LAUSR

BACKGROUND Active warming reduces risk of surgical complications. Implementation of a perioperative thermal care bundle increased use of active warming for surgical patients. OBJECTIVE This study aimed to determine if implementing a thermal care bundle to prevent inadvertent perioperative hypothermia is cost-effective. DESIGN A model-based cost-effectiveness analysis was undertaken using Monte Carlo simulations from input distributions to estimate costs and effects. SETTING Hospitals undertaking between 5,000 and 40,000 surgeries per year, which either implemented or did not implement the thermal care bundle, were modelled. PARTICIPANTS The decision tree guiding the structure of the model was populated with clinical outcomes (surgical site infection, blood transfusion requirement and morbid cardiac events) of a hypothetical cohort of surgical patients. INTERVENTIONS Implementation or non-implementation of the thermal care bundle. MAIN OUTCOME MEASURES Net monetary benefit was calculated by multiplying the health benefits (quality-adjusted life years) by the willingness-to-pay threshold minus the cost. We tested a range of values for willingness to pay per quality-adjusted life year thresholds and plotted results for expected incremental benefits and probability of cost-effectiveness. The incremental cost-effectiveness ratio was also calculated. RESULTS Thermal care bundle implementation simultaneously reduced costs and increased quality-adjusted life years in the majority of simulations (88.1%). The average cost reduction was $689,659 (95% credible intervals spanned from a $2,718,364 decrease in costs to $379,826 increase in costs) and average difference in quality-adjusted life years was 54 (95% CI = 0.4 less to 176 more). This equated to an incremental cost-effectiveness ratio of $12747 saved per quality-adjusted life year gained. CONCLUSIONS It is likely that increasing use of active warming by implementing the thermal care bundle would generate cost-savings and improve the quality of life for surgical patients. It would be good value for hospitals with similar characteristics to those included in our model to allocate the extra resources required for implementation.