ObjectiveTo calibrate the Lower Extremity Functional Scale (LEFS) items into a regional lower extremity physical function (LEPF) item bank and assess reliability, validity, and efficiency of computerized adaptive test (CAT) and short form (SF) administration modes. DesignRetrospective cohort. SettingData were collected from patients treated in outpatient rehabilitation clinics for musculoskeletal impairments of the hip, knee, foot, and ankle that responded to all 20 LEFS items at intake. ParticipantsPatients aged 14 years or older who started an episode of care during January 2016-October 2019 and identified the lower extremity region as the source of a primary musculoskeletal complaint. Total cohort included 78,186 patients (mean age, 53±19y, range, 14-89y). InterventionsNot applicable. Main Outcome MeasuresItem response theory (IRT) model assumptions of unidimensionality, local item independence, item fit, and presence of differential item functioning (DIF) were studied. LEPF-CAT– and LEPF-SF–generated scores were evaluated. ResultsAn 18-item solution was supported for its unidimensionality and fit to the IRT model, with reliability estimates >0.9 for all administration modes. No DIF impact on LEPF scores was identified. Scores discriminated between multiple patient groups in clinically logical ways and were highly responsive to change, with negligible floor or ceiling effects. CAT scores were generated using an average of 4.9 items (median, 4). ConclusionsThe LEPF scores were reliable, valid, and efficient for assessing perceived physical function of patients with musculoskeletal impairments of the hip, knee, foot, and ankle; thus, it was found suitable for research and routine clinical administration. These findings are limited to the type of patients included in this study, with further validation needed to assess their generalizability.