Data analysis
Comparisons of postbiotic preparations using direct and indirect evidence for efficacy and safety was to be conducted. In trials where postbiotics were administered at different doses, the main analysis used the results of the higher dose group to avoid the generation of an auto-loop. Sensitivity analysis on the dose-response was conducted to determine if there was a difference in results when used as high- or low-dose.
STATA 14 software was used for data analysis and synthesis, employing the frequentist approach using random effects model incorporated in STATA 14. Heterogeneity of studies was measured quantitatively using the software Review Manager 5.4.
For studies with missing summary statistics, the study authors were contacted. If they were unresponsive, the guide for data collection in section 5.5.8 of the Cochrane Handbook for systematic reviews of interventions was followed.27 Data from figures reported in journal articles28 was electronically extracted using WebPlotDigitizer. Missing means and/or SD were derived from the available data in the studies following the section on data extraction for continuous outcomes in the Cochrane Handbook (Higginset al., 2022), the formula derived by Wan et al., 29 and using MedCalc Comparison of Means Statistical Software (www.medcalc.org/calc/), as appropriate.
Sensitivity analysis was conducted with studies that reported the efficacy outcome using SCORAD. Results were reported as mean difference with 95% CI. This sensitivity analysis was done to facilitate ease of interpretation of results in the clinical setting since SCORAD has an established minimal clinically important difference (MCID) at 8.7 points.
Publication bias was to be assessed using comparison-adjusted funnel plots.