Epidemiology: What It Is, What It Studies, and Why It Matters

What is epidemiology?

Epidemiology is the science that studies the frequency, distribution, and determinants of diseases in human populations. The term comes from the Greek epí (upon), dêmos (people), and lógos (study): literally, the study of what happens upon the people.

Unlike clinical medicine, which deals with the individual patient, epidemiology reasons about groups of people — hundreds, thousands, sometimes millions — to answer three fundamental questions: who gets sick, where they get sick, and why they get sick. The answers to these questions guide prevention strategies, health policies, and ultimately the decisions every doctor makes in their own practice.

Epidemiology is not an abstract discipline confined to statistics laboratories. Every time a doctor prescribes a drug based on a clinical study, interprets a risk factor, or assesses the effectiveness of a screening program, they are applying epidemiological findings to everyday practice. In my experience as an anesthesiologist, epidemiology guides concrete decisions: the choice of perioperative antibiotic prophylaxis, the stratification of thromboembolic risk before surgery, the decision about which monitoring to adopt in intensive care — all of it rests on epidemiological data.

What epidemiology is for

Epidemiology is a cross-cutting science that touches every field of medicine. It is precisely from epidemiological studies that the medical community is able to draw up therapeutic guidelines and make rational decisions about the treatment and prevention of diseases. As a result, the ultimate goal of epidemiology is to improve the health status of populations.

Epidemiology pursues five main practical aims.

The first is to investigate diseases that are still poorly characterized, asking questions such as: what circumstances favor the disease? Why do some people fall ill while others remain unaffected? This is the domain of etiology, which draws its initial hypotheses from epidemiology.

The second is to trace the origin of a disease: how it spread in a particular territory, and why it may have made a species jump, as happens in some zoonoses.

The third is to gather information on the natural history of the disease, in order to mitigate its main causes and promote prevention.

The fourth is to plan disease-control programs, identifying the most effective strategies to combat them — from vaccination programs to population screening.

The fifth is to analyze the risk-benefit ratio of each health intervention, also assessing its economic impact. This type of analysis is what allows a healthcare system to allocate resources where they produce the greatest benefit.

What epidemiology studies

Epidemiology is organized into three broad, complementary areas, each with a distinct role.

Descriptive epidemiology answers the questions who, where, and when. It measures the frequency of diseases in a defined population, and their distribution by age, sex, geographic area, and time period. Its main tools are incidence and prevalence rates and mortality rates. When you read that the incidence of type 2 diabetes is rising in Western countries, or that lung cancer is more common in men over 60, you are citing descriptive epidemiology data.

Analytical epidemiology answers the question why. It looks for causal links between exposures (risk factors, infectious agents, behaviors) and diseases. It uses specific study designs — cohort studies, case-control studies, and randomized clinical trials — to determine whether an observed association is truly causal or merely a statistical coincidence. The discovery that smoking causes lung cancer, that Helicobacter pylori causes gastric ulcers, that asbestos causes mesothelioma: these are all achievements of analytical epidemiology.

Experimental epidemiology tests interventions. Randomized clinical trials — the most solid reference point in medical research — are experimental epidemiological studies: participants are randomly assigned to a treatment or a placebo, and the outcome is measured. It is the method by which we prove that a drug works, that a vaccine protects, that a surgical procedure is superior to an alternative.

The main types of epidemiological study

Each type of study has a specific role in the hierarchy of evidence.

The cross-sectional study photographs a population at a single moment. It measures the prevalence of a condition and its associations with other factors. It is quick and inexpensive, but it cannot establish the direction of causation: if we observe that depressed patients have lower vitamin D levels, we cannot conclude whether vitamin D deficiency causes depression or vice versa.

The case-control study compares people who have already developed a disease (cases) with similar people who have not (controls), and retrospectively investigates past exposures. It is the ideal tool for rare diseases: if a condition affects one person in a hundred thousand, waiting for it to develop in a prospective cohort would be impractical. It was with a case-control study that Richard Doll and Bradford Hill demonstrated the association between smoking and lung cancer in 1950.

The cohort study follows over time a group of people exposed to a factor and a group that is not exposed, measuring who develops the disease. It is prospective by nature: the exposure is observed first, then the outcome. The Framingham Heart Study, launched in 1948 and still active today, is the most famous cohort study in the history of medicine: it identified the main cardiovascular risk factors — hypertension, cholesterol, smoking, diabetes, obesity — by following thousands of residents of the town of Framingham, Massachusetts, for over seventy years.

The randomized clinical trial is the only type of study capable of demonstrating a causal link definitively. Random assignment to treatment eliminates confounding factors: the differences observed between the two groups are attributable to the intervention and not to pre-existing characteristics of the patients. The pathophysiology of a disease may suggest a treatment, but only the randomized trial can prove that the treatment works in clinical reality.

The systematic review with meta-analysis collects and statistically synthesizes the results of several studies on the same topic. It sits at the top of the pyramid of evidence: when a meta-analysis of ten randomized studies confirms the effectiveness of a treatment, the conclusion is more robust than that of any single study.

The fundamental measures in epidemiology

Epidemiology quantifies disease through precise measures that have distinct meanings.

Incidence measures new cases in a defined period. It is expressed as a rate: the number of new cases divided by the population at risk per unit of time. An incidence of 5 cases per 100,000 person-years means that, in a population of one hundred thousand people followed for one year, 5 new cases are expected.

Prevalence measures all existing cases at a given moment. It includes both new cases and pre-existing ones. A chronic disease such as diabetes has a much higher prevalence than its incidence, because those who fall ill remain ill for years. An acute and rapidly fatal disease, on the other hand, may have high incidence but low prevalence.

Relative risk compares incidence between the exposed and the unexposed. A relative risk of 3 for lung cancer in smokers means that smokers have a three times higher risk than non-smokers. It is the most intuitive measure of the association between exposure and disease.

The odds ratio is the counterpart of relative risk used in case-control studies, where relative risk cannot be calculated directly. For rare diseases, the odds ratio closely approximates relative risk.

The Number Needed to Treat translates statistical results into clinical practice: it indicates how many patients must be treated to prevent one adverse event. An NNT of 20 means that for every 20 patients treated, one will avoid the event. It is a measure I use when communicating with patients: it makes the expected benefit of a treatment tangible.

The relationship between epidemiology and clinical medicine

Epidemiology and clinical practice feed one another.

The etiology of a disease is often discovered through epidemiological studies even before the biological mechanism is understood. John Snow identified contaminated water as the cause of cholera in 1854 London — thirty years before Robert Koch isolated Vibrio cholerae. Epidemiology had already provided the practical solution (closing the contaminated water pump) while microbiology was still searching for the cause.

Diagnosis itself is an epidemiological act. When a doctor takes the medical history and assesses the signs and symptoms during the clinical examination, they are implicitly estimating the pre-test probability of each diagnostic hypothesis — a probability that depends on the prevalence of the disease in the reference population. Chest pain in a 65-year-old man with diabetes and hypertension has a very different pre-test probability of acute coronary syndrome than the same pain in a healthy 25-year-old woman.

Evidence-based medicine is, in essence, the systematic application of epidemiology to clinical decisions: integrating the best available evidence with the doctor's experience and the patient's preferences.

Epidemiology and public health

Epidemiology is the foundation of public health. Screening programs, vaccination campaigns, and food-safety and environmental regulations all rest on epidemiological data.

Epidemiological surveillance — the continuous monitoring of disease frequency — makes it possible to identify epidemics at an early stage and to assess the effectiveness of interventions. Surveillance systems play a crucial role in infectious diseases: it is thanks to surveillance that influenza outbreaks are detected, antibiotic resistance is monitored, and the variants of emerging pathogens are tracked.

Epidemiology also contributes to individual prevention. When a doctor advises a patient to quit smoking, to keep their blood pressure under control, or to exercise, they are translating epidemiological findings into personalized recommendations. The strength of the recommendation depends on the quality of the epidemiological evidence that supports it.

Frequently asked questions

What is the difference between epidemiology and etiology?

Etiology studies the causes of diseases in the individual organism. Epidemiology studies how diseases are distributed across populations and which factors determine their frequency. Etiology asks why a patient became ill; epidemiology asks why more people fall ill in one population than in another. The two disciplines are complementary: it is often epidemiology that discovers the association between a factor and a disease, and then etiological research that explains the mechanism.

What are incidence and prevalence in simple terms?

Incidence counts new cases of illness over a period (how many people fall ill tomorrow). Prevalence counts all cases at a given moment (how many people are ill today). A disease that resolves quickly has high incidence but low prevalence; a chronic disease accumulates prevalence over time even with modest incidence.

Is epidemiology useful to the patient too, or only to the doctor?

It is useful to both. The informed patient understands why the doctor recommends certain screenings and not others, why some risk factors matter more than others, and what weight to give to statistics when making a decision about their own health. Understanding that a relative risk of 2 does not mean certainty of becoming ill, but a doubling of the probability, is a skill that improves the quality of shared decisions between doctor and patient.

References

Doll R, Hill AB. Smoking and carcinoma of the lung: preliminary report. Br Med J. 1950;2(4682):739-748. PubMed

Dawber TR, Meadors GF, Moore FE. Epidemiological approaches to heart disease: the Framingham Study. Am J Public Health. 1951;41(3):279-286. PubMed

Sackett DL, Rosenberg WM, Gray JA, et al. Evidence based medicine: what it is and what it isn't. BMJ. 1996;312(7023):71-72. PubMed

Rothman KJ, Greenland S. Causation and causal inference in epidemiology. Am J Public Health. 2005;95 Suppl 1:S144-S150. PubMed