BACKGROUND: Suboptimal colonoscopy quality is related to a higher risk of interval cancer. Aim of our study was to explore the variability in detection rate of neoplasia among different endoscopic centres in an unselected population. METHODS: Consecutive patients referred for colonoscopy in 28 Italian centres were included. Detection rate for polyp, neoplasia and advanced neoplasia was assessed at both the individual and centre level. Inter-centre variability in detection rate of colorectal lesions was explored after adjusting for patient-related factors at multivariate analysis. RESULTS: 3150 patients were included. Median detection rates for polyp, neoplasia and advanced neoplasia were 35%, 26% and 13%. At multivariate analysis, patient-related factors associated with neoplasia detection were age, sex, alcohol and smoking history. After adjusting for these variables, a statistically significant difference in the observed/expected ratio among different centres was observed (Chi-squared test: p<0.01). Among non-patient-related factors, documentation of withdrawal time was associated with neoplasia detection. However, a statistically significant inter-centre variability also remained after adjusting for this variable. CONCLUSIONS: A wide variability was present in the detection rate of neoplasia and advanced neoplasia at the level of endoscopic centres in an unselected population. The adoption of a centre-related neoplasia detection rate could be suggested as a performance indicator.
BACKGROUND: Suboptimal colonoscopy quality is related to a higher risk of interval cancer. Aim of our study was to explore the variability in detection rate of neoplasia among different endoscopic centres in an unselected population. METHODS: Consecutive patients referred for colonoscopy in 28 Italian centres were included. Detection rate for polyp, neoplasia and advanced neoplasia was assessed at both the individual and centre level. Inter-centre variability in detection rate of colorectal lesions was explored after adjusting for patient-related factors at multivariate analysis. RESULTS: 3150 patients were included. Median detection rates for polyp, neoplasia and advanced neoplasia were 35%, 26% and 13%. At multivariate analysis, patient-related factors associated with neoplasia detection were age, sex, alcohol and smoking history. After adjusting for these variables, a statistically significant difference in the observed/expected ratio among different centres was observed (Chi-squared test: p<0.01). Among non-patient-related factors, documentation of withdrawal time was associated with neoplasia detection. However, a statistically significant inter-centre variability also remained after adjusting for this variable. CONCLUSIONS: A wide variability was present in the detection rate of neoplasia and advanced neoplasia at the level of endoscopic centres in an unselected population. The adoption of a centre-related neoplasia detection rate could be suggested as a performance indicator.