STATISTICAL METHODOLOGIES USEFUL FOR THE ANALYSIS OF DATA FROM RISK-ASSESSMENT STUDIES

In dentistry, the vast majority of studies attempting to predict who is at high risk for getting a disease or condition, or attempting to identify risk factors for a specific disease or condition, have focused either on only one risk factor at a time, or have measured multiple potential risk factors, but analyzed their effects in isolation. Since researchers tend to agree that most dental conditions have a multiple etiology, it is necessary to develop models that consider simultaneously the effect of a number of potential risk factors on the disease or condition of interest if we are to have any understanding of the relative impact of potential risk factors. Many existing statistical techniques will aid dental researchers in identifying risk factors. However, the selection of an appropriate analytic technique depends on a number of conditions. The strategy for this paper is to discuss a wide range of possible statistical techniques that may be applied to the problem of deriving a model for identification of multiple risk factors for dental diseases and conditions. We have approached this task by presenting a number of dental research problems needing an appropriate analytic technique. Next, basic issues that must be considered in choosing an appropriate analytic strategy are discussed. These issues include features of the study design, the data structure of the variables being measured, and the types of assumptions that are applicable to provide valid inferences about the target population of interest. A matrix of possible analytic techniques is presented for various combinations of study-design and data-structure features. After a discussion of each of the techniques, the appropriate statistical techniques for each of the dental examples are described. The issues and examples presented in this paper should be of use to dental researchers who wish to investigate multiple risk factors for a disease or condition of interest.