94 Agreement Between Proxy- and Self-Report PROMIS Scores in Pediatric Burn Survivors

Abstract

Abstract Introduction Reporting by proxy is necessary when patients are not able to report their own experience, such as young children or those too sick to respond. PROMIS pediatric proxy-report item banks are a set of generic measures that facilitate comparisons across populations and studies. Our objectives were to (1) examine agreement in pediatric burn survivors between child self- and caregiver proxy-report on multiple PROMIS domains and (2) examine factors associated with differences between self- and proxy-reports. Methods Data were collected from children 8–17 years with moderate to severe burn injury and their respective caregivers between 6 months and 15 years after injury. The PROMIS-25 and Anger v1.0 short form were completed by pediatric burn participants. Caregivers completed either custom (depression, pain interference) or standard (Physical Function (PF) 8a, Peer 7a, Anger 5a) PROMIS proxy short forms. Self- and proxy-report scores were compared using Wilcoxon sign rank test, Cohen’s effect size, and intraclass correlation coefficients (ICC(2,1)) and by agreement across severity of symptoms based on recommended cutoffs. Regression analyses examined child- (self-report score, age, gender, and ethnicity) and proxy-related (relationship to child) factors associated with score differences. Results A total of 274 child-caregiver pairs completed the PROMIS measures. Mean child age was 13.0 (SD:3) years. Caregivers reported significantly worse scores than the child on PF, pain, and anger (all p≤0.01). Cohen’s d ranged from 0.05 (depression) to 0.25 (PF), with all domains except PF in the small effect size range (< 0.2). Similarly, ICCs were all of moderate agreement and ranged from 0.51 (pain) to 0.69 (depression). The percentage of dyads in agreement by severity groups was high with only 5% (pain), 8% (PF, peer relationships), and 9% (depression, anger) of pairs discordant. Regression models indicated only higher self-report score was associated (all p< 0.05) with greater differences across all domains, though female gender was also associated with greater differences on depression only. Conclusions This study provides support for the use of pediatric proxy PROMIS depression, PF, peer relationships, pain interference, and anger scales in pediatric burn patients. Although agreement was moderate to good, proxy report should only be considered when self-report is not possible or practical. Caregivers typically report slightly worse severity of symptoms than children across all domains.