Article Title

Implementing a Multisite, Transdisciplinary Case-Control Study to Assess Effectiveness of Screening Colonoscopy for Preventing Death From Colorectal Cancer (SCOLAR)

Publication Date



colorectal, cancer, screening


Background/Aims: Colorectal cancer (CRC) is the second leading cause of cancer death in the United States. There is a need for observational studies that assess effectiveness of “screening” colonoscopy by carefully defining indication for exam. “Screening for Colorectal Cancer in Average-Risk Adults (SCOLAR),” is a transdisciplinary study to determine if colonoscopy is as effective as believed, filling a critical void in the evidence on effectiveness of CRC screening interventions in reducing the risk of death in real-world practice.

Methods: We describe implementation of SCOLAR, an ongoing case-control study at Kaiser Permanente Northern and Southern California conducted by investigators from seven institutions. All subjects had 5 years minimum prior health plan enrollment, no personal or strong family history of CRC, or diagnosis of inflammatory bowel disease. Cases are subjects who died from invasive colorectal adenocarcinoma (2006–2012) and were 55–90 years old on date of death. Controls were individually matched to cases on service area, sex, birth date and enrollment years. SAS programs were centrally developed and adapted locally for sampling patients and extracting electronic data to populate a standardized data collection tool. Data collection involves chart reviews of electronic medical records and paper charts to identify patient screening history for a minimum of 10 years prior to cancer diagnosis (or reference date for controls). In addition to data quality safeguards, i.e. inter-rater validation or teleconferences among chart abstractors, expert reviewers perform adjudication of indications to differentiate screening from diagnostic procedures, a key methodological strength to assess true CRC prevention.

Results: To date, 4,753 chart abstractions have been performed, of which 48.8% of subjects were women, 35.3% were 55–64 years of age, 9.5% were Hispanic, and 70.9% were white. We found a high level of accuracy using electronic data to capture receipt of colonoscopy. We are currently performing reviews and adjudication of data to create indication. Preliminary results show that 43% had a screening fecal occult blood test within 2 years of the reference date, 39.7% had a flexible sigmoidoscopy, and 32% had at least one colonoscopy.

Discussion: We believe SCOLAR provides a model for overcoming methodological issues affecting case-control studies of the effectiveness of colorectal cancer screening.




March 27th, 2015


April 28th, 2015