Table 3 Key points and recommendations

General

 • No one adjudication approach works best in all clinical trial scenarios. The adjudication approach should be chosen based on the specific trial characteristics, primarily the expected misclassification rates of the site and central assessors, the expected correlation between assessors, and the resource burden associated with different types of assessment.

 • Regardless of the adjudication approach, trial organizers should ensure that only qualified assessors (i.e. those with adequate experience and expertise) are chosen to perform adjudication

 • Assessors should be trained before performing adjudication to ensure that consistent approaches are used, particularly in how to deal with complicated cases.

Should we use site or central assessors?

 • The choice between using site and central assessors should be based on the comparative misclassification rates of the two; whichever type of assessor has the lowest misclassification rates will generally be the better choice (although differences in resource burden between the two approaches should also be considered). When misclassification rates are unknown, as is a common occurrence, a combination approach involving one site assessor and two central assessors may be beneficial.

Should central assessors adjudicate all outcomes, or only suspected events?

 • Using a two-stage approach where central assessors adjudicate suspected events is typically not a good idea. In situations where central assessors have lower misclassification rates than site assessors, bias will be increased, compared with using a central assessors to adjudicate all outcomes. If a two-stage approach is used, it should be modified so that site assessors pass on all outcomes that could potentially be classified as an event (rather than only outcomes they think are events), as this should reduce their false-negative rates, which will improve the performance of the two-stage approach.

How many assessors should we use?

 • Increasing the number of assessors can reduce bias and increase power, but the size of the benefit depends on the correlation between assessors, as well as how many assessors there are to begin with. In practice, many reported ICCs tend to be >0.50, indicating that there will be rarely much benefit to using more than three assessors (although for trials with ICCs <0.50, using more than three assessors may be beneficial). For outcomes with very high ICCs, using one assessor will typically be sufficient (although using three assessors may still offer some protection against the unlucky scenario in which the one chosen assessor has poor classification rates).

Independent vs. group adjudication for central assessors

 • For central assessment with multiple assessors, the choice of using either independent adjudication (where each assessors conducts the adjudication independently) or group adjudication (e.g. an endpoint review committee) should be made based on whichever approach is expected to minimize the misclassification rates. However, if group adjudication leads to a very small reduction in the misclassification rate, but is also accompanied by a large increase in the correlation between assessors, then this approach may actually be detrimental. As for other scenarios, the increase in the resource burden should be considered along with the expected increase in classification rates.