Most epidemiological investigations of aetiology are observational. They look for associations between the occurrence of disease and exposure to known or suspected causes. In ecological studies the unit of observation is the population or community. Disease rates and exposures are measured in each of a series of populations and their relation is examined. Often the information about disease and exposure is abstracted from published statistics and therefore does not require expensive or time consuming data collection. The populations compared may be defined in various ways.
One common approach is to look for geographical correlations between disease incidence or mortality and the prevalence of risk factors. For example, mortality from coronary heart disease in local authority areas of England and Wales has been correlated with neonatal mortality in the same places 70 and more years earlier. This observation generated the hypothesis that coronary heart disease may result from the impaired development of blood vessels and other tissues in fetal life and infancy.
Many useful observations have emerged from geographical analyses, but care is needed in their interpretation. Allowance can be made for the potential confounding effects of age and sex by appropriate standardisation.
More troublesome, however, are the biases that can occur if ascertainment of disease or exposure, or both, differs from one place to another. For example, a survey of back disorders found a higher incidence of general practitioner consultation for back pain in the north than the south of Britain, which might suggest greater exposure to some causative agent or activity in the north. Closer investigation, however, indicated that the prevalence of back symptoms was similar in both regions and that it was patients' consultation habits that varied. Thus, in this instance correlations based on general practitioner consultation rates would be quite misleading. A study based on rates of admission to hospital for perforated peptic ulcer would probably be reliable as in affluent countries almost all cases will reach hospital and be diagnosed. On the other hand, unbiased ascertainment of disorders such as depression or Parkinson's disease may be difficult without a specially designed survey. When there is doubt about the uniformity of ascertainment, it may be necessary to explore the extent of any possible bias in a validation exercise.
Many diseases show remarkable fluctuations in incidence over time. Rates of acute infection can vary appreciably over a few days, but epidemics of chronic disorders such as lung cancer and coronary heart disease evolve over decades. If time or secular trends in disease incidence correlate with changes in a community's environment or way of life then the trends may provide important clues to aetiology. Thus, the currently increasing incidence of melanoma in Britain has been linked with greater exposure to sunlight (from fashions in dress and holidays abroad); and successive rises and falls in mortality from cervical cancer have been related to varying levels of sexual promiscuity, as evidenced by notification rates for gonorrhoea.
Like geographical studies, analysis of secular trends may be biased by differences in the ascertainment of disease. As health services have improved, diagnostic criteria and techniques have changed. Furthermore, whereas in geographical studies the differences are accessible to current inquiry, validating secular changes is more difficult as it depends on observations made and often scantily recorded many years ago. Nevertheless, the reality - if not the true size - of secular trends can often be established with reasonable certainty. The rise and subsequent fall in the incidence of appendicitis in Britain during the past 100 years is a good example.
The study of migrant populations offers a way of discriminating genetic from environmental causes of geographical variation in disease, and may also indicate the age at which an environmental cause exerts its effect. Second generation Japanese migrants to the USA have substantially lower rates of stomach cancer than Japanese people in Japan, indicating that the high incidence of the disease in Japan is environmental in origin. In first generation migrants rates are intermediate, which suggests that the adverse environmental influences act, at least in part, early in life.
In interpreting migrant studies it is important to bear in mind the possibility that the migrants may be unrepresentative of the population that they leave, and that their health may have been affected directly by the process of migration. Norwegian immigrants into the USA, for example, have been found to have a higher incidence of psychosis than people in Norway. Although this may indicate environmental influences in the USA that led to psychotic illness, it may also have resulted from selective emigration from Norway of people more susceptible to mental illness, or from the unusual stresses imposed on immigrants during their adjustment to a foreign culture.
Despite these difficulties, migrant studies have contributed importantly to our understanding of several diseases.
The other populations for whom statistics on disease incidence and mortality are readily available are occupational and socioeconomic groups. Thus, mortality from pneumonia is high in welders, and the steep social class gradient in mortality from chronic obstructive lung disease is evidence that correlates of poverty, perhaps bad housing, have an important influence on the disease.