User perceptions of machine learning models as decision support for colorectal cancer multidisciplinary team conferences (AID-SIM-2): a qualitative simulation study.

Clinical Snapshot

Key Findings

• P Value: Not reported — qualitative study

• Effect Size: Not applicable — qualitative study; no effect size reported. Quantitative component found similar distributions of risk groups between clinicians and models overall, with notable discrepancies in individual high-risk patient assessments.

• Primary Outcome: Clinicians expressed overall positive perceptions of prediction models and structured data presentation tools as decision-support in colorectal cancer MDT conferences. The primary perceived benefit was increased standardisation of care, independent of individual physician preferences. Clinicians emphasised the necessity of retaining autonomy to overrule tool suggestions when clinical nuances were not captured by the model.

• Nnt Or Sensitivity: Not applicable — qualitative simulation study. No diagnostic accuracy, NNT, or hazard ratio data reported. Risk-score concordance between clinicians and models is described qualitatively as similar in distribution but divergent for complex, higher-risk individual patients.

• Confidence Interval: Not reported — qualitative study

Clinical Application

The simulation-based evaluation framework described is feasible for prospective assessment of decision-support tools prior to clinical deployment. The four-modality comparison design could be adapted by other cancer centres to evaluate local readiness and user acceptance before committing to full implementation. The finding that clinicians support standardisation but require override capability has direct implications for tool design and governance frameworks. In Australia, colorectal cancer MDT conferences are mandated under the Cancer Australia and state-based cancer care frameworks, and are a core component of RACGP-aligned cancer care pathways. The TGA regulates software as a medical device (SaMD) under the Software as a Medical Device framework, meaning any machine learning prediction tool used in clinical decision-making would require TGA conformity assessment before deployment. The findings of this study are directly relevant to Australian cancer centres considering integration of clinical decision-support systems into MDT workflows. The emphasis on clinician autonomy and standardisation aligns with RACGP principles of shared decision-making and evidence-based practice. PBS implications are indirect but relevant — if decision-support tools influence treatment selection (e.g., adjuvant chemotherapy eligibility), this has downstream implications for PBS-listed oncology agents. Australian cancer networks, including the Victorian Comprehensive Cancer Centre and Cancer Institute NSW, would benefit from simulation-based user acceptance studies modelled on this design prior to local implementation. Clinicians participating in colorectal cancer MDT conferences, including surgeons, medical oncologists, radiation oncologists, radiologists, and pathologists at cancer centres considering implementation of machine learning decision-support tools

Abstract

BACKGROUND: Multidisciplinary team (MDT) conferences are considered a cornerstone of decision-making in cancer diagnostics and care. However, the current literature has not demonstrated improved patient outcomes based on the decisions of the MDT conferences. AIM: We aimed to evaluate how four different decision-support modalities impacted the decision-making process and the internal discussions in the MDTs in a multicenter simulation study, with a focus on user perceptions. METHODS: Four colorectal cancer centers with MDTs participated. We performed four simulations in each center. Each simulation used a different decision-support tool: (1) Current standard, (2) Current standard plus a prediction model, (3) A structured data presentation tool, and (4) A structured data presentation tool plus the prediction model. Clinician- and model-estimated risks were compared, the treatment suggestions from each site were compared, questionnaires about user perceptions were conducted after Simulations 2, 3, and 4 using a Google Form link, and a semi-structured interview was conducted at each site after the last simulation. RESULTS: Similar distributions of risk groups between clinicians and models were found; however, distinct discrepancies in predictions arose, particularly with higher-risk patients, highlighting the need for standardization for more complex clinical cases. The primary perceived benefit of decision support was increased standardization of care, independent of the individual physicians' personal views. However, participants emphasized the necessity of clinician autonomy to overrule tool suggestions when identifying clinical nuances not captured by the model. CONCLUSIONS: The colorectal cancer MDTs expressed a positive view regarding the use of prediction models and other forms of decision-support in their workflow. While clinicians and prediction models had similar risk score distributions, they diverged in the assessment of specific individual patients.

References

1. 1.Bräuner, K. B., Bruun, B., Bertelsen, C. A., Dagenborg, V. J., Safir-Hansen, K., Sanko, R., Gögenur, I., & Konge, L. (2026). User perceptions of machine learning models as decision support for colorectal cancer multidisciplinary team conferences (AID-SIM-2): a qualitative simulation study. Acta Oncologica, advance online publication. https://doi.org/10.2340/1651-226X.2026.45754