Study of colorectal cancer resection patterns across the state of Victoria using validated administrative data algorithms.

Abstract

Administrative data provide a unique opportunity to examine whole-of-state colorectal cancer (CRC) data. The purpose of this study was to compare types of CRC resection across Victorian geographical zones, using hospital volume and accredited training-post status. All CRC resections in Victorian public hospitals between 2008 and 2013 were analysed using validated algorithms of administrative data from the Victorian Admitted Episodes Dataset. Hospitals were grouped according to Colorectal Surgical Society of Australia and New Zealand (CSSANZ) training-post status, case-volume (high >200 in 5 years) and remoteness of location. Resection frequency and type were compared. In 44 public hospitals over 6 years, 7596 CRC resections were performed. Patient age, American Society of Anesthesiologists Physical Status Classification System score and tumour stage were similar among groups. CSSANZ accounted for nearly 50% of cases but the lowest percentage of emergencies (16.8%). The ratio of right-sided to left-sided plus rectal resections was greater for low-volume than high-volume centres (56.8% versus 40.4%), while left colon and rectal resections comprised a larger proportion of high-volume workload. High- compared with low-volume favoured ultra-low anterior resections (62% versus 33%) over abdominoperineal resections (38% versus 67%). Work patterns among high-volume hospitals were similar regardless of remoteness or CSSANZ status. This study demonstrated that administrative data can provide granular, clinically relevant information with population-wide coverage. Most public CRC resections in Victoria were performed in metropolitan hospitals. The majority of rectal cancer resections were performed in high-volume metropolitan centres but 15% were performed by low-volume regional hospitals.