Colorectal Cancer resection outcomes using administrative data

Abstract

Background Colorectal cancer is the second leading cause of cancer death in Australia. If found early, a patient can undergo potentially curative surgery. Even in the third of patients that do recur, there can be palliation or even cure with further surgery with or without chemotherapy. Although surgery remains a vital tool in treating colorectal cancer, it is not without complications.

The principal studies of this thesis used administrative data to report on short- and long-term outcomes following resection for colorectal cancer in the state of Victoria over a ten-year period.

Methods Administrative data are collected on all patients admitted to Australian Hospitals. Patient demographics, co-morbidities, type of operation, post-operative complications, histopathology and some staging information are recorded. Trained coders review clinical notes and then assign alphanumeric codes to these data based on the International Classification of Diseases Tenth Revision, Australian Modification (ICD-10-AM). These codes were developed for the purpose of billing and therefore may not be focused on reporting data in a clinically relevant fashion. We have previously shown that the use of algorithms of code combinations can increase the accuracy of this data source for clinical research1. This thesis added laparoscopic detail to these coding algorithms. These algorithms were then applied to a central repository of administrative data in Victoria, to report on short-term outcomes following resection for colorectal cancer comparing regional to metropolitan hospitals. Results were adjusted for potential confounding variables using a multivariable logistic regression analysis. This data source was then linked to death data to report on overall long-term survival following colorectal cancer surgery, comparing regional to metropolitan hospitals. Survival results were presented as a rate, adjusted for potential confounders using a multivariable Cox regression analysis.

Results These studies found strong evidence for lower odds of prolonged length of stay (OR 0.53, 95% CI 0.48 – 0.58, p=<0.001) and inpatient mortality (OR 0.67, 95%CI 0.49 – 0.91, p=0.01) in inner regional hospital compared with metropolitan hospitals. For outer regional hospitals, there was strong evidence of decreased odds of prolonged LOS (OR 0.64, 95%CI 0.52 – 0.77, p=<0.001) and return to theatre (OR 0.67, 95%CI 0.47 – 0.95, p=0.03). There was no difference in overall survival comparing colorectal cancer resection patients from inner or outer regional hospitals to metropolitan ((HR 1.02, 95%CI 0.95 – 1.09, p=0.59) and (HR 0.97, 95%CI 0.85 – 1.11, p=0.68) respectively).

Conclusion These studies demonstrated the strength of administrative data with validated algorithms and data linkage in reporting on outcomes following colorectal cancer resection. This methodology resulted to two of the largest and most detailed studies concerning colorectal cancer resection in Australia. Importantly, they validated current practices in Victoria by revealing similar outcomes in regional and metropolitan centres after resection for colorectal cancer.