Severity assessment.

Objective
To assess the comprehensiveness of the reporting of physical harms in colorectal cancer screening programmes (CRCSPs) in randomised controlled trials (RCTs) and systematic reviews (SRs).
Design
We conducted an overview of reviews, comparing the comprehensiveness of reporting of harms in SRs and RCTs with a recent SR conducted according to the PRISMA-harms extension, identifying 17 types of physical harm potentially resulting from CRCSPs.
Main outcome measures
Proportion of the 17 types of physical harm reported per study (study coverage), across studies (outcome coverage) and the level of harm severity reported in RCTs and SRs.
Results
We identified 24 RCTs and 16 SRs investigating physical harms related to CRCSPs. The median study coverage was 4 and 3 out of the 17 types of harm, varying from 5.9–47.1% and 5.9–52.9% types of physical harm reported in RCTs and SRs, respectively. The median outcome coverage was 4 and 3 across RCTs and SRs, varying from 0–66.7% and 0–87.5% in RCTs and SRs, respectively. Of note, 4 types of harm were not reported in any of the identified SRs. Inconsistent definitions of harm in RCTs and SRs made it difficult to assess which levels of severity of harm that were reported in studies.
Discussion
Poor reporting of harms in RCTs was compounded in SRs. We found poor study and outcome coverage and considerable inconsistencies concerning how physical harms were defined in RCTs and SRs. The inconsistent reporting of harms may result in an underestimation of their magnitude in relation to CRCSPs, raising concerns about our current capacity to evaluate the safety of these programmes. Adequate use of existing guidelines for harm reporting in RCTs and SRs and international consensus on how best to define and measure harms in studies of CRCSPs is warranted.