Validity and Responsiveness of the Pediatric Quality of Life Inventory 4.0 Generic Core Scales in the Pediatric Inpatient Setting

Abstract

<bold>Importance</bold>: Validated patient-reported outcome (PRO) measures sensitive to clinical change are needed to evaluate the effectiveness of quality improvement interventions. <bold>Objective</bold>: To evaluate the responsiveness, construct and predictive validity of the Pediatric Quality of Life Inventory (PedsQL<super>TM</super>) 4.0 Generic Core Scales in the pediatric inpatient setting. <bold>Design/Setting/Participants</bold>: Prospective, cohort study. Parents/caregivers of patients 1 month-18 years-old (N = 4633) and patients 13-18 years-old (N = 359) admitted to Seattle Children's Hospital between 10/11 - 12/13. Of eligible participants invited to complete the survey (N = 7184), 65% (N = 4637) completed the PedsQL<super>TM</super> on admission and of these 58% (N = 2694) completed the follow-up survey 2-8 weeks after discharge. <bold>Main Outcome Measures</bold>: Responsiveness was assessed by calculating improvement scores (difference between follow-up and admission scores). Construct validity was examined by comparing mean improvement scores for known-groups differing by medical complexity. Predictive validity was assessed using Poisson regression to examine associations between admission scores, prolonged length of stay (LOS&ge3 days), 30-day readmissions or return emergency department (ED) visits. Similar models examined the association between improvement scores and risk of 30-day readmissions or return ED visits. <bold>Results</bold>: The mean PedsQL<super>TM</super> improvement scores (0-100 scale) were Total: 22.1 (standard deviation (SD) 22.7), Physical: 29.4 (SD 32.4), Psychosocial: 17.1 (SD 21.0). Mean PedsQL<super>TM</super> improvement scores were lower for medically complex patients compared to patients without chronic conditions (Total: 13.7, 95% confidence interval (CI) 11.6 - 15.8 versus 24.05, 95% CI 22.44 - 25.66; p <0.001). A 10-point decrement in PedsQL<super>TM</super> Total admission score below the established community-based mean was associated with an increase in risk of prolonged LOS (Total: +15%, 95% CI 13%, 17%), 30-day readmission (Total: +8%, 95% CI 3%, 14%), and ED return visit (Total: +13%, 95% CI 6%, 20%). A 5-point decrement in PedsQL<super>TM</super> Total improvement score below the study sample mean improvement score was associated with an increase in risk for 30-day readmission or ED return visit (Total: +11%, 95% CI 1%, 22%). <bold>Conclusions</bold>: The PedsQL<super>TM</super> demonstrated responsiveness, construct and predictive validity in a population of hospitalized pediatric patients, and may be a useful PRO measure for hospital-based clinical effectiveness research.