Validity, inter-rater reliability and responsiveness of the modified Iowa Level of Assistance (mILOA) for acutely hospitalised general medical patients

ObjectiveTo evaluate the psychometric properties of the modified Iowa Level of Assistance (mILOA) in general medical patients aged ≥65 years.DesignProspective observational psychometric measurement study.<b>Main outcome measures</b>: The mILOA was completed at hospital admission (<i>n = </i>246) and discharge (<i>n = </i>121). A second measure was completed within 24-h to establish inter-rater reliability and measurement error (<i>n = </i>50). Criterion predictive validity for discharge destination, floor and ceiling effects, responsiveness, and construct validity, via convergent validity (with the de Morton Mobility Index, DEMMI) and known group validity (discharge home), were assessed using standard psychometric techniques.ResultsThere were 246 included participants (median age = 83.2; interquartile range = 46.1-88.0 years; 47% female). The mILOA took a median of 12.5 [interquartile range = 8.0-18.0] minutes to complete. There were no systematic differences across the range of scores (mean difference = 1.1) with limits of agreement of 7.7 and -5.4 (variability due to measurement error = 2.36). It was able to predict discharge destination (area under the curve = 0.74; 95% CI = 0.66-0.81). No floor or ceiling effects were observed, and it was highly responsive to change (effect size = 0.80). Scores correlated significantly and positively with DEMMI scores (Spearman's rho = 0.86, 95% confidence interval = 0.81-0.90, <i>p</i> < .001) and showed a median difference of 8 points between known groups.ConclusionThe mILOA has good inter-rater reliability, no floor or ceiling effects, strong responsiveness, high correlation with the DEMMI, and the ability to discriminate between known groups in older general medical patients.