Pre-diagnostic colonoscopies reduce cancer mortality - results from linked population-based data in South Australia

Identifying a common data dictionary across colorectal cancer outcome registries: A mapping exercise to identify opportunities for data dictionary harmonisation

IMPORTANCE

The development of colorectal cancer outcome registries internationally has been organic, with differing datasets, data definitions and infrastructure across registries which has limited data pooling and international comparison. Currently there is no comprehensive data dictionary identified as a standard. This study is part of an international collaboration that aims to identify areas of data capture and usage which may be optimised to improve understanding of colorectal cancer outcomes.

OBJECTIVE

This study aimed to compare and identify commonalities and areas of difference across major colorectal cancer registries. We sought to establish datasets comprising of mutually collected common fields, and a combined comprehensive dataset of all collected fields across major registries to aid in establishing a future colorectal cancer registry database standard.

DESIGN AND METHODS

This mixed qualitative and quantitative study compared data dictionaries from three major colorectal cancer outcome registries: Bowel Cancer Outcomes Registry (BCOR) (Australia and New Zealand), National Bowel Cancer Audit (NBOCA) (United Kingdom) and Dutch ColoRectal Audit (DCRA) (Netherlands). Registries were compared and analysed thematically, and a common dataset and combined comprehensive dataset were developed. These generated datasets were compared to data dictionaries from Sweden (SCRCR), Denmark (DCCG), Argentina (BNCCR-A) and the USA (NAACCR and ACS NSQIP). Fields were assessed against prominent quality indicator metrics from the literature and current case-use.

RESULTS

We developed a combined comprehensive dataset of 225 fields under seven domains: demographic, pre-operative, operative, post-operative, pathology, neoadjuvant therapy, adjuvant therapy, and follow up/recurrence. A common dataset was developed comprising 38 overlapping fields, showing a low degree of mutually collected data, especially in preoperative, post operative and adjuvant therapy domains. The BNCCR-A, SCRCR and DCCG databases all contained a high percentage of common dataset fields. Fields were poorly comparable when viewed form current quality indicator metrics.

CONCLUSION

This study mapped data dictionaries of prominent colorectal cancer registries and highlighted areas of commonality and difference The developed common field dataset provides a foundation for registries to benchmark themselves and work towards harmonisation of data dictionaries. This has the potential to enable meaningful large-scale international outcomes research.

DAta Linkage to Enhance Cancer Care (DaLECC): Protocol of a Large Australian Data Linkage Study

Cancer is a leading cause of global morbidity and mortality, accounting for 250 Disability-Adjusted Life Years and 10 million deaths in 2019. Minimising unwarranted variation and ensuring appropriate cost-effective treatment across primary and tertiary care to improve health outcomes is a key health priority. There are few studies that have used linked data to explore healthcare utilisation prior to diagnosis in addition to post-diagnosis patterns of care. This protocol outlines the aims of the DaLECC project and key methodological features of the linked dataset. The primary aim of this project is to explore predictors of variations in pre- and post-cancer diagnosis care, and to explore the economic and health impact of any variation. The cohort of patients includes all South Australian residents diagnosed with cancer between 2011 and 2020, who were recorded on the South Australian Cancer Registry. These cancer registry records are being linked with state and national healthcare databases to capture health service utilisation and costs for a minimum of one-year prior to diagnosis and to a maximum of 10 years post-diagnosis. Healthcare utilisation includes state databases for inpatient separations and emergency department presentations and national databases for Medicare services and pharmaceuticals. Our results will identify barriers to timely receipt of care, estimate the impact of variations in the use of health care, and provide evidence to support interventions to improve health outcomes to inform national and local decisions to enhance the access and uptake of health care services.

Adjustment of Anti-Hyperglycaemic Agents During Bowel Preparation for Colonoscopy in Patients with Diabetes

OBJECTIVE

Due to the growing diabetes pandemic, the number of colonoscopies performed in patients with diabetes is steadily rising. However, recommendations on adjustments of anti-hyperglycaemic agents (AHG) during bowel preparation for colonoscopy are limited.

METHODS

A total of nine articles were revealed on a PubMed search using the search terms "diabetes" and "colonoscopy", "sigmoidoscopy", "endoscopy", "endoscopic intervention", "endoscopic invasive diagnostics", "endoscopic surgery", or "diabetes care in the hospital" and manual screening of the references of the articles reporting on AHG adjustment during bowel preparation.

RESULTS

Regular glucose measurements and the opportunity to contact the diabetes team were commonly advised. Recommendations also agreed that all oral AHG and short-acting insulin should be omitted when patients are on clear fluids. Recent studies suggest discontinuation of sodium-glucose cotransporter-2 (SGLT2) inhibitors even three days before the colonoscopy. In contrast, recommendations differed regarding adjustment of basal insulin depending on diabetes type and time point in relation to the intervention.

CONCLUSIONS

While discontinuation of oral AHG and short-acting insulin during bowel preparation for colonoscopy is generally accepted, recommendations on the adaptation of basal insulin follow different approaches.

AI and Clinical Decision Making: The Limitations and Risks of Computational Reductionism in Bowel Cancer Screening

Advances in artificial intelligence in healthcare are frequently promoted as ‘solutions’ to improve the accuracy, safety, and quality of clinical decisions, treatments, and care. Despite some diagnostic success, however, AI systems rely on forms of reductive reasoning and computational determinism that embed problematic assumptions about clinical decision-making and clinical practice. Clinician autonomy, experience, and judgement are reduced to inputs and outputs framed as binary or multi-class classification problems benchmarked against a clinician’s capacity to identify or predict disease states. This paper examines this reductive reasoning in AI systems for colorectal cancer (CRC) to highlight their limitations and risks: (1) in AI systems themselves due to inherent biases in (a) retrospective training datasets and (b) embedded assumptions in underlying AI architectures and algorithms; (2) in the problematic and limited evaluations being conducted on AI systems prior to system integration in clinical practice; and (3) in marginalising socio-technical factors in the context-dependent interactions between clinicians, their patients, and the broader health system. The paper argues that to optimise benefits from AI systems and to avoid negative unintended consequences for clinical decision-making and patient care, there is a need for more nuanced and balanced approaches to AI system deployment and evaluation in CRC.