Cohort Study
Introduction
Epidemiology aims to identify the causes and determinants of diseases in populations. To understand whether a particular exposure leads to a specific outcome, we must observe individuals over time and compare the occurrence of disease among exposed and non-exposed groups.
Consider a hypothetical situation: In a village of 2,000 people, some individuals smoke and others do not. After 10 years of observation, 40 smokers develop lung cancer compared to 5 non-smokers. This raises important questions:
- Does smoking increase the risk of lung cancer?
- How much greater is the risk among smokers?
- Which study design allows us to follow individuals forward in time to establish this relationship?
To answer such questions, we use a cohort study, which follows groups defined by exposure status and observes the occurrence of disease over time .
Types of Epidemiological Studies
Epidemiological studies are broadly classified into observational and experimental designs.
1. Observational Studies
These studies observe events without intervention.
a) Descriptive Studies
- Case report
- Case series
- Cross-sectional study
- Ecological study
b) Analytical Studies
- Case-control study
- Cohort study
2. Experimental (Interventional) Studies
- Randomized Controlled Trial (RCT)
- Field trial
- Community trial
3. Quasi-Experimental Studies
- Non-randomized trial
- Before–after study
Among analytical observational studies, the cohort study plays a vital role in establishing temporal relationships between exposure and disease .
Cohort Study
A cohort is defined as a group of individuals who share a common characteristic or experience within a defined time period. Examples include:
- Birth cohort
- Marriage cohort
- Occupational cohort (e.g., radiation workers)
- Vaccine-exposed group
A cohort study is also known as:
- Prospective study
- Longitudinal study
- Incidence study
- Forward-looking study
In this design, individuals are grouped according to exposure status and followed over time to observe disease development .
Indications for Cohort Studies
Cohort studies are particularly useful in the following situations:
- When there is prior evidence of association between exposure and disease from descriptive or case-control studies.
- When exposure is rare but the incidence of disease among exposed is high (e.g., survivors of atomic explosions or industrial disasters).
- When sufficient funding and resources are available.
- When loss to follow-up is expected to be minimal.
Such conditions improve validity and feasibility .
Framework of a Cohort Study
The cohort study follows a cause → effect approach.
Key Features
- Two groups:
- Exposed group (Cohort)
- Non-exposed group (Control)
- Both groups must be free from disease at baseline.
- Diagnostic criteria must be predefined.
- Groups should be comparable in all relevant aspects except exposure.
- Equal diagnostic intensity in both groups.
- Identical follow-up procedures.
Basic 2×2 Framework
| Disease Yes | Disease No | Total | |
|---|---|---|---|
| Exposed | a | b | a+b |
| Non-Exposed | c | d | c+d |
Incidence among exposed = a / (a + b)
Incidence among non-exposed = c / (c + d)
If incidence in exposed is higher than in non-exposed, exposure may be causal .
Types of Cohort Studies
1. Prospective Cohort Study
Participants are identified in the present and followed into the future to observe disease development.
2. Retrospective Cohort Study (Historical Cohort)
Both exposure and outcome have already occurred at the time of study initiation. Records are used to reconstruct exposure and follow-up.
3. Combination Cohort Study
Starts retrospectively and continues prospectively.
Each type has its own logistical and financial considerations .
Elements of a Cohort Study
- Selection of study subjects
- Obtaining data on exposure
- Selection of comparison group
- Follow-up
- Analysis
Selection of Study Subjects
Subjects may be selected from:
- General population (residents of a city or village)
- Special groups (occupational cohorts such as doctors, lawyers)
- Exposed groups (radiation workers, pesticide factory workers)
Selection should ensure representativeness and comparability .
Obtaining Data on Exposure
Exposure data can be collected through:
- Personal interviews
- Review of medical or occupational records
- Environmental measurements
- Medical examination and laboratory tests
Objective measurement reduces misclassification bias .
Selection of Comparison Groups
1. Internal Comparison
Comparison within the same cohort (e.g., light smokers vs heavy smokers).
2. External Comparison
Comparison with a separate non-exposed group.
3. Comparison with General Population
Incidence among exposed compared with population rates.
Proper comparison ensures valid estimation of risk .
Follow-Up
Follow-up methods include:
- Periodic medical examinations
- Review of hospital records
- Death registry surveillance
- Mailed questionnaires
- Telephonic interviews
- Home visits
Loss to follow-up should be minimized and ideally kept below 5%, ensuring at least 95% retention to maintain validity .
Analysis
The primary objective is to compare disease incidence between exposed and non-exposed groups and measure the strength of association.
Key Measures
- Incidence
- Relative Risk (RR)
- Attributable Risk (AR)
- Population Attributable Risk (PAR)
Incidence Calculation
Incidence in Exposed (Ie) = a / (a + b)
Incidence in Non-Exposed (Io) = c / (c + d)
Incidence reflects the probability of developing disease over a specified time .
Relative Risk (Risk Ratio)
RR = Ie / Io
Interpretation
- RR = 1 → No association
- RR > 1 → Exposure increases risk
- RR < 1 → Exposure is protective
Example: If RR = 4, exposed individuals have four times higher risk of disease.
Attributable Risk (Risk Difference)
AR = Ie – Io
This represents the excess risk among exposed due to exposure.
Example: If AR = 0.10, there are 10 additional cases per 100 exposed attributable to exposure.
Population Attributable Risk (PAR)
PAR = It – Io
Where:
It = Incidence in total population
PAR indicates the proportion of disease in the population attributable to exposure.
Case-Control Study vs Cohort Study
| Basis | Case-Control | Cohort |
|---|---|---|
| Direction | Backward | Forward |
| Starting point | Disease | Exposure |
| Grouping | Cases & Controls | Exposed & Non-exposed |
| Incidence | Cannot calculate directly | Can calculate |
| Measure | Odds Ratio | Relative Risk |
| Best for | Rare diseases | Rare exposures |
| Time | Short | Long |
| Cost | Less | More |
| Recall bias | Common | Rare |
| Multiple outcomes | Limited | Possible |
| Temporality | Difficult | Clear |
Cohort studies establish temporality clearly and allow direct calculation of incidence and relative risk .
Summary
- A cohort study follows a group sharing a common exposure over time.
- It uses a cause → effect approach.
- Begins with exposure status and observes disease development.
- Can be prospective, retrospective, or combined.
- Particularly suitable for studying rare exposures.
Key Limitations
- Time-consuming
- Expensive
- Loss to follow-up affects validity
- Not ideal for rare diseases
Measures Used
- Incidence
- Relative Risk (RR)
- Attributable Risk (AR)
- Population Attributable Risk (PAR)
Conclusion
Cohort studies are powerful analytical tools in epidemiology. By following exposed and non-exposed individuals over time, they establish temporal relationships and quantify risk. Despite logistical challenges, they remain essential for studying causal associations, especially when exposure is rare and temporality must be clearly demonstrated .
MCQs
1.
In a 10-year cohort study of 5,000 individuals, 1,000 are exposed to asbestos. Lung cancer develops in 50 exposed and 25 non-exposed individuals. What is the Relative Risk (RR)?
A. 2
B. 3
C. 8
D. 5
Answer: C
Incidence exposed = 50/1000 = 0.05
Incidence non-exposed = 25/4000 = 0.00625
RR = 0.05 / 0.00625 = 8
2.
If RR = 3 and incidence in non-exposed is 5%, what is incidence in exposed?
A. 5%
B. 10%
C. 15%
D. 20%
Answer: C
Ie = RR × Io = 3 × 5% = 15%
3.
A cohort study reports AR = 0.12. Interpretation is:
A. 12% of total population disease due to exposure
B. 12 excess cases per 100 exposed due to exposure
C. Exposure increases risk by 12 times
D. 12% exposed are diseased
Answer: B
4.
Which scenario best establishes temporality?
A. Cross-sectional study
B. Case-control study
C. Cohort study
D. Ecological study
Answer: C
5.
In a historical cohort study, both exposure and outcome have already occurred. The main advantage is:
A. No bias
B. Quick and economical
C. No need for records
D. No confounding
Answer: B
6.
Loss to follow-up of 20% in exposed and 2% in non-exposed group most likely causes:
A. Selection bias
B. Information bias
C. Confounding
D. Recall bias
Answer: A
7.
In a cohort study, incidence among exposed = 30/600 and among non-exposed = 10/400. What is AR?
A. 0.025
B. 0.05
C. 0.075
D. 0.10
Answer: C
Ie = 0.05
Io = 0.025
AR = 0.025
(Correct Answer: B – 0.025; question tests careful calculation.)
8.
Population incidence is 20/1000. Incidence in non-exposed is 10/1000. What is PAR?
A. 5/1000
B. 10/1000
C. 15/1000
D. 30/1000
Answer: B
9.
Which measure cannot be directly calculated from a case-control study but can be calculated in cohort study?
A. Odds ratio
B. Relative risk
C. Exposure frequency
D. Matching ratio
Answer: B
10.
Which is most suitable design to study multiple outcomes of a single rare exposure?
A. Case-control
B. Cohort
C. Cross-sectional
D. Case series
Answer: B
11.
In a prospective cohort study, baseline disease screening is important to:
A. Increase sample size
B. Avoid selection bias
C. Ensure both groups disease-free
D. Reduce confounding
Answer: C
12.
If RR = 1 and AR ≠ 0, which is correct?
A. Calculation error likely
B. Exposure protective
C. No association
D. Strong association
Answer: A
13.
Which is NOT a characteristic of cohort study?
A. Forward direction
B. Measures incidence
C. Best for rare diseases
D. Establishes temporality
Answer: C
14.
In occupational radiation exposure study, workers are compared to general population rates. This comparison is called:
A. Internal
B. External
C. Population
D. Matched
Answer: C
15.
In a cohort study, differential diagnostic intensity between groups leads to:
A. Confounding
B. Observer bias
C. Detection bias
D. Berksonian bias
Answer: C
16.
A combined cohort study begins retrospectively and continues prospectively. Major advantage:
A. Eliminates bias
B. Reduces follow-up
C. Extends observation time
D. Removes confounding
Answer: C
17.
Which is TRUE regarding attributable risk?
A. Indicates strength of association
B. Indicates public health impact among exposed
C. Same as RR
D. Used only in case-control
Answer: B
18.
If exposure prevalence is high and RR is moderate, which measure best reflects public health importance?
A. RR
B. AR
C. PAR
D. Odds ratio
Answer: C
19.
In a cohort of 2000, 500 exposed. Disease develops in 100 exposed and 50 non-exposed. What is RR?
A. 2
B. 4
C. 6
D. 8
Answer: B
Ie = 100/500 = 0.2
Io = 50/1500 = 0.033
RR ≈ 6
20.
Which bias is least likely in cohort study compared to case-control study?
A. Recall bias
B. Loss to follow-up
C. Selection bias
D. Confounding
Answer: A
DoctorDP 