Quantifying Uncertainty to Inform Time-Limited Trials of Invasive Mechanical Ventilation

PROJECT SUMMARY/ABSTRACT About 800,000 patients with acute respiratory failure are initiated on invasive mechanical ventilation (IMV) in the United States annually; however; over 1 in 10 patients are subsequently unable to be liberated from IMV and progress to prolonged IMV. Progression to prolonged IMV is viewed by many patients as a state worse than death, but may occur by default (even if inconsistent with the patient’s stated goals) due to a combination of clinical momentum, missed opportunities for shared decision-making, and prognostic uncertainty. Time- limited trials (TLTs) are a promising model of care that directly address each of the barriers to goal-concordant care and can facilitate earlier, more informed shared decision-making. A TLT is a collaborative plan among clinicians and a patient/family to try life-sustaining therapy for a defined duration, after which the patient’s clinical course informs prognostication and guides subsequent decisions. The American Thoracic Society has recently identified major knowledge gaps hindering implementation of TLTs. First, clinicians’ ability to accurately recognize prognostic uncertainty is limited, leading to risk of bias and underuse of TLTs. Second, the optimal duration of TLTs for IMV remains undefined. Suggested TLT durations range widely from 1-7 days, indicating a substantial knowledge gap in setting optimal TLT durations. Third, while clinicians have requested empirical guidance to inform patient selection and duration of TLTs, how clinicians will incorporate this guidance into the complex, dynamic context of critical illness is unknown. We propose to fill these critical knowledge gaps through a novel approach. Using a large, granular database including ~25% of US hospitalizations, we will (1) identify patients with greatest uncertainty in prognostic estimates at the time of IMV initiation to inform TLT patient selection, (2) characterize the temporal profile of uncertainty during the course of IMV to inform a patient’s TLT duration (by defining the point at which prognostic gains plateau with each passing day), and (3) assess the thresholds of uncertainty that affect clinician decision-making. This proposal is in direct response to NHLBI’s Notice of Special Interest in Palliative Care (NOT-HL-23-117), by seeking to facilitate earlier shared decision-making, develop clinical tools that can inform optimal timing and discussions on palliative care, and evaluate the factors influencing adoption and penetration of a TLT clinical tool. Results will synergize with ongoing studies of patient-centered outcomes of TLT and directly inform future clinical trials implementing an improved TLT framework to improve delivery of quality, goal-concordant care to patients with acute respiratory failure.