A Linked Community and Health Facility Intervention to Improve Newborn Health in Cambodia: The NICCI Stepped-Wedge Cluster-Randomized Controlled Trial

Abstract

Background: Newborn mortality in Cambodia remains high, with sepsis and complications of delayed care-seeking important contributing factors. Intervention study objectives were to improve infection control behavior by staff in health centers; improve referral of sick newborns; increase recognition of danger signs, and prompt care-seeking at an appropriate health facility; and appropriate referral for sick newborns by mothers and families of newborn infants. Methods: The stepped-wedge cluster-randomized controlled trial took place in rural Cambodia from February 2015 to November 2016. Sixteen clusters consisted of public health center catchment areas serving the community. The intervention included health center staff training and home visits to mothers by community health volunteers within 24 h of birth and on days 3 and 7 after delivery, including assessment of newborns for danger signs and counselling mothers. The trial participants included women who had recently delivered a newborn who were visited in their homes in the first week, as well as health center staff and community volunteers who were trained in newborn care. Women in their last trimester of pregnancy greater than 18 years of age were recruited and were blinded to their group assignment. Mothers and caregivers (2494) received counseling on handwashing practices, breastfeeding, newborn danger signs, and prompt, appropriate referral to facilities. Results: Health center staff in the intervention group had increased likelihood of hand washing at recommended key moments when compared with the control group, increased knowledge of danger signs, and higher recall of at least three hygiene messages. Of mother/caregiver participants at 14 days after delivery, women in the intervention group were much more likely to know at least three danger signs and to have received messages on care-seeking compared with controls. Conclusions: The intervention improved factors understood to be associated with newborn survival and health. Well-designed training, followed by regular supervision, enhanced the knowledge and self-reported behavior of health staff and health volunteers, as well as mothers’ own knowledge of newborn danger signs. However, further improvement in newborn care, including care-seeking for illness and handwashing among mothers and families, will require additional involvement from broader stakeholders in the community.