4.3.5 Considerations for the meta-analysis of continuous data

For the effect sizes related to differences in continuous data (WMD, SMD), the data regarding the mean response, the standard deviation, and the number of participants in each group are used. The difference in means is the difference between the mean response in the intervention group and the mean response in the control group. This may be the difference in the means between groups at the final measurement of outcomes, or it may be the difference between the means in their changes from baseline. The simple difference in means is also called the mean difference (MD) or the weighted mean difference (WMD). We will use the term the WMD in this chapter. The WMD is used in meta-analysis of continuous data if all studies included in meta-analyses measured the outcome using the same measurement instrument. For meta-analysis computation the difference in means from each individual study are used. The results are expressed in the natural (clinical) units used for the common measurement instrument. If WMD is used, reviewers should provide explanations regarding the interpretation of the results expressed in units used for the common measurement instrument. The minimum score and the maximum score that are possible on the measurement instrument should be specified together with their interpretation. Also, reviewers should specify what change (difference) is considered significant from a practical or clinical point of view. Reviewers should explain the interpretation of a negative or positive difference. The standardized mean difference (SMD) is a difference in means that is standardized by using information on the variability of data (standard deviation). There are three methods (formulas) that are commonly used for the computation of SMD: Cohen’s d, Hedges’ adjusted g, and Glass’s delta. These three formulae use different standard deviations in their computation. Currently, the JBI SUMARI software offers capabilities for the computation of Cohen’s d. The SMD is used in meta-analysis of continuous data if the studies measured the same outcome but with different measurement instruments. For meta-analysis computation the SMD from each individual study are used. The results are expressed in units of standard deviation. Reviewers should provide explanations regarding the interpretation of the results. In order to facilitate the interpretation of the results it is recommended that reviewer’s convert the results into natural (clinical) units by multiplying the results expressed in units of standard deviation with the standard deviation of the scores from a study on a known measurement instrument. The instrument chosen may be the most commonly used instrument or the instrument which has the best psychometric properties. Reviewers should explain the interpretation of differences and justify what is considered a small or medium or large difference; explanations should be provided for negative or positive differences.