Study of Familial Patterns of Reported Diabetes: Evaluation of Questionnaire Data

  1. Franklin W Briese, PhD
  1. University of Maryland, Institute of International Medicine, Division of Epidemiology, Biostatistics Baltimore, Maryland University of Minnesota, School of Public Health, Division of Biostatistics Minneapolis, Minesota

Abstract

This is a study, based on a family history survey for diabetes obtained by questionnaire from 31,046 students who entered the University of Minnesota during the five academic years from 1955–56 to 1959–60. For each individual reported in the questionnaire, the primary item of information obtained was the diabetes status, classified where possible according to parental mating group (i.e., both parents diabetic, one parent diabetic, neither parent diabetic).

The main purpose of the study, in the planning and initial operational stages, was to provide data adequate for testing the correlation of diabetes prevalence in offspring of parental mating groups with the number of diabetic parents and for testing single-gene genetic hypotheses. The initial data collected were on “reported diabetes”—a rather ill defined entity—and were besides, as became clear later, subject to some fairly gross errors of misclassification. The study plan included procedures for rectifying these difficulties, but it was not possible to carry this work through. As a consequence, it was felt that the data would not support the statistical analysis necessary to test the questions mentioned above.

There are, however, numerous familial patterns suggested by various tabulations of the data which, interpreted with caution, are of some interest. The following trends appearworth noting:

1. The average number of offspring per family does not appear to be correlated with mating subgroup, nor does it appear to be correlated with the presence of diabetes among first degree relatives.

2. Prevalence of diabetes appears to be positively correlated with the number of diabetic parents. In fact, the observed prevalence in families with one diabetic parent is about three times the prevalence in families with neither parent affected. This effect appears to persist even when the data are subclassified by parental age.

3. Mortality rates do not appear to be correlated with the number of diabetic parents. Sources of error inherent in the present study, in particular, and the retrospective questionnaire-based study, in general, are extensively analyzed and discussed.

| Table of Contents