14. Hsu, P.H., Mathewson, F.A.L., Abu-Zeid, H.A.H., Rabkin, S.W.: Change in risk factor and the development of chronic disease. A methodological illustration*. J Chron Dis 30: 567-584, 1977.

This report examines and suggests solutions for the methodological problems usually encountered when the relationship between the rate of change in risk factors and the development of chronic disease is studied. The general pattern of changes in body mass index (BMI), systolic blood pressure (SBP) and diastolic blood pressure (DBP) were compared between two groups: one including 525 subjects who developed ischemic heart disease (IHD) and/or stroke (CVA) over 26.5 yr of follow-up (IHD/CVA) and another group of 2877 subjects who remained free from these diseases during the same follow-up period (reference group).

The longitudinal trends of these factors in classes of age, BMI, SBP and DBP at entry were found to be similar in both groups, and it was then concluded that comparisons could be made on parameters characterizing the elevation model of each factor; the attained level and the rate of change. It was also recognized that there were between-individual differences in the number of examinations, the length of exposure time and the time of exposure; and within-individual difference in the rate of change over different time periods. Furthermore, there was statistical artifact to be avoided. To consider and account for these problems, it was suggested that the cohort subjects be grouped retrospectively, and the rate of change in factors under study, be compared prospectively and separately in different time periods within classes classified according to the initial attained levels of all factors of interest. For simpler comparisons a frequency-adjustment method was also provided for summarizing the rates of change in time periods within each class.

The data showed that IHD/CVA group not only had higher initial attained levels of age, BMI, SBP and DBP than the reference group, but also had larger subsequent rate of change in both SBP and DBP before the manifestation of IHD/CVA. The BMI rate of change did not show apparent difference between the two groups. These differences were compared and observed in classes classified according to age, BMI, SBP and DBP at entry.

The differences on the rate of change in blood pressure were further supported by the findings that the incidence of "hypertension" was higher, and among the "hypertensives", the length of time to reach status of "hypertension" was shorter in the IHD/CVA group than in the reference group.

Multiple logistic function analyses showed that the initial attained levels of age, BMI, SBP and DBP were all significantly, positively and independently associated with the risk of "hypertension", and as a risk factor, the development of "hypertension" after entry was closely related to the development of IHD/CVA, independent of the initial attained levels of age, BMI, SBP and DBP.

Further evidence was presented to clarify the issue on the relationship between the attained level and the rate of change in blood pressure. It was concluded that the initial level and the rate of change in systolic blood pressure were two independent factors and they should be treated as such.