4.5. Are the numbers of participants lost to follow-up implausible?

• The reviewer should consider whether numbers of participants lost to follow-up are plausible. This may require domain knowledge, for example about the plausibility of little or no attrition given the context, condition, follow-up duration, and study protocol.
• The reviewer should also consider the role of incentives to minimise attrition in the study and whether they could explain low rates of attrition.
• It may be useful to consider what level of attrition was anticipated in the sample size calculation reported for the study. For example, if a substantial degree of attrition was anticipated, this may lead to concerns if there was actually little or no attrition in the study and no explanation is provided for this.
• Round, equal numbers of participants lost to follow-up, or numbers lost to follow-up resulting in a perfect match with the planned sample size, may be suggestive of problems, but are unlikely to be sufficient to warrant concerns unless other problematic features are also present.
• The answer to this check should contribute to a domain-level judgement.

Example of check 4.5

In a large multicentre trial of psychotherapy versus usual care for people with long-term depression, participants must attend the trial site every three months over an 18-month period to have their outcomes assessed. The trial manuscript reports that there was no loss to follow-up (all 524 participants at all study sites were retained in the study until the end of follow-up). The reviewer judges this to be very unusual for trials conducted in this population, where attrition rates are typically high even in trials with shorter follow-up durations. The reviewer answers “yes” for this check, and this response contributes to the domain-level judgement.