Original ArticleA systematic literature review of attrition between waves in longitudinal studies in the elderly shows a consistent pattern of dropout between differing studies
Introduction
In the investigation of longitudinal effects in population-based studies, two important areas of loss of individuals from the study need to be investigated. One area that has been fairly well investigated is that of initial non-response [1], [2], [3], [4], [5]. The other area of concern is that of dropout between waves, both due to death and other types of dropout such as refusal. In studies of older people loss due to death cannot be avoided, but by spacing interviews close together factors relating to terminal decline will not affect results unduly. Dropout of the living can cause the remaining participants to be a biased sample. Knowledge of this bias, if systematic, could help future researchers to lessen or prevent it, as well as helping to assess the validity of extrapolating results based on the remaining participants. If there are consistent patterns of non-response in large population-based studies in older people, these effects should be given more prominence in consideration for design and analysis.
The aim of this review is to investigate all large population based studies of the elderly that report on attrition between waves of a follow-up in a systematic manner. The review concentrates on that dropout which could in theory be prevented or reduced, that is, the loss to follow-up due to refusal, sickness, inability to locate individuals, or individuals moving away from a defined study area.
Section snippets
Methods
In June 2003, the MEDLINE, PSYCHINFO, and ISI WoS (Scientific and Social) databases were used to find all available articles published up until December 31, 2002. Articles of all languages were investigated, initially by inspection of the translation of the abstract, then by requesting the article. The search terms and exclusions are detailed in the Appendix. A total of 1792 articles were found using these methods of review. In addition to the search criteria, publications of large population
Results
All 25 articles that met the criteria on inspection are detailed in Table 1. These articles were further investigated to ensure that the analyses presented in the article were of sufficient detail to enable comparison of the factors affecting attrition between studies. Seven studies were excluded due to problems with the analysis, four because dropout due to death was included with the attrition [6], [7], [8], [9], one did not have a uniform follow-up procedure [10], one depended on a
Discussion
This is the first reported systematic review of dropout in population-based longitudinal studies of older people. It has been shown that individuals who dropped out for reasons other than death tended to be older and were more cognitively impaired. People who are ill or frail are more likely to drop out, and people who cannot be contacted may be in worse health than continuing respondents. Many of the large longitudinal studies have not produced dedicated articles about attrition, which
Acknowledgments
The authors would like to thank Kikumi Wilson and Yvonne Salis-Trapala for help in translating the articles by Sugisawa, 2000 [24] and Buen, 1993 [19]. M.C. and F.M. are funded by the Medical Research Council as part of the Cognitive Function and Ageing study co-operative.
References (32)
- et al.
Refusal and information bias associated with postal questionnaires and face-to-face interviews in very elderly subjects
J Clin Epidemiol
(1996) - et al.
Who participates in population based studies of visual impairment? The Salisbury Eye Evaluation project experience
Ann Epidemiol
(1999) - et al.
Attrition in the Longitudinal Aging Study Amsterdam. The effect of differential inclusion in side studies
J Clin Epidemiol
(2002) - et al.
The Gloucestershire Longitudinal Study of Disability: outcomes in nonresponders, responders, and subsequent defaulters
J Clin Epidemiol
(1998) - et al.
Loss to follow-up in a longitudinal study on aging in Spain
J Clin Epidemiol
(2001) - et al.
[Characteristics of non-responders in a follow-up study of those over 65 years of age]
Gac Sanit
(1993) - et al.
Transitions in cognitive status among the aged in Japan
Soc Sci Med
(1996) - et al.
Nonresponse pattern and bias in a community-based cross-sectional study of cognitive functioning among the elderly
Am J Epidemiol
(1994) - et al.
Patterns of nonresponse in a national survey of elderly Japanese
J Gerontol
(1993) - et al.
Correlates of nonparticipation in the Canadian Study of Health and Aging
Int Psychogeriatr
(2001)
Using the “Mini-Mental State” Examination to predict elderly subjects' completion of a follow-up interview
Am J Epidemiol
Age, depression, and attrition in the national survey of families and households
Sociological Methods and Research
Total homocysteine and cognitive decline in a community-based sample of elderly subjects: The Rotterdam Study
Am J Epidemiol
Aging and total cholesterol levels: Cohort, period, and survivorship effects
Am J Epidemiol
Method and description of the Aging in Manitoba project: A 20-year longitudinal study
Canadian Journal on Aging – Revue Canadienne Du Vieillissement
Assessment of bias in the SENECA study
Eur J Clin Nutr
Cited by (332)
Healthy ageing from birth to age 84 years in the Helsinki Birth Cohort Study, Finland: a longitudinal study
2023, The Lancet Healthy LongevityLongitudinal study
2023, Handbook for Designing and Conducting Clinical and Translational SurgeryCross-sectional study
2023, Handbook for Designing and Conducting Clinical and Translational Surgery