Impact of colorectal cancer screening programme on survival and employment in Taiwan: A nationwide analysis of real-world data.

Background: Colorectal cancer (CRC) leads to life loss and a significant economic burden, which could be reduced by CRC screening.

Objective: We assessed the potential savings of lives and employment to evaluate the effectiveness of the Taiwan CRC Screening Programme.

Methods: Through interlinkages among Taiwan Cancer Registry, National Mortality Registry, Taiwan CRC Screening Database, and National Health Insurance claim data, we enroled patients with colorectal adenocarcinoma, aged 50-74 years and diagnosed during 2004-2017, and followed them up to 2018. Life expectancy (LE), lifetime employment duration (LED), loss-of-LE and loss-of-LED were calculated, compared with age-, sex- and calendar year-matched cohorts. Assuming no difference within a specific stage for screen-detected versus non-screen detected CRC and weighting them by different stage distributions, we compared the total loss-of-LE and loss-of-LED.

Results: The cohort enroled 77,169 patients with colorectal adenocarcinoma, which included 31,728 women (mean [SD] age, 62.5 [7.1] years) and 45,441 men (mean [SD] age, 62.8 [6.8] years). The mean loss-of-LE and loss-of-LED in women were 6.0 (95% confidence interval [CI] 5.7-6.3) and 1.0 (95% CI 0.8-1.1) year(s), whereas those in men were 5.1 (95% CI 4.9-5.4) and 1.1 (95% CI 1.0-1.2) years, respectively. Among the cohort, 53,678 cases had the screening information. On average, screening potentially saved 2.9 (95% CI 2.6-3.2) years of life expectancy plus 0.5 (95% CI 0.4-0.6) years of employment per case in women and 2.7 (95% CI 2.5-3.0) years plus 0.6 (95% CI 0.5-0.7) years in men, respectively.

Conclusion: The Taiwan CRC Screening Programme is associated with the savings of lives and employment duration. Future studies are warranted to evaluate the cost-effectiveness of beginning screening at a younger age after accounting for saving employment loss and possibly adjusting lead time bias.