Skip to main content

Table 4 Results of MA and data synthesis

From: Methodological steps used by authors of systematic reviews and meta-analyses of clinical trials: a cross-sectional study

VariableNumber (%)
Know pre/post-intervention data and post-intervention data (Yes), n = 364161 (44.2)
Investigating the efficacy of an intervention, and the pre- and post-intervention (pre/post) values for each group with the correlation value were not available, n = 156
 Used pre/post-difference values for each group77 (49.4)
 Used post values for each group40 (25.6)
 Used (pre/post) values for each group and use correlation value39 (25.0)
If the (pre/post) values for each group are the only data available but the correlation value is not available, what did you do? n = 157
 Request authors of original articles53 (33.8)
 Use a default correlation value of 0.5 if no response from authors28 (17.8)
 Analyze several tests with correlation values ranging from 0.1–0.9 if nonresponse from authors32 (20.4)
 I do not know what is meant by the correlation value44 (28)
Using raw data (such as mean, SD) or analyzed results (such as p-values) in the MA, n = 382
 Use raw data221 (57.9)
 Use analyzed result26 (6.8)
 It is dependent on case by case135 (35.3)
MA of adjusted or non-adjusted data, n = 372
 Unadjusted data alone118 (31.7)
 Adjusted data alone58 (15.6)
 Meta-analyze both ways163 (43.8)
 Select data that have higher level of nominal significance9 (2.4)
 I don’t know adjusted and unadjusted data24 (6.5)
Repeating the same data of placebo in many subgroups when analyzing subgroup based on the concentration of the drug used, n = 350
 Yes75 (21.4)
 No, because the same population cannot be pooled twice in one MA. So the author should only perform subtotal results.275 (78.6)
 Ever extracted Pearson correlation for analysis, n = 37096 (25.9)
 Ever extracted Spearman correlation for analysis, n = 36871 (19.3)
Dealing with Pearson and Spearman correlation in the MA (combining or separating), n = 190
 Analyze each method separately in two MAs127 (66.8)
 Use only Pearson method34 (17.9)
 Combine both in one MA23 (12.1)
 Use only Spearman method6 (3.2)
  1. The data is represented by the number and percentage (%). MA; meta-analysis; standard deviation