Analysis of the physiological stress profile: the interrupted time-series design

Most research employing physiological stress profiling procedures has relied upon statistical analyses that are flawed in two ways: These analyses typically average the physiological levels observed across a series of observations during any one phase of the profile, thereby ignoring rate of change as a relevant parameter, and the problem of autocorrelation, or the natural correlation of time-series observations of the same physiologic activity, is unaddressed. We hope to introduce the biofeedback field to the technique of interrupted time-series analysis, which effectively deals with these two flaws. Interrupted time-series analysis additionally permits statistical conclusions based upon the stress profile of a single individual, thereby providing an objective basis for decisions concerning the effectiveness of training or the timing of changes in training for a single client. We describe the application of the technique to the analysis of a stress profile of a subject diagnosed as hypertensive.