New complementary perspectives for inpatient physical function assessment: matched clinician-report and patient-report short form measures from the PROMIS adult physical function item bank

• Published in: Quality of life research 2022-07, Vol.31 (7), p.2201-2212
• Main Authors: Kallen, Michael A., Brown, Heather E., Hatton, Joeffrey R., Doyle, William A., Murphy, Ryan, Elliott, Ryan, Gutierrez, Mark A., Catherwood, Emma L., Pitman, Heather P., Liu, Vincent X., Gershon, Richard C.
• Format: Article
• Language: English
• Subjects: Electronic health records; Hospitalization; Hospitals; Information systems; Medicine; Medicine & Public Health; Patients; Physical therapists; Public Health; Quality of life; Quality of Life Research; Sociology
• ISSN: 0962-9343; 1573-2649
• DOI: 10.1007/s11136-022-03089-z

Purpose To develop two item content-matched, precise, score-level targeted inpatient physical function (PF) short form (SF) measures: one clinician-reported, one patient-reported. Items were derived from PROMIS PF bank content; scores are reported on the PROMIS PF T-score metric. Methods The PROMIS PF item bank was reviewed for content measuring lower-level PF status (T-scores 10–50) with high item set score-level reliability (≥ 0.90). Selected patient-reported (PR) items were also edited to function as clinician-reported (CR) items. Items were reviewed by clinicians and field tested; responses were assessed for meeting PROMIS measure development standards. New CR and PR items were calibrated using patient responses to the original PROMIS PF items as anchoring data. SFs were constructed, based on content and precision. Results Nine PROMIS PF items were candidates for CR and PR inpatient PF assessment; three new items were written to extend content coverage. An inpatient sample ( N  = 515; 55.1% female; mean age = 66.2 years) completed 12 PR items and was assessed by physical therapists (using 12 CR items). Analyses indicated item sets met expected measure development standards. Twelve new CR and three new PR items were linked to the PROMIS PF metric (raw score r  = 0.73 and 0.90, respectively). A 5-item CR SF measure was constructed; score-level reliabilities were ≥ 0.90 for T-scores 13–45. A 5-item PR SF measure was assembled, mirroring CR SF content. Conclusions Two item content-matched SFs have been developed for clinician and patient reporting and are an effective, efficient means of assessing inpatient PF and offer complementary perspectives.