Thursday, 21 August 2003
This presentation is part of : Thursday Poster Sessions

PD-037 The Longitudinal Study on Aging: Relationships Between Self-Reported Confusion in Community Dwelling Elderly and Time to Death

Helen H. Kyomen, Division on Aging and Department of Psychiatry, Harvard Medical School/McLean Hospital, Belmont, MA, USA and John Hennen, Psychiatry, Harvard Medical School/McLean Hospital, Belmont, MA, USA.

OBJECTIVE: To determine whether self-reported confusion in the elderly is predictive of time to death over a six-year period.

DESIGN: Retrospective cohort analysis of the Longitudinal Study of Aging(LSOA) dataset.

MATERIALS AND METHODS:Subjects: Nationally representative (United States) sample of 7,527 people aged >=70 years and living in the community at baseline (in 1984). Follow-up assessments, including determinations of dates of death, if this event occurred, were completed by the LSOA project in 1986, 1988, and 1990.

Procedures:This study evaluates the relationships between self-reported confusion and time to death in the LSOA cohort of elderly>=70 years of age. The LSOA, over a seven-year period, obtained data on functional and living circumstances in a cohort of older individuals who were >=70 years old and living in the community at baseline. Data, at baseline and over time, on demographic, social,economic, health characteristics, and measures of healthcare use, caregiving and social support for individuals also were gathered. Cox proportional hazards regression methods were used to estimate hazard ratios related to risk of mortality over the 84-month period spanned by the LSOA project.

RESULTS: The mortality hazard ratio (HR) associated with the self-reported confusion factor was 1.29 (95%CI 1.17 - 1.41), suggesting an increased mortality risk of approximately 30% for subjects self-identifying as experiencing "confusion." When other factors strongly correlated with mortality, including age, sex, education level, family income, and presence/absence of a large number of important medical problems, including history of angina, cancer, coronary heart disease, coronary artery disease,diabetes, hypertension, myocardial infarction, pneumonia,rheumatic heart disease, and stroke, were added to the Cox models,the hazard ratio for confusion remained statistically significant. The adjusted HR for confusion (with adjustment for all these factors) was 1.13 (95%CI 1.02 - 1.25). These estimates were developed without regard to the LSOA survey plan, which was designed to provide a representative sample of the total U.S. elderly (age>=70 years) population. The hazard ratio and SE estimates will need to be adjusted using weights based on the LSOA complex survey methods in future analyses.

CONCLUSION: An increased mortality risk of approximately 30% was associated with self-reported confusion in community-dwelling elderly. The clinical implications may be important. Many community-dwelling elderly likely experience increased confusion without meeting full criteria for delirium, dementia, or other neuropsychiatric diagnoses. Soliciting information clinically about the subjective experience, among community-dwelling elderly, of episodes of confusion may be a useful way to identify patients who need further evaluation and care for subsyndromal delirium, dementia, or other conditions that may impair cognitive abilities, such as depression, mania, psychosis or anxiety. With early recognition of confusion in the elderly, there may be enhanced opportunities for patients to seek and receive appropriate care for psychiatric and medical problems that may be manifesting themselves through alterations in mental state.

Back to PD Thursday Poster Sessions
Back to The Eleventh International Congress