Validity, responsiveness and the minimal clinically important difference for the de Morton Mobility Index (DEMMI) in an older acute medical population

Abstract

Abstract Background The de Morton Mobility Index (DEMMI) is a new mobility instrument that overcomes the limitations of existing instruments. It is the first mobility instrument that accurately measures the mobility of all older adults. The aim of this study was to provide a detailed report of investigations of the validity, responsiveness to change and minimal clinically important difference (MCID) of the DEMMI during its development in an older acute medical population. Methods This study was conducted using a head to head comparison design in two independent samples of older acute medical patients (development sample, n = 86; validation sample, n = 106). Consecutive patients (≥ 65 years) were assessed using the DEMMI, Barthel Index (BI) and Hierarchical Assessment of Balance and Mobility (HABAM) within 48 hours of hospital admission and discharge. Convergent and discriminant validity were investigated using Spearman's rho and known groups validity was investigated using a independent t test to compare DEMMI scores for patients who were discharged to home compared to inpatient rehabilitation. Criterion and distribution based methods were employed for estimating instrument responsiveness to change and the MCID. Results Significant moderate to high correlations were identified between DEMMI and BI scores (r = 0.76 and r = 0.68) and DEMMI and HABAM scores (r = 0.91 and r = 0.92) in both samples. In both samples, DEMMI scores for patients who were discharged to home were significantly higher than for patients discharged to inpatient rehabilitation and provided evidence of known groups validity. Patients who were discharged to inpatient rehabilitation (n = 8) had a mean DEMMI score of 50.75 (sd = 11.29) at acute hospital discharge compared to patients who were discharged to home (n = 70) with a mean DEMMI score of 62.14 (sd = 18.41). MCID estimates were similar across samples using distribution and criterion based methods. The MCID for the DEMMI was 10 points on the 100 point interval scale. The DEMMI was significantly more responsive to change than the BI using criterion and distribution based methods in the validation sample. Conclusion This study has validated the DEMMI in two independent samples of older acute medical patients. Estimates of its responsiveness and MCID have also been established. This study confirms that the DEMMI overcomes the limitations of the BI and HABAM and provides an advanced method for objectively assessing mobility for older acute medical patients.