2.3.5. Unit of analysis issues
Data from parallel-group studies were included for meta-analysis.
2.3.6. Assessment of heterogeneity
Had a meta-analysis been possible, we could have used the I2 statistic to quantify the inconsistencies among the included studies. An I2 value of >50% was considered indicative of substantial heterogeneity [18]. If heterogeneity was observed, we conducted a subgroup analysis to explore its possible causes.