Center on Population Health and Aging

Pilot Projects

Pilot Project 4: Coordinated Analysis of Change in Longitudinal Studies on Aging

Scott Hofer (PI), Andrea Piccinin, David Almeida, Lesa Hoffman, Carol Gold, Michael Rovine, Steve Zarit

External Collaborators: Gary Andrews (Australian Longitudinal Study of Aging; ALSA), Kaarin Anstey (ALSA), Stig Berg (H-70; OCTO-Twin), Helen Christensen (Canberra Longitudinal Study), Dorly Deeg (LASA), Boo Johansen (H-70; OCTO-Twin), Mary Luszcz (ALSA), Mike Martin (ILSE), Gerald McClearn (OCTO-Twin), Martin Sliwinski (Einstein Aging Study), and Avron Spiro III (Normative Aging Study)

Specific Aims

This Center for Demography and Aging pilot study will provide research support for the development of a project that will bring together researchers associated with longitudinal studies on aging for innovative and collaborative analysis and will provide an exemplar of the key features of collaborative research involving studies from multiple nations. This preparatory work is essential for establishing an international research network focused on cross-national comparability of aging-related changes in population health and cognition.

The goal is to describe population changes in cognition and health, and explain cognitive changes in the context of health-related changes (morbidity, comorbidity, and mortality). A fundamental concern motivating this proposed research is that the failure to consider health indicators is a major limitation for understanding aging-related changes in cognitive functions. Therefore, a key objective is to identify processes of change other than the simple passage of time that account for population and individual-level change in cognitive functions through detailed consideration of within-person change in health status (morbidity) and patterns of change prior to death. This multinational collaboration will permit the evaluation of fundamental questions regarding aging-related changes in cognition and health and their interaction, with the benefit of immediate cross-validation, evaluation of generalizability, and identification of associated demographic, cultural, and risk-related factors. The specific aims of this proposal are:

1. To further develop a relational database documenting the variables in each study. This database will be constructed such that variables similar at different levels (i.e., broad construct, narrow construct, specific measure) can be identified across studies. Such a searchable database is critical for identifying studies for inclusion in pooled data analysis versus those for parallel construct-level analysis. It will also include information regarding samples, response rates, and inter-occasions intervals. An essential aspect of this database development is the identification of study- and individual-level demographic indicators for cross-cultural (cross-study) analysis.
2. To develop a secure centralized web site to facilitate communication and transfer of information. A web-based system for abstracts of intent, permission for data usage, authorship agreements, and project management will be developed.
3. To implement a demonstration project that will address many of the potential logistical problems and will result in the development of appropriate analysis and data management protocol in cooperation with collaborators. The proposed research will focus on comparison and interpretation of results across studies that differ on demographic, sampling and measurement characteristics. This will require the development of a nested hierarchy of models to assess the degree of model invariance in the context of these varied characteristics. Data management and pooled/coordinated analyses will be refined.
4. To compare average population change in semantic memory cross-nationally. Alternative specifications of time (i.e., chronological age, time in study) will be used to evaluate the convergence of within-person changes and between-person age differences, with statistical control of demographic characteristics across studies.
5. To estimate the influence of dementia on between-person age-related differences, variability in rates of change, and within-person correlations in cognitive outcomes. We will demonstrate what will be further applied (in the full R01) to diseases including dementia, diabetes, respiratory disease, hypertension, cardiovascular disease, and cerebrovascular events in the evaluation of independent and joint efforts on the progression of changes in specific cognitive functions. This focus will be on memory changes occurring prior to diagnosis of dementia (i.e., time centered at diagnosis) in a combined multistudy analysis (replication of Sliwinski, Hofer, & Hall, 2003).

Return to Research Activities of CPHA Faculty Associates

Last modified: 09/17/09 | Contact Webmaster