marginally statistically significantP = 0.007
The same pattern was evident for the subgroup of reviews with marginally statistically significant heterogeneity (P-value between 0.05 and 0.10, P = 0.007, Fisher's exact test).
The same pattern was evident for the subgroup of reviews with marginally statistically significant heterogeneity (P-value between 0.05 and 0.10, P = 0.007, Fisher's exact test).
Even so, we find it pretty bold that the authors believed in a result with borderline significance (P = 0.05), and only when using the fixed effect model, with so much heterogeneity, and with unexplained discrepancies between the results of the trials.