Validation of the Harefield Cleansing Scale: a tool for the evaluation of bowel cleansing quality in both research and clinical practice

BackgroundVariations in bowel cleansing quality before colonoscopy can cause confounding of results within clinical trials and inappropriate treatment decisions in clinical practice. A new tool—the Harefield Cleaning Scale—has been developed, which addresses the limitations of existing scales.ObjectiveValidation exercise for the new cleansing scale.DesignRetrospective validation study.SettingVarious colonoscopy units in France.PatientsPatients who had a total of 337 colonoscopies recorded.InterventionVideo-recorded colonoscopy.Main Outcome MeasurementsComparisons of 2 scoring systems to assess direct correlation, interrater reliability, internal consistency, and test-retest reliability, based on assessment of video recordings from 337 colonoscopies.ResultsCorrelation analysis for expert scores by using the 2 scales yielded a Spearman correlation coefficient of 0.833. Similarly, the comparison of the segmental sum score revealed a Spearman correlation coefficient of -0.778. Cross-tabulation for successful colon cleansing was 92.88% versus unsuccessful colon cleansing in 7.12%. Reliability assessment indicated an acceptable internal consistency with a Cronbach alpha coefficient of 0.81. Test-retest reliability demonstrated an overall agreement of 0.639 (kappa statistic). Receiver operating characteristic analysis versus Aronchick Scale scores yielded an area under the curve of 0.945, with sensitivity of 99% and specificity of 83% at the optimum score cut-off point.LimitationsTest-retest reliability was assessed by using a different patient population to the other measures. There were insufficient patient numbers to assess performance by using adenoma detection rate.ConclusionThis validation analysis has demonstrated that the Harefield Cleansing Scale is a robust, reliable, and consistent tool that has the potential to improve the effective standardization of bowel preparation assessment in both clinical and research practice.