State-of-the-science on postacute rehabilitation: measurement and methodologies for assessing quality and establishing policy for postacute care

Duncan PW, Velozo CA. State-of-the-science on postacute rehabilitation: measurement and methodologies for assessing quality and establishing policy for postacute care.We present an overview of commonly used postacute outcome measures and review new methodologies for postacute assessment. We question the impact that current measurement has had on improvement of quality of postacute care (PAC) and its utility in informing health policy. We suggest that Donabedian’s model of health care quality should be endorsed for measurement. Specifically, measurement of outcomes and process should be used jointly in assessment of PAC.     

Utilizing Rasch measurement models to develop a computer adaptive self-report of walking,climbing, and running

Purpose.?The purpose of this paper is to show how the Rasch model can be used to develop a computer adaptive self-report of walking, climbing, and running.

Method.?Our instrument development work on the walking/climbing/running construct of the ICF Activity Measure was used to show how to develop a computer adaptive test (CAT). Fit of the items to the Rasch model and validation of the item difficulty hierarchy was accomplished using Winsteps software. Standard error was used as a stopping rule for the CAT. Finally, person abilities were connected to items difficulties using Rasch analysis ‘maps’.

Results.?All but the walking one mile item fit the Rasch measurement model. A CAT was developed which selectively presented items based on the last calibrated person ability measure and was designed to stop when standard error decreased to a pre-set criterion. Finally, person ability measures were connected to the ability to perform specific walking/climbing/running activities using Rasch maps.

Conclusions.?Rasch measurement models can be useful in developing CAT measures for rehabilitation and disability. In addition to CATs reducing respondent burden, the connection of person measures to item difficulties may be important for the clinical interpretation of measures.     

Assessing arm and hand function after stroke: a validity test of the hierarchical scoring system used in the motor assessment scale for stroke

OBJECTIVE: To evaluate the validity of the scoring hierarchy for the 3 upper-limb items on the Motor Assessment Scale (MAS). DESIGN: Application of Rasch analysis to 3 independent measurement scales, each representing the upper-arm function, hand movements, and advanced hand activities items of the MAS. SETTING: Inpatient and outpatient occupational therapy (OT) programs in a department of rehabilitation of an urban hospital center. PARTICIPANTS: One hundred patients (67 men, 33 women; average age, 54.3+/-14.4 y; average time since stroke onset, 104 d) attending OT for stroke rehabilitation. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURE: The MAS for stroke (upper-arm function, hand movements, and advanced hand activities sections). RESULTS: Rasch analysis provided support for the validity of hierarchical scoring criteria for the upper-arm scale. This analysis, however, identified inconsistencies in the hierarchical scoring criteria for the hand function and advanced hand activities scales and, when considering measurement error, only small differences in difficulty level between several behavioral criteria. CONCLUSIONS: The findings lead to suggestions for changes in the behavioral criteria hierarchy for upper-limb items on the MAS and highlight the importance of using statistical analyses to test the validity of proposed hierarchies of behavioral criteria in functional assessments.     

Psychometrics of the self-report safe driving behavior measure for older adults

We investigated the psychometric properties of the 68-item Safe Driving Behavior Measure (SDBM) with 80 older drivers, 80 caregivers, and 2 evaluators from two sites. Using Rasch analysis, we examined unidimensionality and local dependence; rating scale; item- and person-level psychometrics; and item hierarchy of older drivers, caregivers, and driving evaluators who had completed the SDBM. The evidence suggested the SDBM is unidimensional, but pairs of items showed local dependency. Across the three rater groups, the data showed good person (≥3.4) and item (≥3.6) separation as well as good person (≥.93) and item reliability (≥.92). Cronbach's α was ≥.96, and few items were misfitting. Some of the items did not follow the hypothesized order of item difficulty. The SDBM classified the older drivers into six ability levels, but to fully calibrate the instrument it must be refined in terms of its items (e.g., item exclusion) and then tested among participants of lesser ability.     

Limitations in dose–response and surrogate species methodologies for risk assessment of Cry toxins on arthropod natural enemies

Dose–response assays and surrogate species are standard methods for risk analysis for environmental chemicals. These assume that individuals within a species have unimodal responses and that a surrogate species can predict responses of other related taxa. We exposed immature individuals of closely related aphidophagous coccinellid predators, Cycloneda sanguinea and Harmonia axyridis, to Cry1Ac and Cry1F toxins through uniform and constant artificial tritrophic exposure through Myzus persicae aphids. Both toxins were detected in coccinellid pupae, with individual and interspecific variation. Uptake was significantly higher in H. axyridis than in C. sanguinea, both in the proportion of individuals and the concentrations per individual. We also observed bimodal uptake of the Cry toxins by H. axyridis, which indicated that some individuals had low bioaccumulation and some had high bioaccumulation. This suggests that standard dose–response assays need to be interpreted with caution and future assays should examine the modality of the responses. In addition, the similarity in the biological effects of the Cry toxins in the two predators was due to different biological exposure mechanisms. The majority of H. axyridis were exposed both internally and in the gut, while C. sanguinea was exposed primarily in the gut. Thus, despite their close phylogenetic relatedness, these species would not be good surrogates for each other and the surrogate species methodology should be tested more rigorously.     

Prediction of return to work by rehabilitation professionals

Rehabilitation therapists made predictions of return to work for 57 patients evaluated prior to beginning a work-hardening program. These predictions, along with 14 specific demographic, injury, and physical function measures were entered into a stepwise discriminant analysis to develop a predictive model for return to work. Therapists correctly predicted the eventual return to work for 47 (73.7%) of the 57 subjects. Therapist prediction was the most powerful variable in the model, explaining 29% of the variance (p.0001). Only two other variables, self-report of pain severity and injury type, contributed significantly to the prediction model, accounting respectively for 12% and 9% of the variance (p's<.05). These data highlight the predictive acuity of rehabilitation therapists and suggest that the process involved in formulating clinical predictions merits further study. The findings also suggest that weighing information such as self-report of pain severity and injury type might further enhance the practitioner's ability to predict return to work.