Cost effectiveness of colorectal cancer screening in the Dominican Republic.

Abstract

e23033 Background: Colorectal cancer (CRC) incidence in the Dominican Republic (DR) is increasing, and CRC now ranks as the third-most prevalent cancer in the DR. There are no national CRC screening guidelines in the DR, and CRC screening is not routinely performed. We evaluated the impact of nationwide CRC screening by comparing the cost-effectiveness of four CRC screening strategies to the current practice of no screening. Methods: We developed a Markov model comparing four CRC screening strategies to the current practice of no screening, or natural history (NH): colonoscopy every ten years (Colo); sigmoidoscopy every five years (Sig); fecal immunochemical test biennially (FIT); and fecal occult blood test biennially (FOBT). Model inputs for screening arms were derived from international randomized controlled trials whenever possible. Data from The Rosa Emilia Sánchez Pérez de Tavares National Cancer Institute in the DR was used to estimate baseline CRC incidence, stage distribution and mortality. Screening was implemented from age 45 to 75. We assumed a 60.6% adherence rate for each screening arm. The model was cross validated by comparing the estimated lifetime impact of each screening arm on CRC incidence and mortality to estimates derived from the MIcrosimulation SCreening Analysis (MISCAN) Colorectal Cancer Model in the US. The primary outcome was the incremental cost-effectiveness ratio (ICER) of each screening strategy. Secondary outcomes included life expectancy, total cost, CRC incidence and CRC-related mortality. The willingness to pay threshold (WTP) was set to the 2022 DR gross domestic product per person ($10,111.20/life year gained). Results: In the NH strategy, 2.3% of the population developed CRC, and 1.2% died from CRC, with a lifetime cost of 203.47 per person. Compared to NH, FIT prevented 27% of CRCs and 37% of CRC deaths, with a lifetime cost of 244.76 per person. Colo was more effective than FIT, preventing 33% and 43% of CRC cases and deaths, respectively, compared to NH. However, the ICER for Colo exceeded the WTP threshold. One-way sensitivity analyses showed that the model is most sensitive to test performance characteristics and test costs. Conclusions: Among the evaluated CRC screening strategies, FIT is the cost-effective option, averting a significant number of CRC cases and deaths, at an incremental cost that is well within the WTP threshold for the DR. Inclusion of local data on the effectiveness and acceptability of the tested CRC screening modalities would improve this model, and strengthen the credibility of our findings. [Table: see text]