A Pilot Study of Augmented Intelligence Risk-Based Stratification for Endocrine Surgical Waiting List Prioritisation

Introduction The New Deal for Surgery report encouraged using new technology in healthcare to address the 377,689 patients in England awaiting National Health Service (NHS) hospital treatment in July 2022. During the pandemic's second wave, this pilot study investigated the utility of COMPASS Surgical List Triage (COMPASS SLT; C2-Ai, Cambridge, England), an augmented intelligence-based system, in assisting surgical decision-making on patient prioritisation. Data generated from COMPASS SLT was compared to data from the British Association of Endocrine and Thyroid Surgeons' (BAETS) and Federation of Surgical Specialties Associations' (FSSA) prioritisation guidance. Methods A cohort of thyroidectomy and parathyroidectomy patients on the surgical waiting list at Imperial College Healthcare NHS Trust, London, United Kingdom, was used. COMPASS SLT calculated individuals' mortality and significant morbidity risk (risk >2.5%). Significant morbidity risk was set at 2.5% or above following internal model validation, thus reducing the risk of model overfitting occurring with COMPASS SLT. The additional increase in mortality and morbidity due to treatment delay was calculated. Actual treatment time was aligned to the treatment delay (in weeks) experienced by each patient. Results Twenty-nine patients, with a median age of 43 years and a waiting time of 18 weeks at the onset of the second wave, were enrolled. Non-statistically significant differences (p=0.937) between the FSSA and BAETS classifications were identified. However, cohort size could promote a type II error. An increase in median mortality and morbidity risk (p<0.001) arising from treatment delay and decisions based on the FSSA and BAETS classifications were identified. Conclusion COMPASS SLT can supplement clinical decision-making. An augmented intelligence tool can provide clinicians objectivity and flexibility in prioritising patients, with information on individual morbidity and mortality.