Variation of Quality of Life Data Collection Across INTERMACS Sites

• 43% of Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) sites have incomplete health-related quality of life (HRQL) Kansas City Cardiomyopathy Questionnaire data collection before mechanical circulatory support device implantation and 41% have incomplete data collection during follow-up after implant.
• INTERMACS sites vary significantly in HRQL data collection, with a 3–5 times greater odds that statistically similar patients would have incomplete HRQL data collection at 1 site vs another.
• Critically ill and busy patients are less likely to complete HRQL data before implantation and busy coordinators are significantly less likely to collect HRQL data at follow-up.
• Improving the consistency of HRQL data collection can improve the value of INTERMACS in defining the patient-centered benefits of mechanical circulatory support treatment.

BackgroundGiven the importance of patients' health-related quality of life (HRQL) after mechanical circulatory support (MCS) device implantation, the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) collects the Kansas City Cardiomyopathy Questionnaire (KCCQ) before and after MCS. The success of data collection and potential implications of missing data on HRQL analyses are unknown.MethodsWe examined the frequency and reasons for not collecting baseline and 3-month KCCQ data across INTERMACS sites from May 2012 to December 2013. Hierarchical logistic regression was used to examine site variability (median odds ratios [MOR]) for not collecting KCCQ data to demonstrate the degree to which some sites can collect more complete data than others.ResultsAmong 3960 and 3523 patients participating in INTERMACS before and 3 months after left ventricular assist device implantation, the KCCQ was not collected in 43.1% at baseline (range across sites, 0–100%) and 40.9% (range, 0–100%) at follow-up. The most common reasons for incomplete KCCQ data at baseline were that the patient was too sick (13.9%) and consent to participate in research was not obtained (12.4%). Significant variation across sites was observed for missingness because of patient (MOR, 2.8; P < .001) and administrative (MOR, 4.8; P < .001) reasons. The most variable patient reasons were that the patients were too stressed (MOR, 7.2; P < .001) and too busy (MOR, 10.6; P < .001). The most variable administrative reasons were that the coordinator was too busy/forgot (MOR, 7.1; P < .001) and miscellaneous reasons (MOR, 8.7; P < .001). At 3 months, significant variation persisted for both patient (MOR, 2.7; P < .001) and administrative (MOR, 3.5; P < .001) reasons. The most variability across sites was that the patient was too busy (MOR, 6.0, P < .001) and that the coordinator was too busy/forgot (MOR, 5.8; P < .001).ConclusionsSites vary substantially in collecting KCCQ data, and many of these variations seem addressable. Improving the consistency of HRQL data collection can improve the value of INTERMACS in defining the patient-centered benefits of MCS treatment.