The increasing numbers of gastrointestinal cancers have contributed significantly to cancer related morbidity and mortality in Australia. Appropriate and timely utilization of surgical treatment improves the survival of all gastrointestinal cancers. A benchmark rate for optimum surgical utilization at population level has not been derived before in Australia or elsewhere. Therefore, it is unknown whether the population level utilization of cancer directed surgery in Australia is appropriate. This study developed models to estimate the optimum rates of curative and palliative surgical utilization for all major gastrointestinal cancers in Australia and estimated the different levels of complexity of surgical intervention required at a population level.
To develop the models of optimum surgical utilization, latest clinical guidelines were examined, and decision trees were created using the Tree Age Pro software to map treatment pathways for individual cancers. Epidemiological data from Australia and other settings were used in the trees to predict the optimum surgical utilization rate for each cancer. Univariate sensitivity analyses around input parameters were conducted to demonstrate the uncertainty over the optimum rate predicted. The optimum rates were then compared with the actual rates of utilization.
The benchmark optimum surgical utilization rates for oesophagus (42%), stomach (59%), colon (86%), rectum (89%), pancreas (41%) and liver (29%) cancers were estimated. The published actual utilization rate for cancer surgery in oesophagus (29%) and liver (22%) was considerably lower than the predicted benchmark, while for stomach cancer (90%) surgery was over utilized. Colon (95.6%), rectum (93%) and pancreas (39%) cancer surgery utilization pattern was only marginally different from the benchmark.
These model outputs could be used to establish benchmarks for surgery provision and reorientate the health services and investment to optimize cancer surgical utilization in Australia and internationally.