Epidemiology is the study of the distribution of disease in human populations and the factors that determine that distribution. There are two main types of epidemiologic research descriptive and analytic. Descriptive studies focus on the incidence of disease regarding person, place, and time. Such studies deal with identifying the characteristics of persons most susceptible, such as age, sex, race, ethnic group, as well as defining geographic variation in incidence and changes in incidence over time. Analytic epidemiology investigates specific etiologic factors such as exposures to carcinogens and family history of cancer. Wilms tumor has been studied by both descriptive and analytic methods.
In order to determine the incidence of a disease in a geographic region one needs to ascertain all instances of disease occurring in a defined geographic area as well as the population at risk, that is, of the area. The Greater Delaware Valley (GDV) surrounding Philadelphia is such a defined region; it has a population base in which the distribution by age, race, sex, and socioeconomic status is known. The Greater Delaware Valley Pediatric Tumor Registry has collected data on all childhood cancers in the GDV since 1970 [l].
Descriptive epidemiology is useful in providing clues to etiology, an understanding of the latter being the ultimate purpose of epidemiologic studies. The clues serve as the basis of more detailed analytic studies. Several known or suspected environmental causes of Wilms tumor such as radiation, drugs, tobacco, alcohol, and the occupational exposures of parents have been considered as etiologic factors. In addition one might suspect postnatal viruses, dietary factors, pesticides, or electromagnetic fields. These appear to be the favorite culprits in modern epidemiologic studies.
How would one go about studying the effect of specific risk factors on the etiology of disease? One could adopt a prospective, cohort approach or use a case-control method. In the cohort method one would follow children with a given characteristic, such as aniridia or hemihypertrophy, and compare their incidence of Wilms tumor to that of children without aniridia or hemihypertrophy. The advantages of such a method are (1) a straightforward study design, and (2) the ability to directly assess the exposure characteristic of interest before the disease occurs. Unfortunately, because of the rarity of the disease, in order to detect any differences between the