Functional scale discrimination at admission and discharge: Rasch analysis of the Level of Rehabilitation Scale-III.

OBJECTIVE: To determine the construct validity of the Level of Rehabilitation Scale-III (LORS-III) with a special focus on this instrument's capability to discriminate rehabilitation inpatient activities of daily living (ADL)/mobility and communication/cognition ability at admission and discharge. DESIGN: Rasch analysis of existing data sets in the LORS-III American Data System (LADS). PATIENTS: Existing admission and discharge data from 3056 rehabilitation inpatients (musculoskeletal injury, cerebrovascular accident, multiple injuries/diseases, brain injury, neuromuscular disorder, and spinal cord injury) entered into LADS between April 1992 and January 1993. MAIN OUTCOME MEASURE: LORS-III consists of 17 measurement areas representing abilities in ADL, mobility, communication, cognition, and memory. Fourteen of the measurement areas are concurrently scored by a nurse and a specified rehabilitation therapist, resulting in a total of 31 items. RESULTS: Consistent with findings reported for other functional status measures, the analysis indicated that the LORS-III consists of two unidimensional scales, an ADL/mobility scale, and a communication/cognition scale. Although all scales fit the Rasch measurement model, the ADL/mobility scale used at admission was most appropriately targeted to the ability level of the sample. At discharge, the ADL scale was generally too easy because the ability level of the sample moved upward towards functional independence. The communication/cognition scale at both admission and discharge showed a similar "ceiling" effect. CONCLUSIONS: These findings indicate the importance of determining the measurement qualities of functional status measures for both admission and discharge ratings. Analyses, such as Rasch, can provide a logical direction for instrument refinement.