UCSF

OBJECTIVES: To evaluate acceptability and effectiveness of midwives as trainers for NoviGuide, a neonatal clinical decision support system (CDSS). DESIGN: A 20-months, mixed-methods open cohort study. SETTINGS AND PARTICIPANTS: Nurse-midwives at four rural health facilities in eastern Uganda. METHODS: We developed a midwife-led trainer programme and instructed two midwives as NoviGuide Trainers in three 3-hour-long sessions. Trainers trained all nurse-midwives at each site in single 3-hour-long sessions. Using the Kirkpatrick model, we evaluated acceptability at level 1 for participants reaction and level 3 for participants attitudes towards the programme. We evaluated effectiveness at level 2 for newly learnt skills, and level 3 for participants uptake of NoviGuide and perception of newborn care practices. We used surveys and focus groups at baseline, 3 months and 6 months and viewed usage data from September 2020 through May 2022. RESULTS: All 49 participants were female, 23 (46.9%) owned smartphones, 12 (24.5%) accessed the internet daily and 17 (34.7%) were present by study end following staff changes. All participants perceived the use of midwives as NoviGuide Trainers to be an acceptable approach to introduce NoviGuide (mean 5.9 out of 6, SD 0.37). Participants reported gaining new skills and confidence to use NoviGuide; some, in turn, trained others. Participants reported improvement in newborn care. Uptake of NoviGuide was high. Of 49 trained participants, 48 (98%) used NoviGuide. A total of 4045 assessments of newborns were made. Of these, 13.8% (558/4045) were preterm, 17.5% (709/4045) weighed under 2.5 kg and 21.1% (855/4045) had a temperature <36.5°C. CONCLUSION: This midwife-led programme was acceptable and led to self-reported improvement in newborn care and high uptake of NoviGuide among nurse-midwives. Task shifting CDSS expert roles to midwives could facilitate large-scale implementation. However, resources like internet coverage, reliable electricity and mobile devices should be considered in low-resource settings.