Welcome to our next lesson. In our previous session, we established that cohort studies are powerful because they follow individuals over time, allowing us to observe who develops a disease. This temporal structure is what enables us to move from simply describing disease frequency to quantifying the association between an exposure and an outcome.
Today, we will build directly on that foundation. Your goal is to learn how to identify, calculate, and interpret the two primary measures of association used in cohort studies: the risk ratio and the rate ratio. These metrics are the quantitative heart of a cohort study, telling us how much more (or less) likely an exposed group is to develop an outcome compared to an unexposed group.
From Data to Insight: The 2x2 Table
At the conclusion of a cohort study, after all participants have been followed, the raw data needs to be structured for analysis. The standard format for this is the 2x2 contingency table, a simple yet powerful tool for organizing count data.
You can think of this as analogous to creating a pivot table from a raw log file in data analysis. We start with a long list of individuals and their final status, and we summarize it into a compact grid that shows the relationship between two categorical variables: exposure status (Yes/No) and outcome status (Yes/No).
The conventional layout is as follows:
| Disease (D+) | No Disease (D-) | Total | |
|---|---|---|---|
| Exposed (E+) | A | B | A+B |
| Unexposed (E-) | C | D | C+D |
| Total | A+C | B+D | A+B+C+D |
- A: Number of exposed individuals who developed the disease.
- B: Number of exposed individuals who did not develop the disease.
- C: Number of unexposed individuals who developed the disease.
- D: Number of unexposed individuals who did not develop the disease.
The following reading explains how a cohort study's results naturally lead to this table.
Introduction to 2 x 2 Tables, Epidemiologic Study Design, and ...
This resource from Oregon State University's epidemiology textbook illustrates how a cohort begins with a disease-free sample and ends with a completed 2x2 table after a follow-up period.
Start at the heading Studies That Use Incidence Data and read through the description of a cohort study. Pay attention to how the study begins with a sample where the "D+" (diseased) column is empty, as shown in Table 4-6. Then, see how after the follow-up period, the data populates the final 2x2 table, as shown in Table 4-7.
With our data organized, we can now calculate the measures of association.
Measuring Association with the Risk Ratio
The first and most intuitive measure is the risk ratio, often called the relative risk. It directly compares the risk of developing the disease in the exposed group to the risk in the unexposed group.
First, we calculate the risk for each group separately. In epidemiology, this risk is formally known as cumulative incidence: the proportion of a group that develops the outcome over a specified time period.
- Risk in the exposed group (IE+):
- Risk in the unexposed group (IE-):
The risk ratio is simply the ratio of these two risks:
The following video provides an excellent visual and conceptual walkthrough of this calculation, focusing on the logic rather than just the formula.
This video from "Global Health with Greg Martin" breaks down the concept of risk and demonstrates how to calculate and interpret the risk ratio using a clear, step-by-step example.
Watch the entire video. Pay attention to these key segments: The explanation of risk in a cohort and how it's calculated separately for exposed and unexposed groups. The calculation of the risk ratio and its initial interpretation. The presenter does a great job of showing how it's just a simple division of the two risks. The interpretation of the final value: what it means if the RR is 1, greater than 1, or less than 1. The application to a real-world study on smoking and low birth weight, which solidifies the concept.
As the video explained, the interpretation of the risk ratio hinges on the number 1.0, which is the null value (meaning no association).
- RR > 1: The exposure is associated with an increased risk of the outcome. (e.g., RR = 2.0 means the exposed group has twice the risk of the unexposed group).
- RR < 1: The exposure is associated with a decreased risk of the outcome; it is a "protective" factor. (e.g., RR = 0.5 means the exposed group has half the risk of the unexposed group).
- RR = 1: There is no difference in risk between the two groups.
A precise way to state the interpretation is: "The risk of [the outcome] was [RR value] times as high in the [exposed group] compared to the [unexposed group] over the study period."
Accounting for Time: Person-Time and the Rate Ratio
The risk ratio works perfectly when all individuals in the cohort are followed for the same amount of time. But what happens in a long-term study where people enter at different times, drop out, or are lost to follow-up? A person followed for 1 year does not contribute the same amount of information as a person followed for 20 years.
To handle this, we use the concept of person-time. Instead of just counting people in the denominator, we sum up the total time each individual was at risk and under observation. This gives us an incidence rate, which measures the speed at which new cases occur.
- Unit of measurement: person-years, person-months, etc.
- Calculation: Incidence Rate = (Number of new cases) / (Total person-time at risk)
This is similar to how a system administrator might measure server failures not just by the number of crashes, but by the rate of crashes per 1000 hours of operation, accounting for the fact that different servers have different amounts of uptime.
This short video clearly explains the concept of person-time and its use in calculating rates.
This video by "prof smith" introduces why person-time is needed and demonstrates how to calculate it to find an incidence rate.
Watch the segment from the beginning to the explanation of person-years. This will show you how individual follow-up times are summed to create the denominator for an incidence rate.
Once we have the incidence rates for our exposed and unexposed groups, we can calculate the rate ratio, which is conceptually identical to the risk ratio but uses rates instead of risks.
- Incidence Rate in exposed:
- Incidence Rate in unexposed:
The interpretation is also very similar: "The rate of [the outcome] was [RR value] times as high in the [exposed group] compared to the [unexposed group]."
The following image provides a helpful summary of the formulas for these key measures.
Risk Ratio vs. Rate Ratio: Which to Use?
You may have noticed that both the risk ratio and the rate ratio are often abbreviated as "RR" and sometimes referred to collectively as "relative risk." While they are conceptually similar, they are derived differently and answer slightly different questions.
- Risk Ratio (from Cumulative Incidence): Answers "What is the total risk of developing the disease over this period?" It's best for "closed" cohorts where everyone is followed for the full duration.
- Rate Ratio (from Incidence Rate/Density): Answers "How quickly are new cases of the disease appearing?" It's necessary for "dynamic" cohorts with variable follow-up times.
This reading clarifies the distinction and provides examples for calculating both.
Introduction to 2 x 2 Tables, Epidemiologic Study Design, and ...
This part of the textbook explicitly contrasts the two types of ratios.
Please read the section on calculating the rate ratio, which shows how a person-time denominator is used. Then, read the short section titled "Relative Risk", which explains the overlapping terminology.
For some hands-on practice, this next resource provides concise examples and a couple of practice problems.
[PDF] Risk and Rate Measures in Cohort Studies
This document from the University of North Carolina gives clear, worked examples for both risk ratios and rate ratios, and then provides problems for you to solve.
First, review the definitions and examples under the headings Risk ratios and Rate ratio. Then, read the paragraph on interpretation. Finally, try your hand at the two problems under Practice Questions. You can check your work against the Answers to Practice Questions at the end of the document.
Conclusion
In this lesson, we moved from the design of cohort studies to their analysis. You learned how to quantify the strength of an association between an exposure and an outcome using the two key measures derived from cohort studies.
Here are your key takeaways:
- Risk Ratio (Relative Risk): Compares the cumulative incidence (risk) in an exposed group to an unexposed group. It's calculated as .
- Rate Ratio: Compares the incidence rates in an exposed group to an unexposed group, using person-time to account for variable follow-up. It's calculated as .
- Interpretation: For both measures, a value of 1.0 means no association, >1.0 means increased risk/rate, and <1.0 means decreased risk/rate (a protective effect).
In our last two lessons, we've focused on study designs that start with an exposure and look for an outcome. However, this isn't always feasible, especially for rare diseases. In our next lesson, we will explore the case-control study, a design that works in reverse: it starts with the outcome (cases) and looks backward for the exposure. This different logic will require a different measure of association: the odds ratio.