When a group of studies measures a change in a continuous variable (such as blood-pressure readings, cholesterol levels, or pain scores), the most accessible way to analyze the change over all the studies is to use a weighted mean difference (WMD). “Difference” refers to the difference in outcomes between the intervention and control groups at the end of the study and ideally is measured with the usual units for this outcome (e.g., mm Hg for blood pressure). The studies are weighted according to their sample sizes, and the mean refers to the average of these differences across the included studies.

A WMD can be calculated if the studies use the same units of measurement for a given outcome and if the mean, sample size, and standard deviation for each of the studies are known. Using WMD in a meta-analysis has the advantage of expressing outcomes in measurement units that clinicians can readily understand and is therefore the Cochrane Collaboration's preferred method of reporting outcomes for continuous variables.2