The Bioanalytics Core (BC) will support the implementation of a uniform, innovative approach to the analyses of the Program's research Projects. The Core consists of Dr. Timothy Houle and Carol Aschenbrenner, both of whom are experienced biostatisticians with expertise in the proposed methods. The goals of the BC are to evaluate the statistical inferences posed in each of the Projects using the most appropriate statistical model. To accomplish these goals we will utilize several approaches including: growth curve modeling, generalized linear models, generalized estimating equations and other multivariate procedures (e.g., multidimensional scaling). The BC will work closely with the individual investigators of the Projects (PIs: Peters, Martin, Eisenach) to condition the data and address the specific hypotheses inherent in the research. All of the Projects involve studies that repeatedly measure some element of recovery over time after a surgical/experimental insult. Growth curve modeling, sometimes referred to as mixed-effects modeling, or hierarchical linear modeling, will allow us to specify a change trajectory (i.e., healing) that is unique to each individual/animal. The nature of the common form of changes in pain/behavior will be modeled using curvilinear forms (e.g., polynomial regression). Through the use of fixed and random effects, we will then be able to examine the influences on the changes in pain measurements in both the human and animal studies. In this way, we can examine the predictors immediately after surgery, the factors that predict delayed/absent recovery, patterns in these changes across individuals, and examine the relationships between initial measurements and expected change. This approach to the data analysis is highly innovative in this setting, and will extract novel information from these important data.
