The performance of PROMIS computer adaptive testing for patient-reported outcomes in hip fracture surgery: a pilot study

Abstract

Introduction Patient-reported outcome measures (PROMs) are essential to patient-centered care in orthopaedics. PROMIS measures have demonstrated reliability, validity, responsiveness, and minimal floor and ceiling effects in various populations of patients receiving orthopaedic care but have not yet been examined in hip fracture patients. This pilot study sought to evaluate the psychometric performance of the PROMIS Physical Function (PROMIS PF) and Pain Interference (PROMIS PI) computer adaptive tests and compare these instruments with legacy outcome measures in hip fracture patients. Methods This study included 67 patients who were 27–96\xa0years old (median 76) and underwent osteosynthesis for a proximal femoral fracture. At 3, 6, and/or 12\xa0months follow-up, patients completed both legacy (mHHS, SF-36-PCS, and VAS for pain) and PROMIS questionnaires (PROMIS PF and PROMIS PI). Respondent burden and floor/ceiling effects were calculated for each outcome measure. Correlation was calculated to determine concurrent validity between related constructs. Results A strong correlation was found between PROMIS PF and mHHS (rho\u2009=\u20090.715, p \u2009<\u20090.001) and moderately strong correlation between PROMIS PF and SF-36 PCS (rho\u2009=\u20090.697, p \u2009<\u20090.001). There was also a moderately strong correlation between the VAS and the PROMIS PI (rho\u2009=\u20090.641, p \u2009<\u20090.001). Patients who completed PROMIS PF were required to answer significantly fewer questions as compared with legacy PROMs (mHHS, SF-36). For the PROMIS measures, 1% of patients completing PROMIS PF achieved the highest allowable score while 34% of patients completing PROMIS PI achieved the lowest allowable score. Of the legacy outcome measures, 31% of patients completing the VAS for pain achieved the lowest allowable score and 7% of patients completing the mHHS achieved the highest allowable score. Conclusions The results of this study support the validity of PROMIS CATs for use in hip fracture patients. The PROMIS PF was significantly correlated with SF-36 PCS and mHHS while requiring fewer question items per patient relative to the legacy outcome measures.