The word ‘epidemic’ means something that falls upon people, explains Djøra Soeteman, PhD, MA, of the Center for Health Decision Science at the Harvard T.H. Chan School of Public Health. ‘Epi’ means upon and ‘demos’ means people. This definition reflects one of the basic ideas of modern epidemiology, which is to look at diseases at the level of populations. Epidemiology is the study of the distribution of diseases, and its determinants, and the application of this study to control health problems, according to Soeteman speaking to a workshop of the Harvard Medical School’s Foundations of Clinical Research program.

In the mid-1800s, John Snow performed what would become the foundational epidemiological studies while identifying the source of local cholera outbreaks in London, England. His study methods and conclusions earned him the moniker of the Father of Epidemiology.

“John Snow started with a hypothesis about the source of a cholera outbreak and used data points to confirm it,” says Soeteman, whose research focuses on applied decision analytic methods to evaluate the clinical benefits and cost-effectiveness of health care interventions and policies. Dr. Soeteman also serves as an associate director of the Global Clinical Scholars Research Training program of the Harvard Medical School.

Snow evaluated his hypothesis with increasingly more rigorous study designs. “Without any knowledge of the existence of any microorganism, he demonstrated through these epidemiological studies that water actually could serve as a vehicle for transmitting cholera. And that information from his study could be used to take action and to improve public health in general.”

Today’s epidemiology has a broader mission from its origins in the 1800s. “Epidemiology is not only to a limited extent to the study of epidemics and infectious diseases,” Soeteman says, “In reality, it's a much broader area of study with implications in a wide variety of human disease.”

The focus now is on a robust understanding of the burden—or extent—of disease but also on how to control diseases and conditions.

Types of Epidemiological Study Designs

Epidemiological study designs fall into two broad categories: observational designs and experimental designs.

In an observational study, the researcher or the epidemiologist simply observes the exposure and the disease status, or outcome, of each study participant. “We know that the observational study designs are sensitive to bias and confounding,” says Soeteman. For example, selection bias is a danger when one chooses a particular population that will affect a certain outcome.

Observational study designs:

• Cross-sectional studies are “cheap and rapid” but less suitable for non-chronic diseases
  • Exposure and outcome are measured at the same time
  • The lack of time dimension limits the ability to make causal inference
  • Weaker than either cohort or a case-control study
• Ecological studies use aggregate data for two or more groups rather than for individuals
  • Uses variables in the group as potential determinants of disease
• Case control: A study that selects patients on the outcome or disease and compares the exposure in this group to a very similar group of people who do not have the outcome or disease
• Prospective cohort: A cohort is followed over time; exposure is measured before outcome
• Retrospective cohort: A cohort is followed over time; exposure is measured after outcome

In contrast, in experimental studies, the investigator determines the exposure for each individual through randomization. “The researcher has an active role in sending patients to different treatment arms,” adds Soeteman.

Regardless of the method, comparisons between different groups must be fair. In addition, the role of chance must be accounted for (statistics).

Core terms to understand:

• Ratio: A comparison of two different things. Can be for different categories of the same variable OR for different variables.
• Proportion: The numerator is a fraction or part of the denominator
• Rate: Has a crucial time dimension (e.g., incidence rate: new cases per year per 100,000 in the population)

Epidemiological Measures

According to Soeteman, the most basic task of epidemiology is to describe the occurrence of diseases and to note differences in the occurrence of diseases between different groups. To do so requires epidemiologic specific terms to quantify this disease burden.

• Frequency: How much disease is present? Includes prevalence (number of people who have the disease at a specific point in time) and incidence (number of people who get the disease over a specified period of time; number of new cases)
• Association: Association measures allow one to answer if is there more disease among the exposed compared to the non-exposed group? This measure involves risk difference, risk ratio, and odds ratio.
• Potential impact: How important is a certain risk factor to a disease or a certain outcome? Includes assessment of attributable fraction
• Measures of survival: What is the time interval to a particular outcome? Includes measures of absolute survival time or absolute risk reduction, relative risk reduction, the number needed to treat, progression, or event-free survival time

What Type of Epidemiological Trial to Choose?

“When starting to design your study for a particular research question, you have to select a study design based on this research question but also consider other factors that may influence the optimal choice,” says Soeteman, such as resources and data availability. “You always have to be vigilant about biases and confounders that may threaten the true relationship between the exposure and the outcome.”

Soeteman’s presentation represented just a portion of the information students taking the Foundation in Clinical Research can expect to learn over the course of the program. Just like John Snow lay the foundation of modern epidemiology through his early studies on cholera, the Foundations of Clinical Research program will develop the learner’s foundation to become skilled clinical researchers with all the epidemiological and biostatistical tools to design, conduct, and implement high-quality research studies.

Written by Alice McCarthy