Table 6. Results of qualitative investigation into reasons for high and low values of health facility readiness indicators and practice outcomes.

Quantitative Results	Qualitative Themes		Recommendations
	Facilitators	Barriers
Health center readiness and capacity building
Study area health centers did not have adequate supply of injectable gentamicin nor oral amoxicillin and 89.5% did not have functioning equipment at baseline.	Guidelines provided a discrete list of commodities that required minimal inputs from project partners for procurement and distribution to sub-district level stores.	Distribution of drugs and equipment was not instantaneous as it was being integrated into existing supply chains from the sub-district level.	The MOHFW has incorporated plans for training providers under the recent National Newborn Health Program and there is provision in the budget for drugs and equipment.
Government supervision visits to health centers were infrequent during the study period, whereas at least 79% of providers reported attending monthly meetings at the sub-district level.	Monthly meetings served as small group mentoring sessions to discuss problems and develop local solutions.	Onsite government supervision was infrequent reportedly due to human resource constraints at the managerial level.	Monthly meetings provide a regular opportunity for mentoring, which could include skills assessment and correction.
Practice outcomes: Classification & antibiotic treatment
Providers correctly classified 85.1% of infants based on the clinical algorithm. 85.6% of all classification errors were identified in infants presenting with signs of CSI.	Providers reported comprehension of the algorithm and appreciate the job aides as decision-making tools. Providers requested practical demonstrations be integrated in training sessions.	Assessment of a young infant is more complex and time-consuming than other pediatric patients. Some providers expressed confusion around classifying infants with multiple signs of PSBI that overlapped classifications.	Training and supervision should include case scenarios incorporating challenges specific to assessment and classification of young infants and when possible observations of care.
For infants that received antibiotic treatment, we identified errors in 22.9% of the records for antibiotic dosage.	Many providers report prescribing fewer doses of first-line antibiotics closer to the community is a positive change.	New methods for calculating dosage with digital scales and the dosing chart required practice and time to learn. Some providers expressed their preference for using broader spectrum antibiotics at higher doses to treat PSBI.	Record review with antibiotic dosage chart may aid in identification and correction of dosage errors. Future research should examine providers’ assessment of effectiveness of simplified treatment and address drivers of antibiotic misuse in outpatient settings.
Provider performance on the guidelines varied by facility with three facilities contributing 39% of the errors in our study area. Provider errors in classification and antibiotic dosage decreased over the study period.	Providers reported fewer challenges as they gained practice with the guidelines and received feedback in supervision and refresher trainings.	—	Given human resource constraints limiting frequent supervision, targeting poor performing facilities for additional support could reduce the overall error rate. Increased supervision in the beginning of rollout may accelerate the learning curve.