Explanation

For each primary and secondary outcome, the number of participants included in each group is an essential element of the analyses. Although the flow diagram (item 22a) should indicate the numbers of participants included in the analysis of the primary outcome, the number of participants with available data will often vary for different outcomes and at different time points.

Missing data can introduce potential bias through different types of participants being included in each treatment group. It can also reduce, through loss of information, the power to detect a difference between treatment groups if one exists (item 21c) and reduce the generalisability of the trial findings [404]. It is therefore important to report the number of participants with available data for each primary and secondary outcome and at each timepoint. Where possible, it is also important to report the reason data were not available, for example, if the participant did not attend follow-up appointments, or if data were truncated because the participant died [404]. The extent and causes of missing data can vary. For example, a systematic review of palliative care trials estimated that 23% of primary outcome data were not available [462]; this compares to a recent review of trials published in four top general medical journals where the median percentage of participants with a missing outcome was around 9% [392].

Trial results are often more clearly displayed in a table rather than in the text, as shown in table 11 and table 12. For each outcome, results should be reported as a summary of the outcome in each group (eg, the number of participants included in the analysis with or without the event and the denominators, or the mean and standard deviation of measurements), together with the contrast between the groups, known as the effect size. For binary outcomes, the effect size could be the risk ratio (relative risk), odds ratio, or risk difference; for survival time data, it could be the hazard ratio or difference in median survival time; and for continuous data, it is usually the difference in means.

For all outcomes, authors should provide a CI to indicate the precision (uncertainty) of the estimated effect size [364, 463]. A 95% CI is conventional, but occasionally other levels are used. Most journals require or strongly encourage the use of CIs [464]. They are especially valuable in relation to differences that do not meet conventional statistical significance, for which they often indicate that the result does not rule out an important clinical difference. The use of CIs has increased markedly in recent years, although not in all medical specialties [465]. A common error is the presentation of separate CIs for the outcome in each group rather than for the treatment effect [465]. Although P values may be provided in addition to CIs, results should not be reported solely as P values [466, 467]. Results should be reported for all planned primary and secondary outcomes and at each time point, not just for analyses that were statistically significant or thought to be interesting. Selective reporting within studies is a widespread and serious problem [51, 468].

When the primary outcome is binary, both the relative effect (risk ratio (relative risk) or odds ratio), and the absolute effect (risk difference) should be reported (with CIs) (table 13), as neither the relative measure nor the absolute measure alone gives a complete picture of the effect and its implications. Different audiences may prefer either relative or absolute risk, but both clinicians and lay people tend to overestimate the effect when it is presented solely in terms of relative risk [469-471]. The magnitude of the risk difference is less generalisable to other populations than the relative risk since it depends on the baseline risk in the unexposed group, which tends to vary across populations. For diseases where the outcome is common, a relative risk near unity might nonetheless indicate clinically important differences in public health terms. In contrast, a large relative risk when the outcome is rare may not be so important for public health (although it may be important to an individual in a high risk category). For both binary and survival time data, expressing the results also as the number needed to treat for benefit or harm can be helpful [472, 473].