In studies of long duration, results may be presented for several periods of follow-up (for example, at 6 months, 1 year and 2 years). Results from more than one time point for each study cannot be combined in a standard meta-analysis without a unit-of-analysis error. Some options are:
· To obtain individual patient data and perform an analysis (such as time-to-event analysis) that uses the whole follow-up for each participant. Alternatively, compute an effect measure for each individual participant which incorporates all time points, such as total number of events, an overall mean, or a trend over time. Occasionally, such analyses are available in published reports;
· To define several different outcomes, based on different periods of follow-up, and to perform separate analyses. For example, time frames might be defined to reflect short-term, medium-term and long-term follow-up;
· To select a single time point and analyse only data at this time for studies in which it is presented. Ideally this should be a clinically important time point. Sometimes it might be chosen to maximize the data available, although authors should be aware of the possibility of reporting biases;
· To select the longest follow-up from each study. This may induce a lack of consistency across studies, giving rise to heterogeneity.