• Journal Article

Personalized medicine for prevention: can risk stratified screening decrease colorectal cancer mortality at an acceptable cost?

Citation

Subramanian, S., Bobashev, G., Morris, R. J., & Hoover, S. (2017). Personalized medicine for prevention: can risk stratified screening decrease colorectal cancer mortality at an acceptable cost? Cancer Causes and Control, 28(4), 299-308. DOI: 10.1007/s10552-017-0864-4

Abstract

Purpose Tailored health care interventions are expected to transform clinical practice. The objective of this study was to develop an innovative model to assess the effectiveness, cost, and harms of risk stratified colorectal cancer screening.

Methods We updated a previously validated microsimulation model consisting of three interlinked components: risk assessment, natural history, and screening/treatment modules. We used data from representative national surveys and the literature to create a synthetic population that mimics the family history and genetic profile of the US population. We applied risk stratification based on published risk assessment tools to triage individuals into five risk categories: high, increased, medium, decreased, and low.

Results On average, the incremental cost of risk stratified screening for colorectal cancer compared to the current approach at 60% and 80% compliance rates is $18,342 and $23,961 per life year gained. The harms in terms of false positives and perforations are consistently lower for personalized scenarios across all compliance rates. False positives are reduced by more than 47.0% and perforations by at least 9.9%. There is considerable uncertainty in the life years gained, but the reduction in harms remains stable under all scenarios.

Conclusion A key finding is that risk stratified screening can reduce harms at all levels of compliance. Therefore, selection of screening scenarios should include comprehensive comparisons of mortality, harms from screening, and cost. This study provides guidance for evaluating risk stratified cancer screening and further research is required to identify optimal implementation approaches in the real-world setting.