LAUSR

BACKGROUND AND AIM This study aims to systematically evaluate adherence-to-colonoscopy and related factors in cascade screening of colorectal cancer (CRC) among average-risk populations, which is crucial to achieve the effectiveness of CRC screening. METHODS We searched PubMed, Embase, Web of Science, and Cochrane Library for studies published in English up to October 16, 2020 and reporting the adherence-to-colonoscopy following positive results of initial screening tests. A random-effects meta-analysis was applied to estimate pooled adherence and 95% confidence intervals (CIs). Subgroup analysis and mixed-effects meta-regression analysis were performed to evaluate heterogeneous factors for adherence level. RESULTS A total of 245 observational and 97 experimental studies were included and generated a pooled adherence-to-colonoscopy of 76.6% (95% CI: 74.1-78.9) and 80.4% (77.2-83.1), respectively. The adherence varied substantially by calendar year of screening, continents, CRC incidence, socioeconomic-status, recruitment methods and type of initial screening tests, with the initial tests as the most modifiable heterogeneous factor for adherence across both observational (Q=162.6, P<0.001) and experimental studies (Q=23.2, P<0.001). The adherence-to-colonoscopy was at the highest level when using flexible sigmoidoscopy as an initial test, followed by using guaiac-faecal-occult-blood-test, quantitative or qualitative faecal-immunochemical-test and risk-assessment. The pooled estimate of adherence was positively associated with specificity and positive predictive value (PPV) of initial screening tests, but negatively with sensitivity and positivity rate. CONCLUSIONS Colonoscopy adherence is at a low level and differs by study-level characteristics of programs and populations. Initial screening tests with high specificity or PPV may be followed by a high adherence-to-colonoscopy.