Test Properties

Internal Reliability: All items are approximately normally distributed with no significant skew or kurtosis. 

Content Validity: Cronbach’s alpha for the medical and behavioral subscales were .84 and .70 respectively. This indicates good internal consistency for the medical subscale and adequate internal consistency for the behavioral subscale. 

Construct Validity: All items have adequate factor loadings (> 0.4), with the exception of aggression which is just barely under the desired threshold (.395). Item loadings support the hypothesis that acuity is made up of two different, but related components that are labeled medical and behavioral acuity. 

Concurrent Validity: Significant correlations at admission and discharge from transitional rehabilitation with functional performance measures, such as the Mayo Portland Adaptability Index-4 (MPAI-4) (Malec, 2005), and assigned supervision levels, using the Supervision Rating Scale (SRS) (Boak, 1996), are strong enough to show relationship, but not so strong as to suggest they are measuring the same construct. MPAI-4 total score with HAS total score at admission (r =.80, p<.001), and at discharge (r =.81, p<.001). SRS score with HAS total score at admission (r_s=.53, p<.001), and at discharge (r_s=.66, p<.001). 

Discriminant Validity/Sensitivity: The HAS total score demonstrated sensitivity to change by showing significant change from admission (M = 11.40) to discharge (M = 7.97; t(101) = 7.04, p < .001).  The medical subscale also showed significant change from admission (M = 6.36) to discharge (M = 4.17; t(101) = 8.27, p < .001). Similarly, the behavioral subscale showed significant change from admission (M = 5.04) to discharge (M = 3.8; t(101) = 4.28, p <.001). This demonstrates a reduction in care needs for both medical and behavioral scales over the course of a transitional rehabilitation admission (mean admission = 74 days).