Colorectal cancer (CRC) is the second leading cause of cancer mortality in the United States. CRC risk differs appreciably across population subgroups. To reduce this risk, the US Preventive Services Task Force recommends general population screening beginning at age 50. Although several screening methods are available, colonoscopy predominates in the United States. Perceived advantages of colonoscopy include accuracy and thoroughness in detection of target lesions (collectively termed advanced neoplasia). However colonoscopy is also an invasive procedure that carries a non-negligible risk of complications. Colonoscopy is costly and resource-intensive. Its widespread use for primary screening has strained the nation's endoscopy capacity and could divert resources from other important health care needs. Moreover, screening the general population by colonoscopy is inefficient in that most procedures do not detect target lesions. Concern has been raised that this important but limited resource is being misallocated to low-risk patients and not appropriately targeted to higher-risk patients, who stand a greater chance to benefit. Efficiency of screening colonoscopy could be increased and the procedure targeted to those who need it most if patients could be risk-stratified in advance based on their chances of having clinically significant neoplasia detected at colonoscopy. Persons at lower risk could then be screened by other, less intensive methods or could defer colonoscopy until a later age, when risk is higher. Although theoretically promising, effective methods to stratify the general population of screening patients into risk categories do not yet exist. Prevalence of advanced neoplasia is known to differ according to age and gender. Limited evidence suggests that prevalence also differs according to factors such as race/ethnicity, body mass index (BMI), and smoking. However, very few studies have been conducted in large, diverse, and well-characterized US clinical populations or analyzed differences in prevalence of advanced neoplasia across combinations of factors. To address these limitations, this study will evaluate prevalence of advanced neoplasia according to personal characteristics and behaviors using electronic data from a large, current, well-defined, Midwestern primary care population. These factors include age, gender, race, BMI, family history, smoking, diabetes, and non-steroidal anti-inflammatory drug use. It will distinguish subjects undergoing a first screening colonoscopy (Aim 1) from those having a repeat screening colonoscopy following an initial negative colonoscopy (Aim 2). The primary care study setting and diverse patient population will allow ready generalizability to US screening populations. This study will be among the first to characterize differences in advanced neoplasia prevalence across combinations of personal characteristics and behaviors in a large, diverse primary care population. It will inform the current debate as to whether screening the general population can and should be customized to individual patient characteristics and will lay the groundwork for development of future risk-prediction models.
Colorectal cancer is the second leading cause of cancer death in the United States, but several screening methods are available and effective in reducing this risk. Colonoscopy is the most common screening method, and many consider it the best;but it is also expensive, invasive, burdensome to the patient, and it carries some risk o harms. This study sets the stage for development of methods to predict which persons are most likely to benefit most from screening colonoscopy;and which persons are at lower risk, and may wish to choose a less invasive method or delay colonoscopy until older.
|Adams, Kenneth F; Johnson, Eric A; Chubak, Jessica et al. (2015) Development of an Algorithm to Classify Colonoscopy Indication from Coded Health Care Data. EGEMS (Wash DC) 3:1171|