Comparative effectiveness of adjunct non-pharmacological interventions on maternal and neonatal outcomes in gestational diabetes mellitus patients: A systematic review and network meta-analysis protocol of randomized controlled trials

Background: Gestational diabetes mellitus (GDM) in pregnancy leads to a range of perinatal complications. Although several randomized controlled trials (RCT) have tested the effect of non-pharmacological standard GDM care adjuncts on these outcomes, there is no agglomerated statistical evidence on how their occurrence risk varies across interventions and with placebo. Therefore, a systematic review and network meta-analysis (NMA) protocol is proposed here to address this evidence gap. Materials and methods: A search for above RCTs published in the English language will transpire in PubMed, Embase, and Scopus databases irrespective of date and geographic boundary. The RCTs must test nutritional supplementation, digital intervention, structured exercise program, educational program, counseling service, or a combination of these prenatally in GDM patients. These should report ≥1 of the following outcomes- cesarean section, pre-eclampsia, polyhydramnios, preterm birth, macrosomia, prolonged labor, gestational hypertension, premature rupture of membranes, congenital anomaly, Apgar scores, birth weight, birth length, gestational age at birth, neonatal hypoglycemia, neonatal hyperbilirubinemia, and neonatal Corpulence Index. The risk of bias assessment of the recruited trials will transpire using the Revised Cochrane risk-of-bias tool. Determination of the comparative effectiveness between interventions will occur by the frequentist method NMA for respective outcomes. The categorical and continuous outcomes effect size will get calculated in risk ratio and weighted or standardized mean difference, respectively. For each NMA model, network maps and league tables will show the connections between interventions and effect sizes with their 95% confidence intervals for each intervention pair compared, respectively. The publication bias assessment will occur using comparison-adjusted funnel plots. Best intervention prediction for NMA models with statistically significant intervention effect will happen by determining the surface under the cumulative ranking curve values. Statistical analysis will ensue using Stata software (v16). The statistical significance estimation will happen at p