Missingness of Quality of Life Assessments Before Left Ventricular Assist Device Implant: A Novel Hospital Quality Metric

Purpose While improved health-related quality of life (HRQOL) is an important durable left ventricular assist device (LVAD) implant outcome, 50% of HRQOL data are missing in the STS-Intermacs registry. Pre-implant HRQOL missingness may reflect patient and hospital characteristics and thus serve as a novel hospital quality metric. We hypothesized that hospital pre-implant HRQOL missingness independently predicts 90-day outcomes. Methods The study cohort included 14,063 patients receiving a primary LVAD (4/2012 - 8/2017) with or without RVAD at 169 hospitals. Patient risk factors of pre-implant HRQOL (using the EQ-5D visual analog scale) missingness were examined by a stepwise logistic model. Direct standardization method was used to calculate hospital adjusted missing rates using a mixed effects logistic model. Hospitals were dichotomized as low/high based on median adjusted missing rates. Cox models, adjusting for patient risk, were used to associate a hospital's pre-implant HRQOL adjusted missing rate with 90-day infection, renal dysfunction and mortality. Results Patients with missing pre-implant EQ-5D (47.1% of patients) were younger (mean age: 56.2 vs. 57.9), more often non-White (36.3% vs. 32.0%), with an Intermacs Profile 1 (25.1% vs. 7.3%), compared to those with complete EQ-5D (all p<.0001). Probability of 90-day composite outcome differed by high and low adjusted missing rate hospitals (p<.0001, Figure). A 10% increase in hospital EQ-5D adjusted missingness was significantly associated with 90-day infection (HR: 1.05; 95%CI: 1.03-1.07), infection mortality (HR: 1.09; 95%CI, 1.01-1.18), renal dysfunction (HR: 1.03; 95%CI: 1.004-1.05) but not all-cause mortality (adjusted HR: 1.01; 95%CI, 0.98-1.04). Conclusion Higher hospital adjusted pre-implant HRQOL missingness was predictive of higher 90-day infection, infection mortality and renal dysfunction. Hospital pre-implant HRQOL missingness may serve as a novel LVAD quality metric. While improved health-related quality of life (HRQOL) is an important durable left ventricular assist device (LVAD) implant outcome, 50% of HRQOL data are missing in the STS-Intermacs registry. Pre-implant HRQOL missingness may reflect patient and hospital characteristics and thus serve as a novel hospital quality metric. We hypothesized that hospital pre-implant HRQOL missingness independently predicts 90-day outcomes. The study cohort included 14,063 patients receiving a primary LVAD (4/2012 - 8/2017) with or without RVAD at 169 hospitals. Patient risk factors of pre-implant HRQOL (using the EQ-5D visual analog scale) missingness were examined by a stepwise logistic model. Direct standardization method was used to calculate hospital adjusted missing rates using a mixed effects logistic model. Hospitals were dichotomized as low/high based on median adjusted missing rates. Cox models, adjusting for patient risk, were used to associate a hospital's pre-implant HRQOL adjusted missing rate with 90-day infection, renal dysfunction and mortality. Patients with missing pre-implant EQ-5D (47.1% of patients) were younger (mean age: 56.2 vs. 57.9), more often non-White (36.3% vs. 32.0%), with an Intermacs Profile 1 (25.1% vs. 7.3%), compared to those with complete EQ-5D (all p<.0001). Probability of 90-day composite outcome differed by high and low adjusted missing rate hospitals (p<.0001, Figure). A 10% increase in hospital EQ-5D adjusted missingness was significantly associated with 90-day infection (HR: 1.05; 95%CI: 1.03-1.07), infection mortality (HR: 1.09; 95%CI, 1.01-1.18), renal dysfunction (HR: 1.03; 95%CI: 1.004-1.05) but not all-cause mortality (adjusted HR: 1.01; 95%CI, 0.98-1.04). Higher hospital adjusted pre-implant HRQOL missingness was predictive of higher 90-day infection, infection mortality and renal dysfunction. Hospital pre-implant HRQOL missingness may serve as a novel LVAD quality metric.