Two major obstacles to the routine application of age-period-cohort models are (1) the identification problem, and (2) the fact that separate interpretation of the coefficients of the model is seldom possible. We offer a practical solution to these obstacles that involves plotting the relation between the variable of interest and the age, period, and cohort variables in such a manner that nontrivial age, period, or cohort effects are readily recognized as particular types of features in the graph. These features remain recognizable in the presence of normal sampling variability. Examples are given for applying the technique to previously published mortality data.
Very often criteria by which subjects are selected for epidemiological studies are associated in some manner with their health. The Healthy Worker Effect (HWE) or Healthy Person Effect (HPE) is well known. Little has been said about the converse case in which selection is associated with decreased health status, the Sick Person Effect (SPE). The SPE may introduce a bias for some cohort, most clinical follow-up, and some case-control studies when risks are standardized against an inappropriate referent. We demonstrate the existence of the SPE in two studies. Study 1 compares the incidence of a number of different diseases among individuals who were selected as children for medical treatment with that among their siblings. Study 2 computes the Standardized Morbidity Ratios (SmRs) for various acute and chronic diseases for individuals who have reported particular chronic symptoms. The SPE is clearly apparent for all instances where the general population is taken as the referent. The HPE and SPE may present serious problems for the validity of conclusions with respect to risk levels.
Patterns of prevalence, amount, and cessation of smoking are computed for occupations by socioeconomic class, sex, and race, based on a probability sample of 39,011 households collected by the 1970 Health Interview Survey. Smoking is most prevalent in blue-collar occupations, while a high proportion of professionals and managers who smoke, stop smoking. Within industries, substantially higher percentages of individuals smoke in lower prestige paying jobs, while more smokers quit in the higher prestige paying jobs. Smoking is most prevalent among women managers and professionals, and least among those employed in traditional work. One surprising and possibly very important finding is that white smokers smoke about 20% more cigarettes per day than black smokers. Not only would it seem unreasonable to ascribe the larger rate of lung disease among blacks than whites (especially cancer), to smoking when blacks smoke significantly fewer cigarettes than whites, but this same negative relationship points to occupational exposure as the possible major cause for lung cancer.