Why Study Populations?
Keyes and Galea make the point that individuals become ill or not, but it is the study of populations that will be most helpful in identifying the causes of health and disease. If we were to carefully compare an individual with high blood pressure to an individual from the same population without high blood pressure, we might gain clues about why hypertension occurred in one but not in the other. If we were to take a larger sample of people from a population and identify those with hypertension and those without, we would likely obtain additional insights about the causes of hypertension. However, if we were to obtain samples of people from two or three different populations, we might find that the populations differed in the frequency of hypertension, and this provides an opportunity to obtain a more complete understanding of the causes of hypertension.
Keyes and Galea cite a paper by Geoffrey Rose comparing the distribution of systolic blood pressures in middle aged men in two populations: Kenyan nomads and London civil servants.
Source: Rose, G. (1985). "Sick individuals and sick populations." Int J Epidemiol 14(1): 32-38. As cited by Keyes and Galea.
Valuable information might be obtained by studying the causes for variation in blood pressure within each population, but greater insight would come from also understanding what caused the differences between the two populations. The two distributions are similar in shape, but the entire population of London civil servants is shifted upward from the Kenyan population, including the civil servants in the lower tail of the London population. One is then forced to consider characteristics that differ between populations in order to achieve a more complete understanding of the causes of hypertension.
Keyes and Galea conclude:
"... a focus on individual cases may render the scientist unable to identify and prevent the causes of disease that are most important in terms of health of the entire population. Epidemiologists and other public health practitioners are often utilitarian in our approach to population health by asking what causes we can intervene upon to produce the best health for the greatest number of people in the population. To do so requires studying not individuals with good or poor health, but rather the overall incidence and prevalence of health indicators across groups, asking why poor health is more common in some groups than in others. Thus, a central principle in epidemiology, and the first step in our epidemiologic rubric, is to define a population or group of scientific or public health interest."
[From Epidemiology Matters - A New Introduction to Methodological Foundations by Katherine M. Keyes and Sandro Galea, Oxford University Press, 2014]
Seven Steps to Answer Public Health Questions
As a general outline of how one addresses questions like these, Keyes and Galea list seven general steps.
You can also download the seven steps as a word file.