One approach to overcoming potential biased results, due to dropout from longitudinal clinical studies, is to capture additional data once a marker of health downturn is observed but before the patient leaves the study. We denote this study design feature as “triggered sampling” (TS).
Study Design and Setting
We formally define TS, describe some mechanisms for incorporating TS in longitudinal studies, and present the results from a 2-year longitudinal observational study of treatment preferences, measured on a 1–7 scale, of patients with advanced illness from cancer, congestive heart failure, or chronic obstructive pulmonary disease. We examined the utility of TS through multiple analyses, including mixed effects models.
One hundred forty-eight of 226 participants experienced at least one triggered interview. Those who did not drop out after their first trigger had no noticeable change in their mean preferences (6.20 pretrigger, 6.16 trigger, P=0.76), whereas those who dropped out after their first trigger did (6.29 pretrigger, 5.69 trigger, P=0.04). The mixed effects models conveyed similar results, providing support for the efficiency and efficacy of TS.
TS can help alleviate bias due to impending dropout and potentially be a valuable addition to the designs of longitudinal studies of persons with elevated mortality risk.
To read this article in full you will need to make a payment
Purchase one-time access:Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
One-time access price info
- For academic or personal research use, select 'Academic and Personal'
- For corporate R&D use, select 'Corporate R&D Professionals'
Subscribe:Subscribe to Journal of Clinical Epidemiology
Already a print subscriber? Claim online access
Already an online subscriber? Sign in
Register: Create an account
Institutional Access: Sign in to ScienceDirect
- A systematic literature review of attrition between waves in longitudinal studies in the elderly shows a consistent pattern of dropout between differing studies.J Clin Epidemiol. 2005; 58: 13-19
- Practice and drop-out effects during a 17-year longitudinal study of cognitive aging.J Gerontol B Psychol Sci Soc Sci. 2004; 59: 84-97
- Neglect of dropout underestimates effects of death in longitudinal studies.J Gerontol B Psychol Sci Soc Sci. 2005; 60: 106-109
- State-of-the-Science Conference Statement: Improving End-of-Life Care.U.S. Department of Health and Human Services, Bethesda, MD2004
- Patient preferences: integrating treatments and outcomes.J Gerontol B Psychol Sci Soc Sci. 2002; 57: S348-S354
- Statistical analysis with missing data.2nd edition. John Wiley & Sons, Hoboken2002
- Inference and missing data.Biometrika. 1976; 63: 581-592
- Bayesian and profile likelihood change point methods for modeling cognitive function over time.Comput Stat Data Anal. 2003; 42: 91-109
Published online: September 29, 2006
Accepted: June 2, 2006
© 2007 Elsevier Inc. Published by Elsevier Inc. All rights reserved.