A primary function of the Biostatistics Core will be to collaborate with investigators on appropriate designs for studies of multifactorial geriatric health conditions, including power and sample size determinations, proper methods of sampling and/or randomization (when applicable), control of bias, robust methods for estimating treatment effects, use of proxies and of novel and efficient designs, development of plans for interim monitoring of study progress, and considerations about cross-study analyses. A major issue in these designs will be accounting for loss of information from the experimental participants. The appropriate use of proxies and other techniques for recording information will be introduced into the study designs. A second primary function of the Biostatistics Core will be to collaborate with investigators on the analytic plans for OAIC studies prior to their initiation. These plans will include in the case of clinical trials, considerations of recruitment, baseline comparability of treatment arms, completeness of follow-up, adherence to assigned therapy, data quality, monitoring of safety and treatment efficacy, appropriateness of subgroup analyses and control for multiplicity (multiple comparisons and outcomes). Since the analysis flows from proper statistical design and appropriate hypotheses, the Core will ensure that the analytic plans and study designs are consistent with meeting the primary objectives of each study and of the OAIC. The Core will also conduct all analyses for OAIC studies according to the established project timelines. A major role of the Biostatistics Core will be the application and development of new statistical methodology to enhance OAIC research. Since many of the OAIC projects will involve longitudinal analyses, informative censoring because of death or the incapacity of the study participants is likely; thus, this will be a particular area of focus for the Core. The Core will keep abreast of new methodological developments and extend techniques from applied statistics. Development of new methodology will be accomplished through development projects. The first two such projects proposed include: 1) developing analytic approaches to determine the mechanisms of action of multifactorial interventions (Peduzzi, PI), 2) developing analytic approaches for determining the effect of precipitating events on outcomes and for handling bidirectionality (feedback loops) between risk factor or treatment and outcome (Dubin, PL).
Showing the most recent 10 out of 691 publications
