Clinically Important Differences Are Useful for Interpreting Patient-Reported Outcomes Measurement Information System (PROMIS) Measures After Left Ventricular Assist Device (LVAD) Implantation: Findings from the Mechanical Circulatory Support: Measures of Adjustment and Quality of Life (MCS A-QOL) Study

Abstract

<h3>Purpose</h3> Interpreting patient-reported outcomes is enhanced by estimating clinically important differences, i.e., clinically significant score differences likely to have implications for patient treatment or care. <h3>Methods</h3> Adult heart failure patients at 12 U.S. sites enrolled in MCS A-QOL (<i>n</i>=648). Post-LVAD implant they completed 12 PROMIS physical, mental and social health measures. Other measures were: Kansas City Cardiomyopathy Questionnaire (KCCQ-12), EuroQOL (EQ-5D-3L), and NYHA Functional Classification. Clinically important differences were estimated by mean PROMIS score differences between distinct, clinically relevant categories for multiple indicators. Tukey-Kramer method adjusted for multiple comparisons. Cohen's <i>d</i> effect sizes were calculated. Minimally important differences (MIDs) were estimated by significant (adjusted <i>p</i><0.05) differences between adjacent, minimally different categories, e.g., NYHA I vs. II. <h3>Results</h3> Mean age was 58 years and the majority were male (78%), Non-Hispanic White (68%), with dilated cardiomyopathy (55%), long-term implant strategy (57%) and NYHA Class I-II (54%). Large (4-10 points) clinically important differences were detected in nearly all pairwise comparisons for all PROMIS measures (see Table for Fatigue results). MIDs for four PROMIS measures were: Fatigue (3 to 5 points; see Table), Physical Function (2 to 3 points), Ability to Participate in Social Roles and Activities (3 points), and Satisfaction with Social Roles and Activities (3 to 5 points). <h3>Conclusion</h3> Patients with an LVAD, their caregivers and their clinicians should find it useful to be able to interpret the meaning of their PROMIS scores. MIDs are especially useful for power calculations when designing new studies.