The effect of Expanding Maternal and Neonatal Survival interventions on improving the coverage of labor monitoring and complication prevention practices in hospitals in Indonesia: A difference‐in‐difference analysis

To assess whether the Expanding Maternal and Neonatal Survival (EMAS) program was associated with improved care provided during hospital‐based childbirth.


| INTRODUCTION
Highfacility-basedchildbirthratesandincreasesinskilledattendance at birth do not preclude negative health outcomes, especially when there are gaps in the quality of care provided. 1,2 The World Health Organization's(WHO)qualityofcareframeworkdefinesqualitycare across multiple domains under two general areas: provision of care andexperienceofcare. 3 Ideally,suchcareprovisionisguidedbybest practicesandisstandardsbased.

| Procedures
Data were collected in each province by a different data collection

| Data management and quality control
Data collected on the tablets were transferred directly to online servers. A field manager conducted quality control checks for data completeness.Themanageralsoreviewedcodingandflaggeditems requiringfollowup.

| Statistical method
As the number of patients enrolled at health facilities differed substantially between the study arms and between the study rounds (Table2), we used inverse probability weighting for adjusting these differencestoreduceselectivitybiasduringtheanalysis.
To assess overall effective management of labor and childbirth practices, we created composite scores of related practices for key assessment areas (i.e. selected labor monitoring practices, resuscitationreadiness,andinfectionprevention),expressedaspercentilesof a perfect score. The composite score for selected labor monitoring incorporated provider partograph use,which included monitoring of frequency and strength of uterine contractions, maternalvital signs, and fetal heart rate. Difference-in-difference (DID) analysis adjusting for hospital type and provincewas conducted using each of the three composite scores to examine changes between EMAS interventionandcomparisonhospitalsatbaselinecomparedwithendline. SignificancewassetatP=0.05.

| Infection prevention practices
The use of infection prevention practices by providers improved by a larger margin and to a higher degree among intervention sites thancomparisonsites.Certainpractices,includingwearingapronsor glovesforpersonalprotectionwerehighandremainedhighinboth arms while other practices such as handwashing or wearing a mask werelowatbaselineinbotharmsbutimprovedonlyintheintervention sites. EMAS intervention sites showed greater improvement in infection prevention measures between baseline and endline comparedwithcomparisonsites (Fig.3).

| LIMITATIONS
The EMAS evaluation study utilized a nonrandomized quasiexperimental design, rather than a randomized control trial design.
Nonrandom selection of facilities introduced the risk of selectivity bias in the quasi-experimental EMAS study, which influenced the analysistechniquesused.
Thesecondroundofdatacollectionwasscheduledjustbeforethe EMAS project ended. In some cases, this was after only 11months of intervention exposure,which potentially limited the ability of the studytomeasurethetrueimpactoftheEMASprogram.
There were more observations at baseline compared to endline This study was able to perform direct clinical observations, providing an accurate assessment of selected aspects of the quality of careprovidedinfacilities.However,thestudydidnotmeasuredirect EMASinterventioninputs(e.g.clinicalmentoring,useofperformance standards,orthefidelityofprojectimplementation),whichmayhave accountedforsomedifferencesinobservedoutcomes.

| CONCLUSIONS
Improved labor monitoring, resuscitation equipment readiness, and infectionpreventionpracticescontributetoenhancedcareforwomen and newborns. Overall, observed outcomes indicate that EMASsupported approaches made a significant difference in improving selectedpracticesduringlaborandchildbirth,whichinturnmaylimit preventabledeathsarisingfrommaternalandnewborncomplications.

ACKNOWLEDGMENTS
We thank the health providers and the women and their newborns whose participation made this study possible. The authors would like to acknowledge Anne Hyre for her support with study designandforhersubstantivecontributions,alongwiththatofSari HusniatiandAliZazri,inoperationalizingthestudy.Additionally,we are grateful for the many contributions of Young Mi Kim, Gayane Yenokyan, and Cynthia Geary in supporting manuscript planning, analysis,andwriting.

CONFLICTS OF INTEREST
Theauthorshavenoconflictsofinteresttodeclare.