Learning before leaping: integration of an adaptive study design process prior to initiation of BetterBirth, a large-scale randomized controlled trial in Uttar Pradesh, India

Abstract

Background: Pragmatic and adaptive trial designs are increasingly used in quality improvement (QI) interventions to provide the strongest evidence for effective implementation and impact prior to broader scale-up. We previously showed that an on-site coaching intervention focused on the World Health Organization Safe Childbirth Checklist (SCC) improved performance of essential birth practices (EBPs) in one facility in Karnataka, India. We report on the process and outcomes of adapting the intervention prior to larger-scale implementation in a randomized controlled trial in Uttar Pradesh (UP), India.

Methods: Initially, we trained a local team of physicians and nurses to coach birth attendants in SCC use at two public facilities for 4-6 weeks. Trained observers evaluated adherence to EBPs before and after coaching. Using mixed methods and a systematic adaptation process, we modified and strengthened the intervention. The modified intervention was implemented in three additional facilities. Pre/post-change in EBP prevalence aggregated across facilities was analyzed.

Results: In the first two facilities, limited improvement was seen in EBPs with the exception of post-partum oxytocin. Checklists were used <25 % of observations. We identified challenges in physicians coaching nurses, need to engage district and facility leadership to address system gaps, and inadequate strategy for motivating SCC uptake. Revisions included change to peer-to-peer coaching (nurse to nurse, physician to physician); strengthened coach training on behavior and system change; adapted strategy for effective leadership engagement; and an explicit motivation strategy to enhance professional pride and effectiveness. These modifications resulted in improvement in multiple EBPs from baseline including taking maternal blood pressure (0 to 16 %), post-partum oxytocin (36 to 97 %), early breastfeeding initiation (3 to 64 %), as well as checklist use (range 32 to 88 %), all p < 0.01. Further adaptations were implemented to increase the effectiveness prior to full trial launch.

Conclusions: The adaptive study design of implementation, evaluation, and feedback drove iterative redesign and successfully developed a SCC-focused coaching intervention that improved EBPs in UP facilities. This work was critical to develop a replicable BetterBirth package tailored to the local context. The multi-center pragmatic trial is underway measuring impact of the BetterBirth program on EBP and maternal-neonatal morbidity and mortality.

Clinical trials identifier: NCT02148952 .

Fig. 1

Change in observed essential birth practices performed by birth attendants following implementation of the BetterBirth Program in the first two facilities in Uttar Pradesh following initial adaptation. Trained observers collected data at four predetermined observation time points (OPs) during the perinatal process (OP1: at admission; OP2: before pushing; OP3: immediate post-delivery; OP4: within 1 h post-delivery). Numbers observed OP1: pre 20, post 33; OP2: pre 23, post 23; OP3: pre 23, post 23; OP4: pre 23, post 23

Fig. 2

Change in observed essential birth practices performed by birth attendants following implementation of the BetterBirth Program in three facilities in Uttar Pradesh after the second adaptation. Trained observers collected data at four predetermined observation time points (OPs) during the perinatal process (OP1: at admission; OP2: before pushing; OP3: immediate post-delivery; OP4: within 1 h post-delivery). Numbers observed per observation point (OP). OP1: pre 624, post 335; OP2: pre 521, post 402; OP3: pre 523, post 403; OP4: pre 522, post 409. *p < 0.001. Rates are adjusted for clustering by site