Use of a patient consultation questionnaire and weighted numerical scoring system for the prediction of colorectal cancer and other colorectal pathology in symptomatic patients: A prospective cohort validation study of a Welsh population

Abstract

There is currently no system in widespread use that accurately prioritizes colorectal referrals in symptomatic patients with an acceptable degree of sensitivity and specificity. We have validated a weighted numerical scoring system for the prioritization of such colorectal referrals in an attempt to rectify this, with detection of colorectal cancer (CRC) the primary outcome. We conducted a prospective study of symptomatic patients referred by primary care to the colorectal service in a district general hospital. A computer-generated weighted numerical score (WNS) was derived from the primary symptoms and symptom combinations. Patients underwent colorectal investigations and a final diagnosis was established. Sensitivity, specificity and accuracy of CRC detection as determined by the WNS, Department of Health (DOH) and National Institute for Health and Clinical Excellence guidelines was determined. Primary Care compliance with guidelines was analysed. A definitive diagnosis was established in 3457 patients. One hundred and eighty-six (5.4%) had CRC. The mean score for the cancer patients (76.9, 95%CI 72-81) was significantly higher than that of non-cancer patients (52, 95%CI 52-53) P < 0.001. Receiver Operator Curve analysis demonstrates a high discriminatory power for the Patient Consultation Questionnaire (PCQ) with an area under curve of 0.76. Compliance by primary care with the nationally recommended referral guidelines was poor with only 55% and 58% compliance with DOH and National Institute for Clinical Excellence referral guidelines for suspected cancer respectively. The PCQ and the WNS is an efficient, objective system that allows the accurate prioritization of colorectal referrals with a high sensitivity for cancer and other serious colorectal pathologies.