Premium
Reliability of the de Morton mobility index (DEMMI) in an older acute medical population
Author(s) -
de Morton Natalie A.,
Davidson Megan,
Keating Jennifer L.
Publication year - 2011
Publication title -
physiotherapy research international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.509
H-Index - 49
eISSN - 1471-2865
pISSN - 1358-2267
DOI - 10.1002/pri.493
Subject(s) - confidence interval , reliability (semiconductor) , medicine , inter rater reliability , kappa , population , physical therapy , psychology , statistics , rating scale , mathematics , power (physics) , physics , geometry , environmental health , quantum mechanics
Background. The de Morton Mobility Index (DEMMI) is an instrument that accurately measures the mobility of older people across clinical settings. Purpose. To report the multiple reliability studies conducted during the development and validation of the DEMMI. Methods. Intra‐rater and inter‐rater reliability studies were conducted for the DEMMI in two independent samples (development and validation samples) of older acute medical patients (aged 65 years or older). Inter‐rater reliability studies were conducted between the test developer (a physiotherapist) and another experienced physiotherapist. Order of assessor administration was randomized by a coin toss. Patients who were fatigued after the first assessment were excluded from the inter‐rater reliability study. Intra‐rater reliability studies included participants with ‘unchanged’ mobility status between hospital admission and discharge. Scale reliability estimates were expressed as the minimal detectable change with 90% confidence (MDC 90 ). Item reliability was calculated using Kappa statistics and absolute percentage agreement. Results. The MDC 90 for the DEMMI development sample was 9.51 points (95% confidence interval [95% CI], 5.04–13.32; n = 21) and 7.84 (95% CI, 4.34–11.65; n = 16) on the 100‐point interval DEMMI scale for the inter‐rater and intra‐rater reliability studies, respectively. Similar estimates were obtained for the DEMMI validation samples of 8.90 (95% CI, 6.34–12.69; n = 35) and 13.28 points (95% CI, 8.08–20.87; n = 19). Items were not excluded from the DEMMI based on the results of item reliability. Conclusion. Reliability estimates for the DEMMI were consistent across independent samples of older acute medical patients using different reliability study methodology. Error represents approximately 9% of the DEMMI scale width. Copyright © 2010 John Wiley & Sons, Ltd.