Psychometric evaluation of Patient-Reported Outcome Measurement Information System (PROMIS) measures in children and adolescents with lower limb orthopaedic conditions : construct validity, reliability, and responsiveness.

Abstract

The Patient-Reported Outcomes Measurement Information System (PROMIS) has not been broadly validated for use in paediatric patients with lower limb musculoskeletal conditions. The aim of this study was to evaluate the psychometric properties of PROMIS administered using computerized adaptive testing. This was a single-centre prospective study involving consecutive patients, aged between 11 and 17 years, with orthopaedic conditions affecting the lower limbs. The convergent and divergent construct validity of PROMIS domains with corresponding subscales of the Paediatric Outcomes Data Collection Instrument (PODCI) were assessed in Group A. The reliability was assessed by retesting a sub-group (Group B) of patients from Group A with the same PROMIS domains in those whose condition was not expected to vary within a month of their initial assessment. PROMIS domain responses preoperatively, and > six months postoperatively, were assessed in Group C. The construct validity was assessed in 280 patients in Group A, and convergent validity was achieved between the PROMIS and PODCI Mobility domains (<i>r</i> = 0.73; p < 0.001), Upper Extremity domains (<i>r</i> = 0.52; p < 0.001), PROMIS Physical Activity domain and PODCI Sports and Physical Function subscale (<i>r</i> = 0.57; p < 0.001), PROMIS Depressive Symptoms domain and PODCI Happiness subscale (<i>r</i> = -0.54; p < 0.001), and PROMIS Pain domain and PODCI Pain/Comfort subscale (<i>r</i> = -0.71; p < 0.001). Divergent validity criteria were met with a correlation of < 0.30. The median number of questions required to answer on PROMIS was less than on PODCI. Reliability was assessed in 51 patients in Group B, in whom the PROMIS domains were collected again following a mean of 20 days. Test-retest reliability was highest for Mobility (intraclass correlation coefficient (ICC) = 0.84), Depressive Symptoms (ICC = 0.75), and Pain (ICC = 0.74). Responsiveness was assessed in 150 patients in Group C in whom PROMIS domains were collected preoperatively and at a mean of ten months postoperatively. The largest effect sizes were seen in the Pain and Mobility domains, with smaller improvements in the Physical Activity, Depressive Symptoms, and Upper Extremity domains. PROMIS domains of Mobility, Physical Activity, Depressive Symptoms, and Pain may be useful for assessing patient-reported health status in paediatric patients with lower limb orthopaedic conditions.