ABSTRACT
In this chapter, we describe various statistical methods for monitoring healthy subjects or outpatients after treatment. In the first case-monitoring healthy subjects-within-subject variances play the central role. Observations are generally scheduled at long intervals (at least every 6 months, often annually or even less frequently), and deviations from the appointed times are relatively minor. The immediate problem is that of judging the significance of a change between the initial measurement of an analyte and the next observation. To resolve this we must use information about within-subject variances obtained from repeated measurements in other healthy subjects. When several serial measurements of the analyte become available while the individual remains in a steady state, we can turn to methods of deriving predictive ranges that rely solely on the person's own past record. Such methods, and the statistical models behind them, are necessarily relatively simple.
