Improving Newborn Care in Ethiopia: Evaluation of a Quality Improvement Workshop

Introduction: Almost 3 million neonates die annually, making up nearly 50% of under-5 mortality, with 98% of these deaths occurring in low and lower-middle income countries (LMICs). With the goal of reducing neonatal mortality, the American Academy of Pediatrics (AAP) developed Helping Babies Survive (HBS), an educational program to teach evidence-based newborn practices in low-resource settings. While HBS has been widely disseminated in LMICs, barriers to implementation often challenge translation of this knowledge into practice. To address these barriers, the AAP and partners developed a simplified guide, Improving Care of Mothers and Babies (ICMB), to teach quality improvement (QI) methods to facility-based providers in LMICs. Its effectiveness remains untested. Consistent with trends in other LMICs, neonatal mortality in Ethiopia has not improved recently despite a dramatic decrease in under-5 mortality. To improve neonatal mortality in collaboration with the Ethiopian Pediatric Society, we recently completed HBS training in 169 primary and secondary hospitals using a train-the-trainer cascade. A subset of these hospitals was selected for further training using the ICMB guide. A central aim was to determine the gain in knowledge about QI methods resulting from this guide-directed training. Methods: In December 2017, participants from 18 hospitals in Addis Ababa attended a 2-day workshop to learn QI methods using the ICMB guide. Instructors presented key background information on the QI process outlined in the guide, and participants completed practice exercises in small groups. An 18-item multiple-choice questionnaire (MCQ) was given to participants before and after the workshop to assess knowledge acquisition. Course evaluations provided data on learner perceptions and areas for potential program improvements. Results: We had 34 participants complete both pre- and post-workshop MCQs. One of the questions did not perform well, so was dropped from the questionnaire before final analysis of results. The percentage of correct answers on the 17-item MCQ significantly increased from a mean of 63 ± 15 % to 81 ± 14% (p