Predictors of variation in colorectal cancer care and outcomes in New South Wales: a population‐based health data linkage study

Abstract

Patient characteristics should be included in risk-adjustment models for comparing outcomes between hospitals and for quantifying hospital variation. Further exploration of the reasons why certain hospitals and patients appear to be at risk of poorer care is needed.

Short‐term outcomes after surgical resection for colorectal cancer in South Australia

“…Advanced stage, older age, emergency admission and comorbidity increased the risk of postoperative deaths following resection for CRC. Thirty‐day postoperative mortality was also higher among non‐metropolitan than metropolitan residents, which contrasts with the findings of Jorgensen et al (Australia) and Helewa et al (Canada) who found no regional disparities in 30‐day mortality, after case‐mix adjustment. International studies suggest higher volume hospitals are better placed to manage complications during surgery .…”

“…Emergency admission was consistently associated with poorer outcomes for other short‐term outcomes which is consistent with other studies . Obstruction, bleeding and perforation are the main causes of emergency admission , making surgical intervention more urgent and more complex.…”

Changing roles of population-based cancer registries in Australia

“…Linked NSW hospital and death data are commonly used to monitor surgical resection rates and associated mortality rates. [36][37][38] The validity of these data needs to be checked when service agencies use them to monitor service performance. This validation process is illustrated with pancreatic cancer data.…”

“…41 Registry data linkage with administrative data enables assessment of patterns of care and other performance indicators for health-system monitoring. [36][37][38][39][40] Administrative data may lack the quality of customised registry data, but they complement the registry data and are adequate for showing broad populationwide patterns of care. 34,[36][37][38] Clinical registries may be used to validate linked administrative data in instances where these data cover common patient groups.…”

“…[36][37][38][39][40] Addition of tumour/node/metastases (TNM) cancer stage and other prognostic indicators to registry databases is important for these analyses and is facilitated by the increased use of structured pathology reporting advocated by the Royal College of Pathologists of Australasia. 41 Registry data linkage with administrative data enables assessment of patterns of care and other performance indicators for health-system monitoring.…”