? DESCRIPTION (provided by applicant): Colorectal cancer (CRC) screening is effective at saving lives in the elderly when it is targeted to individuals who are most likely to benefit and avoided in individuals who are most likely to be harmed. In general, at least five years of life expectancy is thought to be needed for an individual to benefit from CRC screening. Current risk classifications using the Charlson Comorbidity Index (CCI) and Elixhauser Classification alone or in conjunction with self-reported functional status measures fall short of an 80% 5-year predicted mortality threshold, a typical benchmark used by physicians for clinical decision making. The long-term goal of this research is to promote effective use of colonoscopy in elderly populations. The primary objective of this application is to develop a novel claims-based model that combines frailty and disease severity indicators with the CCI and Elixhauser conditions to identify a frail, elderly population with a sufficiently high predicted probability of 5-year mortaity that clinicians would be comfortable systematically excluding them from colonoscopy. Aim 1: Develop and validate a predictive model of 5-year mortality that includes demographics, claims-based indicators of frailty and disease severity, and comorbidities included in the Charlson Comorbidity Index (CCI) and Elixhauser classification within a large cohort of North Carolina Medicare beneficiaries older than 65 years of age. Sub-Aim 1a.Compare the predictive ability of the model that includes frailty and disease severity (Comorbidity Plus model) to the model that includes only demographics and the comorbidities included in the CCI and Elixhauser classification. Aim 2: Among individuals undergoing colonoscopy, determine the clinical utility of the Comorbidity Plus model to identify overutilization and characterize patterns of health care use and cause of death among over-utilizers. We will develop our model using claims data from a large cohort of over 900,000 elderly North Carolinians. We will evaluate its clinical utility by identifying a subset of patients who underwent colonoscopy but were unlikely to benefit, given that their predicted probability of 5-year mortality was ?80% at the time of colonoscopy. We expect is research will improve on previous models designed to promote targeted screening in the elderly and result in an algorithm that can be used to identify individuals who are unlikely to benefit from CRC screening. The proposed research has the potential to make a significant contribution by decreasing inefficiencies in colonoscopy use and contributing new knowledge in the area of de-implementation.

R21CA191454

2016-01-05

2018-12-31