2.5. Data Synthesis and Statistical Analysis

The differences in treatment costs were analyzed by WMD (95% CI) using STATA 15.1. The heterogeneity was analyzed by I². When I² ≤ 50% and p ≥ 0.1, homogeneity is identified, and the analysis was performed using a fixed effects model. I² > 50% and p < 0.1 suggested heterogeneous studies which were further analyzed by sensitivity analysis. The random effects model was chosen when the heterogenicity could not be decreased; otherwise, the combination of results was discarded and descriptive analysis was used. The publication bias was analyzed by the funnel plot.