonomy does not represent antenatal care provision worldwide. It offers a stable and reproducible approach to describing the purpose and content of models of antenatal care which have been tested in a trial. It highlights a lack of reported detail of trial interventions and usual care processes. It provides a baseline for future work to examine and test the salient characteristics of the most effective models, and could also help decision-makers and service planners in planning implementation.