Concept Architecture
Concept
Theoretically, Cohort Design is an observational epidemiological study design in which a defined group of individuals is followed over time to compare the occurrence of outcomes between exposed and unexposed populations. The concept is founded on longitudinal observational research, probability theory and causal inference. It exists to estimate the incidence of disease, quantify associations between exposures and outcomes, and investigate temporal relationships that cannot be established using cross-sectional study designs.
Mathematically, Cohort Design is represented by comparing outcome incidence between exposure groups using measures such as cumulative incidence, incidence rates, relative risk and hazard ratios. Time-to-event cohort studies frequently employ survival analysis, including Kaplan?Meier estimation and Cox proportional hazards regression, while regression models are used to adjust for confounding variables. These methods enable estimation of both crude and adjusted measures of association.
In practice, Cohort Design may be prospective, with participants followed from exposure to outcome, or retrospective, using existing clinical records, registries or administrative databases. Participants are classified according to exposure status before outcome assessment, and follow-up continues until the outcome occurs, loss to follow-up or study completion. Cohort studies are widely used in health economics to estimate disease incidence, treatment effectiveness, healthcare utilisation and epidemiological parameters for decision-analytic models.
Purpose
Used to estimate the association between exposures and subsequent outcomes by following defined populations over time, providing incidence estimates and measures of treatment or risk that inform clinical, epidemiological and health economic decision-making.
Mathematical Formulae
Primary Formula
Relative Risk:
RR = [a / (a + b)] � [c / (c + d)]
where:
- a = exposed individuals with the outcome
- b = exposed individuals without the outcome
- c = unexposed individuals with the outcome
- d = unexposed individuals without the outcome
Supporting Formulae
Incidence Rate:
IR = Events / Person-Time
Hazard Function:
h(t) = lim(?t?0) P(t � T < t + ?t � T � t) / ?t
Cox Proportional Hazards Model:
h(t�X) = h?(t) ? exp(??X? + ??X? + ? + ??X?)
Related Mathematical Methods
- Relative Risk Estimation
- Incidence Rate Estimation
- Kaplan?Meier Estimation
- Cox Proportional Hazards Regression
- Poisson Regression
- Survival Analysis
- Maximum Likelihood Estimation
Example
A retrospective cohort study evaluates the effectiveness of a new diabetes management programme. Among 1,200 participants receiving the programme, 60 experience a cardiovascular event during follow-up. Among 1,200 comparable participants receiving usual care, 90 experience the event.
Risk in exposed group = 60 / 1,200 = 0.05
Risk in unexposed group = 90 / 1,200 = 0.075
RR = 0.05 � 0.075 = 0.67
The programme is associated with a 33% reduction in the risk of cardiovascular events. This estimate may subsequently be incorporated into a cost-effectiveness model.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| COUNTIFS | =COUNTIFS(B:B,"Exposed",C:C,"Event") | Count outcome events in the exposed cohort. |
| COUNTIFS | =COUNTIFS(B:B,"Unexposed",C:C,"Event") | Count outcome events in the unexposed cohort. |
| IFERROR | =IFERROR((D2/E2)/(F2/G2),"") | Calculate the relative risk. |
| SUM | =SUM(H2:H1201) | Calculate total person-time of follow-up. |
| AVERAGE | =AVERAGE(I2:I1201) | Calculate average follow-up duration. |
VBA (Optional)
VBA can automate cohort summaries, calculate incidence measures and generate comparative risk estimates for longitudinal observational studies.
Sources
- Rothman KJ, Greenland S, Lash TL. Modern Epidemiology.
- Kleinbaum DG, Klein M. Survival Analysis: A Self-Learning Text.
- Hosmer DW, Lemeshow S, May S. Applied Survival Analysis.
- Harrell FE. Regression Modeling Strategies.
- NICE. Health Technology Evaluation Manual.
- Drummond MF, et al. Methods for the Economic Evaluation of Health Care Programmes.
Related Concepts (2)
Library
Publications
1
Good Practices for Real-World Data Studies of Treatment and/or Comparative Effectiveness: Recommendations from the Joint ISPOR-ISPE Special Task Force on Real-World Evidence in Health Care Decision Making — Berger, Sox, Willke, Brixner, Eichler, Goettsch, Madigan, Makady, Schneeweiss, Tarricone, Wang, Watkins & Mullins, Vol. 20, No. 8 ed., 2017 (Value in Health)
The joint ISPOR-ISPE recommendations on good procedural practice for real-world data studies (observational studies and registries) used to inform healthcare decisions — study registration, replicability and stakeholder involvement — the reference for RWE credibility in HTA.
Journal ArticleView source →
Frequently Asked Questions (6)
What is a cohort design?
An observational design following a group sharing a common starting characteristic forward in time to observe a subsequent outcome of interest.
Source: Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. 3rd ed. Lippincott Williams & Wilkins; 2008.
Why does a cohort design follow people forward in time?
A cohort study begins with a group defined by a shared characteristic or exposure and follows them forward to see who develops the outcome, so it observes cause before effect in the order they occur. This forward direction lets it measure how often the outcome arises in the exposed and unexposed, and study several outcomes of one exposure at once. Its cost is that rare or slow outcomes may need large numbers and long follow-up. It watches the future unfold rather than reconstructing the past. Rothman and colleagues (2008) describe this design.
Source: Rothman et al. 2008
How does a cohort study work?
A cohort study works by defining a group based on a starting characteristic, typically an exposure, and following the participants over time to record who develops the outcome of interest, then comparing the incidence of the outcome between exposed and unexposed groups. This yields measures such as relative risk. The follow-up may be prospective, tracking participants forward, or retrospective, reconstructing it from records. Because it starts with exposure and observes outcomes as they occur, the cohort design can establish the time order of exposure and outcome and measure incidence directly.
Source: Rothman, Greenland & Lash 2008
What are the advantages of a cohort design?
The advantages of a cohort design include its ability to establish the temporal sequence, since exposure is ascertained before the outcome occurs, strengthening causal inference; its capacity to measure incidence and the risk of the outcome directly; its suitability for studying multiple outcomes of a single exposure; and its usefulness for rare exposures, which can be sampled specifically. Prospective cohorts can collect high-quality exposure data before outcomes are known, reducing recall bias. These advantages make the cohort design valuable for studying the effects of exposures over time.
Source: Rothman, Greenland & Lash 2008
What are the limitations of a cohort design?
The limitations of a cohort design include that prospective cohorts can be lengthy and costly, especially for rare or slowly developing outcomes requiring long follow-up and large samples; loss to follow-up can introduce bias; and, as an observational design, it is subject to confounding, since exposure is not randomised, so measured and unmeasured confounders can distort associations. Retrospective cohorts depend on the quality of existing records. These limitations mean cohort studies require attention to follow-up, confounding, and data quality, and their causal conclusions are weaker than those from randomised trials.
Source: Friedman, Furberg & DeMets 2015
How does a cohort design differ from a case-control design?
A cohort design differs from a case-control design in direction and starting point: a cohort study starts with exposure and follows participants forward to observe outcomes, while a case-control study starts with the outcome, comparing prior exposures in those with and without it, looking backward. Cohort studies can measure incidence, establish time order clearly, and study multiple outcomes, but are less efficient for rare outcomes, whereas case-control studies efficiently study rare outcomes and yield odds ratios. So the two differ in efficiency for different situations and in the measures and inferences they support.
Source: Rothman, Greenland & Lash 2008
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 13 Nov 2025
Content version: 1.0.0
Canonical Identity
- Persistent URI
- https://healtheconomics.wiki/concept/cohort-design
- Term code
- HE-ES-CTM-016
Stable URI · Machine-readable · Resolvable · CC BY 4.0