A randomized controlled trial of an evidencebased toolbox and guide to increase primary care clinicians' rates of colorectal cancer screening in diplomates of the abim

BACKGROUND: Colorectal cancer screening (CRCS) is effective, costeffective and consistently recommended by clinical guidelines, yet only 64.5% of Americans aged 50‐75 years have been screened. Recommendation from a physician is the most influential factor in determining whether a patient is screened for CRC. This study was undertaken to determine whether the Evidence‐Based Toolbox and Guide to Increase Primary Care Clinicians' Rates of CRCS, developed by the Centers for Disease Control and American Cancer Society, could help diplomates of the American Board of Internal Medicine (ABIM) recertifying for the Internal Medicine boards increase their practice rates of CRCS. The main study aim was to compare practice rates of CRCS in the the control (PIM only) and study (PIM + toolbox) arms. METHODS: In this randomized controlled trial diplomates who are primary care providers for patients over age 50 and who were enrolled in ABIM's Maintenance of Certification program were invited to participate. Participants had to enroll in the Cancer Screening Practice Improvement Module (PIM), which requires physician‐directed measurement of their performance and design of a quality improvement plan. Diplomates were randomized to the PIM or the PIM + toolbox arm. The toolbox consists of four essential sections and evidence‐based tools: Your Recommendation, An Office‐based Policy, a Reminder System and an Effective Communication System. Analysis was based on the intention‐to‐treat principle. An external auditor determined the validity of self‐reported CRCS rates in a random sample of 20 % of the participants. RESULTS: A total of 2288 recruitment emails were sent to ABIMdiplomates who met the inclusion/exclusion criteria of whom 160 diplomates expressed interest in participation and 144 enrolled in the study. Of these, 79 diplomates completed the study. The CRCS rates declined in 17 of the 79, but improved in the remaining 62. We conducted an analysis of covariance, a GEE and mixed logistic regression models. All final CRCS were adjusted for the baseline performance and all results showed a non‐significant treatment effect. In the final mixed logistic regression model adjusted for the baseline performance rate and treatment group of the physician and shown in Table 1, we determined the strength of various physician characteristics in predicting the probability of a successful screening that was performed post‐treatment. Medical school country was the only predictor of improved screening rates post‐treatment. CONCLUSIONS: These results demonstrated that while most diplomates improved their CRCS rates from baseline during the study, the improvement was not significantly different in the control arm (PIM only) compared to the study arm (PIM plus toolbox). Medical school country was the only predictor of success rate, with those physicians who attended medical school abroad having an odds ratio of 4.32 (95 % CI: 1.75‐10.65) of performing a successful screening than those physicians who attended medical school in the U.S.