Colorectal cancer screening programs, level of organization, testing strategy, and their impact on participation

Abstract

Early detection of colorectal cancer (CRC) greatly increases the chances of cure. However, uptake of regular CRC screening is relatively low, with an estimated range of 20% to 50% in most OECD countries, which inhibits the overall impact of screening. The key is therefore to introduce effective strategies aiming at increasing screening participation. Several different screening programs exist that vary with the level of organization of the program (i.e. the number of criteria present, related to governance, health workforce qualification, access to services, service delivery provision, information system, and quality insurance) and the testing strategy itself (stool test or colonoscopy). Thanks to the specific setup of Austria, we can study the impact of three different colorectal cancer screening programs that (co-)exist in some provinces, on participation. We use regional and age eligibility variations in the screening programs and employ difference-in-differences and regression discontinuities to identify the causal effects of the programs. The analyses are based on claims data from the largest public health insurance provider in Austria. The results show that all three screening programs increase participation, the effect is the strongest for the population-based screening program characterized by the highest level of organization including the regular sending of a stool test kit a home. Participation increases by 30 percentage points on average among the 40 to 80 years old, and by 20 percentage points at the youngest age cutoff of 40. The impact of the low-organized program where CRC screening is offered within the general health checkup at the GP office is 13 percentage points at age 50 when a stool test is used and 1.3 percentage points when a colonoscopy is used as the first testing tool. The middle-organized program that offers colonoscopy only without an invitation system, but a quality insurance system and the guarantee of a qualified workforce, increases the take-up of a colonoscopy by 1 percentage point, an increase that does not differ from the impact of the low-organized program. However, screening participation increases further when two programs co-exist, therefore combining various program types could be beneficial in reaching as large a share of the population as possible with screening programs. The heterogeneity analysis shows that participation is larger for women. By testing strategy, women are more likely to be screened with a stool test, and men with a colonoscopy as the first testing.