ABSTRACT
In this chapter, we introduce some of the basic concepts of model building that will set the stage for the more detailed treatment given in Chapters 3 through 5.
The concept of individual variability is as old as human observation. Different people respond differently to the same dose of drug, to the same decibel noise level, or to the same temperature changes. Physiological parameters such as cardiac output (CO), diastolic blood pressure, and blood glucose level, to name just a few, all vary widely from person to person. To help us make sense of these variations and to provide a framework in which we can operate in a logical way, we usually think in terms of a standard person. Our standard person is a hypothetical statistical construct that allows us to use average values of physiological parameters in our models. When we use single values for physiological variables (i.e., 120/80 mm Hg for systolic/ diastolic blood pressure), we implicitly recognize that we are using average values that only apply under a limited set of conditions. In fact, all physiological variables have a range of values when measured in clinical laboratories using similar techniques. These are called normal or reference values [1,2]. By convention, in the biomedical sciences, the normal range is reported as the 95% conšdence interval [3] for a given variable. When there are variations in the values as measured in different laboratories, reference values are reported as measured in a particular laboratory (e.g., Stanford University and Massachusetts General). The mean value and the conšdence interval for any variable may change with age, sex, weight, race, or one or more environmental, socioeconomic, or hereditary factors. Table Appendix A lists adult reference values for a number of routine laboratory tests performed in hospitals, clinics, and physicians’ ofšces.
