CORE--DATA MANAGEMENT AND STATISTICS

University Of Pittsburgh At Pittsburgh, Pittsburgh PA

Investigators

Linked publications & trials

Abstract

Core C comprises the data management and statistics core of the AgeWise program. It will handle all aspects of data storage, retrieval and analysis (including statistical analysis) downstream from actual data collection. The premise and fundamental hypotheses to be tested by projects in AgeWise are that protecting sleep in later life will contribute to preventing or slowing declines in functional status and will promote the preservation of good physical and mental health for as long as possible in elderly vulnerable to functional decline by the challenges of bereavement, care-giving, insomnia in medical illness, and age itself as they enter the final years of life. Core C will not only be involved in the data management and analysis of data from individual component projects, but also from the AgeWise program as a whole. Program-wide data will be used to explore the mediating role of sleep disruption in determining the health and well-being subsequent to common late-life challenges. The AgeWise program seeks to perform an ambitious program of research involving a large number of data from polysomnography, intervention sessions and the collection of a broad range of health, psychosocial and functioning domains. Our laboratories have had three decades of experience in handling such data. Core C will be composed of two interrelated components: Data Management and Statistics. Under the expert guidance of Dr. Victoria Grochocinski, the data management component will handle all aspects of data processing, storage, and analysis downstream from data collection which will be carried out in Core B (Data Collection - Buysse). Consistency across data collection and management methods will allow investigators prompt access to the data pertaining to their component projects. The statistical component of Core C will be lead by Dr. Sati Mazumdar and Dr. Hernando Ombao, who will work closely with Drs. Buysse, Hall, Monk, Nofzinger and Reynolds in their component projects. An appropriate level of statistical standardization and sophistication in the conduct of statistical analyses will be maintained. Our analytic approach consists in examining the data carefully through descriptive analysis prior to any statistical modeling or inference. A critical look at all data is also necessary for the program-wide research agendas which are part of all five component projects. Data reductions, efficiency in analysis, controlling of type I error and parsimony in analysis are important considerations in our approach. Cutting-edge statistical methodologies will be used in all our analyses.