부산대학교 도서관

Background: Little evidence exists to guide continuation of screening beyond the recommended ages of national guidelines for breast, cervical, and colorectal cancers, although increasing age and comorbidity burden is likely to reduce the screening benefit of lower mortality.Objective: Characterize screening after recommended stopping ages, by age and comorbidities in a large, diverse sample.Design: Serial cross-sectional.Participants: All individuals in the PROSPR-I consortium cohorts from 75 to 89 years of age for breast cancer screening, 66–89 years of age for cervical cancer screening, and 76–89 years of age for colorectal cancer screening from 2011 to 2013. The lower age thresholds were based on the guidelines for each respective cancer type.Main Measures: Proportion of annual screening by cancer type in relation to age and Charlson comorbidity score and median years of screening past guideline age. We estimated the likelihood of screening past the guideline-based age as a function of age and comorbidity using logistic regression.Key Results: The study cohorts included individuals screening for breast (n = 33,475); cervical (n = 459,318); and colorectal (n = 556,356) cancers. In the year following aging out, approximately 30% of the population was screened for breast cancer, 2% of the population was screened for cervical, and almost 5% for colorectal cancer. The median number of years screened past the guideline-based recommendation was 5, 3, and 4 for breast, cervical, and colorectal cancer, respectively. Of those screening  > 10 years past the guideline-based age,15%, 46%, and 25% had  ≥ 3 comorbidities respectively. Colorectal cancer screening had the smallest decline in the likelihood of screening beyond the age-based recommendation.Conclusions: The odds of screening past guideline-based age decreased with comorbidity burden for breast and cervical cancer screening but not for colorectal. These findings suggest the need to evaluate shared decision tools to help patients understand whether screening is appropriate and to generate more evidence in older populations.