The carbon footprint of end-to-end healthcare deliveries by the National Health Service in England totalled 25.0 megatons of carbon dioxide equivalent (CO₂e) in 2019. Optimal and sustainable healthcare can lead to better health outcomes as well as a lower environmental footprint.

To evaluate the potential impact of prevention and effective management of type 2 diabetes mellitus (T2DM) in adults on both the clinical outcomes and greenhouse gas (GHG) emissions in the UK healthcare setting.

We incorporated an environmental module into the existing IQVIA core diabetes model to estimate the impact of improving clinical outcomes on GHG emissions over a lifetime horizon. We assessed two hypothetical scenarios: (1) preventing progression from pre-diabetes to T2DM through diet and exercise versus no intervention and natural disease progression to T2DM; and (2) well-controlled T2DM using interventions with clinical benefit on glycosylated haemoglobin (HbA1c), and renal and cardiovascular outcomes versus uncontrolled T2DM.

Preventing progression to T2DM led to 6.357 additional undiscounted life years and 67% less kg CO₂e emissions compared with subsequent natural progression to T2DM for a person with pre-diabetes over a lifetime (emissions of 9586 kg CO₂e over 37.115 years vs 28 716 kg CO₂e over 30.758 years, respectively). Well-controlled T2DM led to 1.947 additional undiscounted life years and 21% less kg CO₂e emissions per patient over a lifetime compared with uncontrolled T2DM (emissions of 14 545 kg CO₂e over 22.772 years vs 18 516 kg CO₂e over 20.825 years, respectively). In both scenarios, the GHG emission savings were primarily due to reduced emissions related to avoidance of treating complications of T2DM including cardiovascular, renal and eye diseases.

Effective prevention and management of T2DM through implementation of evidence-based clinical guidelines can improve patient outcomes while reducing the healthcare-related environmental impacts.

Few artificial intelligence (AI) clinical decision support systems (CDSSs) are ever evaluated in practice. Although some signal of clinical effectiveness may be needed to justify AI deployment and testing, such data are typically unavailable in early-stage research. This conundrum is especially relevant in the intensive care unit (ICU), where conditions like sepsis and acute respiratory distress syndrome (ARDS) require high-stakes decisions. Our group developed the AI ventilator assistant (AVA), a novel AI CDSS for patients with sepsis ARDS receiving invasive mechanical ventilation. But the promising results of predictive performance estimates are not sufficient to assess AVA’s clinical safety and appropriateness prior to future evaluation and deployment. Therefore, we propose a Clinician Turing Test as a novel validation approach to determine whether clinicians can distinguish AVA-generated treatment recommendations from those enacted by real human clinicians. If AVA’s recommendations are consistently indistinguishable from those of real clinicians, thereby ‘passing’ this Turing test, this would provide a strong preclinical signal of safety and appropriateness.

This multisite, randomised, electronic, vignette-based Phase 1b study will use a Clinician Turing Test design. We aim to recruit 350 critical care clinicians, including physicians and advanced practice providers from six US hospitals. Participants will review nine clinical vignettes of patients with sepsis and ARDS derived from the Molecular Epidemiology of Severe Sepsis in the ICU cohort and an associated profile of a suggested treatment plan. For each participant–vignette combination, the source of the treatment profile will be randomly assigned (AI-generated by AVA vs the actually enacted treatment from real human clinicians) in a 1:1 allocation. The primary endpoint is the participants’ accuracy in identifying whether a treatment profile was AI-generated or human-generated, assessed using equivalence testing through a mixed-effects logistic regression model with random effects for participants and vignettes. Secondarily, a fitted binary classifier will assess discrimination ability using the C-statistic. Secondary endpoints include clinicians’ perceptions of the safety and appropriateness of the treatment profiles, confidence in distinguishing AI-generated and human-generated recommendations, interest in AI CDSSs for sepsis and ventilator management and the time to complete the survey. This novel Phase 1b design provides preliminary but essential information about an AI CDSS’s clinical appropriateness without the risk or cost of actual deployment, thereby informing decisions about future clinical implementation and evaluation in real clinical environments.

This protocol was approved by the Institutional Review Board of the University of Pennsylvania (Protocol #858201). Results are expected in 2026 and will be submitted for publication in peer-reviewed journals and presented at scientific conferences.

NCT07025096.