Missing data at follow-up: The case of the interRAI home care assessment instrument in Belgium

• Published in: European geriatric medicine Vol. 7; no. 4; pp. 322 - 327
• Main Authors: Vanneste, D, De Almeida Mello, J, Macq, J, Van Audenhove, C, Declercq, A
• Format: Journal Article
• Language: English
• Published: Elsevier Masson SAS 01.07.2016
• Subjects: BelRAI; Dropouts; Geriatric assessment; Internal Medicine; InterRAI; Loss to follow-up; Missing data; BelRAI; Missing data; Geriatric assessment; Dropouts; Loss to follow-up; InterRAI
• ISSN: 1878-7649; 1878-7657
• DOI: 10.1016/j.eurger.2016.03.004

Abstract Background In Belgium, a web application BelRAI was developed to support the use of interRAI Home Care assessments. However, not all clients have had a follow-up assessment. Objectives To examine the prevalence of incomplete assessments at 6-month follow-up and to compare the profiles of persons with and without missing data. Design Observational study of interRAI Home Care data and data from the Belgian Intermutuality Agency. Setting Home care in Belgium. Participants Healthcare professionals using BelRAI in the context of an evaluation of home care interventions to delay institutionalization of frail older people. Methods Descriptive statistics were calculated based on demographic characteristics and outcome measures. Poisson regressions for the relative risk of institutionalization and for the relative risk of death were performed for a period of 6 and 12 months. Results In a sample of 4987 persons (mean age 79.90), two groups were determined: a group without and a group with missing data at 6-month follow-up. At baseline, the group with missing data at follow-up has a worse health status than the group without missing data. Persons with missing data at 6-month follow-up had a 3.63 higher chance of dying between 6 and 12 months after baseline, a 3.43 higher chance for institutionalization at 6 months, and a 2.87 higher chance for institutionalization at 12 months. Conclusion In this study, missing data at follow-up is not a random occurrence. The findings are important for healthcare professionals and for researchers working with longitudinal data. Further study is required to research possible solutions.