Do Patient-Answered Versus Caregiver-Answered Patient-Reported Outcomes Differ in Pediatric Upper Extremity Fracture Care?

INTRODUCTION

Patient-reported outcomes in children can be assessed using self-report or caregiver proxy-report; however, self-reported and proxy-reported outcomes often vary between respondents, a phenomenon called cross-informant variance. This phenomenon has not been studied in pediatric fracture care. This study compares self-reported and proxy-reported patient-reported outcomes in children being treated for upper extremity fractures.

METHODS

Children aged 8 to 17 with upper extremity fractures completed pediatric self-report questionnaires and their caregivers completed parent proxy-report questionnaires composed of items from three Patient-Reported Outcomes Measurement Information System (PROMIS) domains (Physical Function-Upper Extremity, Psychological Stress Experiences, and Pain-Interference). Mean patient and caregiver T-scores for each domain were compared using paired T-tests. Correlation between scores was assessed using scatterplots and Pearson correlation coefficient. Agreement between child and caregiver T-scores was assessed using Bland-Altman plots.

RESULTS

One hundred child-caregiver dyads were included. Patients' mean age was 12.3 years, and 68% were male. Statistically significant differences were found between mean patient and caregiver T-scores in all PROMIS domains. Caregivers overestimated patient pain and psychological stress and underestimated upper extremity function. However, only the disagreement in the Pain-Interference domain met the threshold of clinical significance, a difference of three or more T-score points. Bland-Altman analysis revealed proportional bias in the Psychological Stress Experiences and Pain-Interference domains. With higher T-scores, caregivers tended to overestimate psychological stress to a greater extent but tended toward agreement with their children for pain.

DISCUSSION

This study identified cross-informant variance between children and caretakers in the setting of pediatric upper extremity fractures. Although notable differences were observed in all PROMIS domains included, only one met the level of clinical significance, suggesting that not all domains are equally susceptible to child-caretaker disagreement. As patient-reported outcomes become more used in pediatric settings, the possibility of cross-informant variance must be considered when choosing to use self-report or proxy-report instruments.

LEVEL OF EVIDENCE

Level III.