WHY DO PRIMARY CARE PHYSICIANS FAVOR DIFFERENT COLORECTAL CANCER SCREENING STRATEGIES?

Purpose: Implementation of current colorectal cancer screening involves a choice among several recommended screening options. We evaluated multi-criteria analyses performed by 41 primary care physicians to determine why the results favored different screening strategies.
Methods: After a brief review of current guidelines, the physicians completed a multi-criteria analysis comparing recommended colorectal cancer screening options using the Analytic Hierarchy Process (AHP).  The criteria included effectiveness in preventing cancer, risk of serious screening-related side effects (perforation and major bleeding), false positive rates, screening frequency, screening test preparation, and the nature of the screening test. The analysis assumed regular screening through age 80 using the same strategy. Age-specific data or descriptions for each currently recommended screening option were presented in a blinded fashion except for the descriptions of the procedures. We assessed the reliability of the AHP results using the consistency ratio, which measures the internal consistency of judgments made during the analysis.
Results: The mean age of the study physicians was 44.7 years; 63% were male. Colonoscopy (CS) every 10 years was the AHP-best strategy for 32 (78%) physicians, and flexible sigmoidoscopy every 5 years (FLEXSIG) was the AHP-best strategy for 7 (17%). Two physicians were excluded from further analysis due to a tie between these 2 strategies. There were no differences between the groups in the priorities assigned to the decision criteria or consistency of judgments. They differed in their interpretations of the clinical significance of variations in test performance and characteristics. Physicians in the FLEXSIG group favored shorter versus longer screening frequencies and assigned lower scores to screening every 10 years (0.12 vs. 0.43, p < 0.001) and higher scores to annual screening (0.32 vs 0.14, p = 0.011). Despite using the same outcome estimate, they also judged flexible sigmoidoscopy relatively more effective in preventing cancer: 0.03 vs 0.019, p = 0.008. There were also statistically significant differences between the groups, all with p values < 0.001, in their judgments of several other screening options regarding effectiveness in cancer prevention, risk of serious screening side effects, and risk of false positives.
Conclusion: Physicians' interpretations of clinical information can vary enough to affect decisions regarding colorectal cancer screening. If confirmed, this finding suggests that subjective clinical judgments, as well as decision priorities, can account for clinical practice variations.