CRITICAL APPRAISAL OF THE CLINICAL TRIAL: EFFECT OF COLONOSCOPY SCREENING ON RISKS OF COLORECTAL CANCER AND RELATED DEATH

First, we would like to congratulate the authors of this study1 for stimulating reflection on public health policy involving the diagnosis of colorectal cancer (CRC), a topic of extreme relevance, given that it is the third most common cancer in men and the second most common cancer in women. In 2020, there were more than 1.9 million new cases of CRC6.

The benefits of CRC screening were recognized four decades ago when the American Cancer Society started recommending it. The screening was responsible for the decline in CRC incidence observed since the 1980s2.

Randomized studies have shown that screening people at medium risk, that is, those with no family history of CRC, reduces the incidence and mortality resulting from this neoplasm2,3.

In contrast to screening programs for other neoplasms, CRC screening allows the diagnosis of lesions at an early stage and the detection of pre-malignant lesions that, if removed, can prevent cancer4.

The importance of CRC screening is based not only on the possibility of early diagnosis but also on the impact of endoscopic polypectomy, which reduces mortality related to this neoplasm by more than 50%7.

The comments below discuss the methodology used in the study (NordiCC Study) and the negative impact on a diagnostic technique established in several publications due to structural errors added to the article.

Pragmatic (real-life) randomized controlled trials (RCTs) can be considered observational studies (cohorts) in which, although randomization is present, it does not give these trials the character of experimentation (associated with the term randomized clinical trial) – the rigorous individual eligibility criteria are a fundamental part of the methodology. Even when these eligibility criteria are “relaxed,” the analysis in these (pragmatic) trials must necessarily consider prognostic differences between participants, which are essential to avoid confounding and selection bias. In addition, the term pragmatic may be misused as the cohort departs from usual practice because, despite the randomization of participants being performed at the group level, the proposed interventions do not correspond to the conventional care that these patients would receive, as an invitation for colonoscopy, periodic contact, or even the creation of a control group without care (called usual care). Also, the absence of care in the usual care group (comparison) calls into question the classic concept of randomization since there is no control in this group regarding losses or migration (crossover) to colonoscopy.

Intention-to-treat analysis is prohibitive due to the extensive loss of adherence (non-compliers) of participants in the invited (screened) group, decreasing the sample in this group after randomization by more than 50%. The only possible analysis is per protocol. Furthermore, non-complier patients (who did not accept the invitation) and conventional care patients should be analyzed within the same group (not screened) and compared with participants who actually underwent colonoscopy.