An epidemiology study examines how diseases and health conditions spread through populations, what causes them, and how to control them. The CDC defines epidemiology as “the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to the control of health problems.” It’s the science behind figuring out why some groups of people get sick and others don’t, then using those answers to protect public health.
What Epidemiologists Actually Do
Epidemiologists are essentially disease detectives. They look at patterns of illness across groups of people, sorted by time, place, and personal characteristics like age or occupation. But epidemiology isn’t purely academic. The whole point is to take what’s learned and turn it into practical public health action, whether that means identifying a contaminated water source, linking a habit to a disease, or evaluating whether a vaccine works.
The CDC outlines a ten-step process for field investigations that captures how this works in practice. Investigators first confirm that a real outbreak exists, then identify and count cases, organize the data by when and where people got sick, develop hypotheses about the cause, design systematic studies to test those hypotheses, and finally implement control measures and communicate findings. Not every epidemiology study follows an outbreak format, but the logic is the same: observe a pattern, form a theory, test it with data, and act on the results.
Observational vs. Experimental Studies
Epidemiological studies fall into two broad categories: observational and experimental. The distinction matters because it determines how strong the evidence is and what conclusions you can draw.
In observational studies, researchers watch what happens without intervening. They can’t ethically expose people to things suspected of being harmful, so they instead find people who are already exposed (smokers, for example) and compare their health outcomes to people who aren’t. The researcher’s role is to measure and analyze, not to change anyone’s behavior or treatment.
In experimental studies, the researcher actively controls who receives an intervention. The most rigorous version is the randomized controlled trial, where participants are randomly assigned to receive either a treatment or a placebo. Randomization, combined with techniques like blinding (where participants don’t know which group they’re in), protects against biases that could skew the results. These trials sit at the top of the evidence hierarchy because they’re the best tool for proving that a treatment actually works, rather than just appearing to work because of some other factor.
Common Observational Study Designs
Cohort Studies
A cohort study follows a group of people over time to see who develops a disease and who doesn’t, then looks back at what differed between the two groups. In a prospective cohort study, researchers recruit participants before anyone has developed the outcome and follow them forward. In a retrospective cohort study, researchers go back in time using existing records to reconstruct exposure histories for people whose outcomes are already known.
Prospective studies are generally considered stronger because the data is collected in real time, reducing the chance of gaps or errors. Retrospective studies rely on historical records that may be incomplete, and they’re more vulnerable to recall bias (where people misremember past exposures). That said, a well-conducted retrospective study can be more reliable than a poorly designed prospective one. The Framingham Heart Study, which has tracked cardiovascular disease risk factors in residents of Framingham, Massachusetts since the mid-20th century, is one of the most famous prospective cohort studies ever conducted. It helped establish the links between heart disease and factors like high blood pressure, cholesterol, and smoking.
Case-Control Studies
Case-control studies work in reverse. Researchers start with people who already have a disease (the cases) and compare them to similar people who don’t (the controls), then look backward to see which exposures differed between the two groups. Each control is matched to a case based on criteria like age or sex to make the comparison fair. The primary measure in these studies is the odds ratio, which estimates how much more likely the cases were to have been exposed to a particular risk factor compared to the controls.
These studies are especially useful for investigating rare diseases, since you can start with known cases rather than waiting for them to appear in a large population. They’re also faster and cheaper than cohort studies. The tradeoff is that they depend heavily on participants accurately remembering past exposures.
Cross-Sectional Studies
A cross-sectional study captures a snapshot of a population at a single point in time. It measures both exposures and health outcomes simultaneously, making it useful for estimating how common a condition is right now. What it can’t do is establish which came first, the exposure or the disease, so it’s limited in its ability to suggest cause and effect.
Key Measures: Incidence and Prevalence
Two numbers come up constantly in epidemiology, and they answer different questions. Incidence measures how many new cases of a disease appear during a specific time period. It tells you the rate at which people are getting sick. The formula divides the number of new cases by the total time the population was observed and at risk.
Prevalence measures how many people have a disease at a given moment (point prevalence) or over a given period (period prevalence), counting both new and pre-existing cases. A disease can have low incidence but high prevalence if people who get it live with it for a long time. Think of type 1 diabetes: relatively few new cases each year, but because it’s a lifelong condition, the total number of people living with it is large. Conversely, a rapidly fatal disease might have high incidence but low prevalence because patients don’t survive long enough to accumulate in the population count.
What Can Go Wrong: Bias and Confounding
Bias is any systematic error that pushes a study’s results away from the truth. It doesn’t mean the researchers were dishonest. It means something in the study design or data collection consistently tilted the results in one direction. There are several common types.
Selection bias happens when the people who end up in a study differ in important ways from those who don’t. If healthier people are more likely to volunteer for a study, the results may look more optimistic than reality. In workplace studies, a related problem called the healthy worker effect occurs because people who are employed tend to be healthier than the general population, making an occupational exposure look safer than it actually is. Loss to follow-up is another form: in long cohort studies, participants who drop out may be sicker (or healthier) than those who stay, distorting the findings.
Information bias comes from errors in how data is collected. Recall bias is a classic example in case-control studies. People who have a disease tend to think harder about past exposures than healthy controls do. Someone diagnosed with lung cancer may remember every instance of chemical exposure at work, while a healthy person asked the same questions might forget or dismiss similar experiences. Interviewer bias can creep in when the person asking questions knows the hypothesis and unconsciously steers responses. Social desirability bias occurs when participants underreport behaviors they find embarrassing, like alcohol consumption, or overreport behaviors they think make them look good, like exercise.
Confounding is a different problem entirely. A confounding variable is something that’s related to both the exposure and the outcome, creating a false appearance of a direct connection. A classic example: studies once suggested that coffee drinking was linked to heart disease. But coffee drinkers were also more likely to smoke. Smoking was the confounder, creating a misleading association between coffee and heart problems. Researchers handle confounding through study design (like randomization in trials) or statistical adjustment after the data is collected.
Landmark Studies That Shaped Public Health
Some of the most consequential public health discoveries came from epidemiological studies. In 1854, John Snow investigated a cholera outbreak in London by mapping cases and tracing them to a single water pump on Broad Street. He found that cholera patients all had one thing in common: they drank water from that pump. When officials removed the pump handle, the outbreak ended. Snow went further, comparing cholera death rates in districts served by two different water companies. One company had moved its water intake upstream from London’s sewage outflows; the other hadn’t. The difference in death rates between the two was striking, and it happened decades before the germ theory of disease was widely accepted.
In the mid-20th century, Richard Doll and Austin Bradford Hill used epidemiological methods to establish the link between smoking and lung cancer, a finding that eventually transformed public health policy worldwide. Around the same time, the Framingham Heart Study began tracking thousands of residents to identify cardiovascular risk factors. During the 1960s and 1970s, epidemiological methods were applied to eradicate naturally occurring smallpox globally, an achievement the CDC calls unprecedented in applied epidemiology.
How Study Quality Is Evaluated
Not all epidemiological studies are created equal, and the medical community has developed tools to assess their quality. The STROBE initiative (Strengthening the Reporting of Observational Studies in Epidemiology) is an international collaboration of epidemiologists, statisticians, and journal editors that created a checklist of items that should be included when reporting cohort, case-control, and cross-sectional studies. The goal is transparency: readers need to know what was planned, what was done, what was found, and what the results mean. STROBE doesn’t dictate how to design a study, but it sets expectations for how clearly and completely the results should be reported. Major medical journals endorse these guidelines, and incomplete reporting is a red flag when evaluating whether a study’s conclusions are trustworthy.