HARMONIC REGRESSION-ANALYSIS OF THE EFFECT OF DRUG-TREATMENT ON THE DIURNAL RHYTHM OF BLOOD-PRESSURE AND ANGINA

Harmonic regression is a method for analysis of diurnal rhythms which we apply to blood pressure and angina attack rate. We show that population amplitude and phase as previously defined are not the expected values of amplitude and phase in the population unless patients are in phase and that the estimates of these population parameters are the amplitude and phase of the mean measurements of the observation times. We show that the amplitude of the mean measurements underestimates the magnitude of the diurnal rhythm of individual patients due to significant differences among patients in the phase parameter. Thus, the mean of the individual-patient amplitudes and phases, rather than the amplitude and phase of the mean measurements, is necessary to describe the diurnal rhythm in the population and to assess the effect of drug treatment. Harmonic regression analysis yields specific hypotheses that relate directly to the consistency of drug effect over 24 hours.