An Electronic Clinical Decision-Making Tool for Patients with Suspected Colorectal Cancer-Preliminary Evaluation in Patients Presenting with Rectal Bleeding.

BACKGROUND AND OBJECTIVES

The electronic clinical decision-making application was designed to assist physicians with evaluation of patients with suspected colorectal cancer (CRC). The physician completes an interactive checklist of evidence-based clinical parameters, and a recommended referral urgency is generated based on the post-test probability of CRC. This study aimed toward validation of the tool in symptomatic patients presenting with rectal bleeding.

METHODS

The medical records of a sample of patients with histologically confirmed CRC from 2010 to 2014 were reviewed. The tool was applied retrospectively to all patients who initially presented with rectal bleeding, to determine its sensitivity for detecting CRC in this population. A generated recommendation of 'immediate referral' (referral ≤24 hours, expected endoscopy ≤2 weeks) or 'urgent referral' (expected consultation and endoscopy ≤4 and ≤8 weeks) was considered a positive test result. An a priori sensitivity of 90% was deemed adequate, based on test characteristics of the tool's individual clinical criteria.

RESULTS

The tool was applied to 281 patients. A total of 69 (24.6%) and 211 (75.1%) patients met criteria for immediate and urgent referral, respectively. The remaining patient (0.4%) met criteria for 'possible priority referral', while none met criteria for 'no specific action recommended'. This resulted in a calculated sensitivity of 99.6% (95% confidence interval 98.0 to 99.9%).

CONCLUSIONS

The tool is sensitive in the prediction of CRC in patients presenting with rectal bleeding. A prospective cohort study is being designed to allow for acquisition of comprehensive test performance characteristics and full validation of the instrument.