CMED 305
Cohort Studies
Nurah Alamro, MD. MPH. DrPH.
Assistant Professor - Community Medicine Unit,
Fa m i l y & C o m m u n i t y M e d i c i n e D e p a r t m e n t
[email protected] Learning Objectives: By end of this session students will
be able to:
1. Describe the types of cohort studies
2. Describe the design of cohort studies
3. Identify steps for conducting cohort studies
4. Identify issues in the design of cohort studies
5. Describe the strengths and weaknesses of cohort studies
{1
Types of cohort studies
}
All Studies
Descriptive (PO) Analytical (PICO)
Generates Hypotheses
Case report Experimental Observational
Tests Hypotheses
Randomized
Case series Clinical Trials Group data Individual
(RCTs)
Cross-sectional Cross-sectional
Ecological study
(survey) (analytical)
Qualitative Cohort
Case-Control
A cohort study is an analytical observational study in
which a group of people with a common characteristic
is followed over time to find how many reach a certain
health outcome of interest (disease, condition, event,
death, or a change in health status or behavior).
• Term "cohort" is defined as a group of people, usually
100 or more in size, who share a common
characteristic or experience within a defined time
period (e.g., age, occupation, exposure to a drug or
vaccine, pregnancy, and insured persons).
• The comparison group may be the general
population from which the cohort is drawn, or it may
be another cohort of persons thought to have had
little or no exposure to the substance in question, but
otherwise similar.
Two types of cohort
studies have been
distinguished on the
basis of the time of
occurrence of disease in
relation to the time at
which the investigation
is initiated and
continued
When to Conduct a Cohort Study
• When there is good evidence of an association between
exposure and disease (If we observe an association
between an exposure and a disease or another
outcome, the question is: Is the association causal?)
• When exposure is rare, but the incidence of disease
high among exposed, e.g. special exposure groups like those
in industries, or exposure to X-rays
• When attrition (loss during follow up) of study population
can be minimized, e.g. follow-up is easy, cohort is stable,
cooperative and easily accessible
• When funds and time are available
{2 }
Design of a Cohort Study
Time
Direction of Inquiry
Disease
Exposed
No Disease
Source People Without
Population the Disease
Disease
Unexposed
No Disease
{3
How to conduct a cohort
study? }
1- Define a source population
2- Select Study Populations (subjects &
controls): two methods: based on exposure
status OR based on factor other than
Steps in exposure e.g. geographic location
conducting a
cohort study 3- Measure the exposure
4- Follow up at intervals to get
accurate outcome data
5- Analyze data
Measuring Exposure Measuring Outcome
• Levels of exposure (e.g. packs of cigarettes • Sources for outcome data: routine
smoked per year) are measured for each surveillance of cancer registry data,
individual at: death certificates, medical records or
1. baseline at the beginning of the study and directly from the participant.
2. assessed at intervals during the period of • Method used to ascertain outcome
follow-up. must be identical for both exposed and
• A particular problem occurring in cohort unexposed groups.
studies is whether individuals in the control
group are truly unexposed. For example, study
participants may start smoking or they may fail
to correctly recall past exposure. Similarly,
those in the exposed group may change their
behaviour in relation to the exposure such as
diet, smoking or alcohol consumption.
• Sources for Exposure data: medical or
employment records, standardized
questionnaires, interviews and by physical
examination.
Analysis in Cohort Studies
The data are analyzed in terms of:
1. Incidence rates of outcome among exposed and non-exposed
2. Estimation of risk:
• Relative Risk (also knows Risk Ratio) (RR)
• Attributable Risk (AR)
Incidence Rates: RR: AR:
Incidence Rate among Incidence rate among exposed Incidence rate among exposed -
exposed= a/a+b Incidence rate among Incidence rate among unexposed
X 100
unexposed
Incidence among exposed
Incidence Rate among
unexposed= c/c+d = a/a+b
“Whatc/c+d
is the ratio of the risk of “How much the disease can be
disease in exposed individuals to prevented if we have an effective
the risk of disease in unexposed measure of eliminating the exposure?”
individuals?”
Vaping and Pulmonary “illness”
Cohort study of vaping and pulmonary
illness followed for 1 year.
Exposure: vaping Outcome: pulmonary
illness Pulmonary No Pulmonary Total
Illness Illness
vaping 42 27,000 27,042
No vaping 7 63,000 63,007
Total 49 90,000 90,049
Pulmonary No Pulmonary Total
Illness Illness
vaping 42 27,000 27,042
No vaping 7 63,000 63,007
Total 49 90,000 90,049
Incidence Rates: RR AR
Incidence Rate among
exposed= = 15 = 93%
1.5/1000/year
What does 15 mean? What does 93% mean?
Incidence Rate among The risk of pulmonary 93% of the morbidity from
unexposed= illness is 15 times higher pulmonary illness among vapers
0.1/1000/year among vapors than non- may be attributable to vaping and
vapers could be prevented by elimination
of vaping
{4 }
Issues in the design of case-
control studies
Loss to Follow Up
• Cohort members may die, migrate, change jobs
or refuse to continue to participate in the study.
• In addition, losses to follow-up may be related
to the exposure, outcome or both.
• For example, individuals who develop the
outcome may be less likely to continue to
participate in the study.
Differential Misclassification of Subjects
• A major source of potential bias in cohort
studies arises from the degree of accuracy
with which subjects have been classified
with respect to their exposure or disease
status.
• Differential misclassification can lead to an
over or underestimate of the effect between
exposure and outcome
Selection Bias
• Selection bias is more common in case-control studies.
• However, it can happen in cohort studies if:
1. The completeness of follow-up is different among exposed and
unexposed.
2. Outcome ascertainment differs between exposed and
unexposed.
Confounding
• Confounding is a distortion (inaccuracy) in the estimated measure of association
that occurs when the primary exposure of interest is mixed up with some other
factor that is associated with the outcome.
• In the figure above, the primary goal is to ascertain the strength of association
between physical inactivity and heart disease. Age is a confounding factor because
it is associated with the exposure (meaning that older people are more likely to be
inactive), and it is also associated with the outcome (because older people are at
{5
Strengths & Weaknesses
}
Strengths Weakness
• Multiple outcomes can be • Costly and time consuming.
measured for any one • Prone to bias due to loss to follow-up.
• Prone to confounding.
exposure. • Participants may move between one
• Can look at multiple outcomes. exposure category.
• Exposure is measured before • Knowledge of exposure status may bias
the onset of disease (in classification of the outcome.
• Being in the study may alter
prospective cohort studies). participant's behavior.
• Good for measuring rare • Poor choice for the study of a rare
exposures. disease (rare outcome).
• Demonstrate causality. • Classification of individuals (exposure
or outcome status) can be affected by
• Can measure incidence. changes in diagnostic procedures.
Thank you!
Office Hours (by appointment via
email):
Mondays & Wednesdays
11 AM – 1 PM
College of Medicine West Building
Level 1 - Office 4011034
[email protected]
References:
• Celentano, David D., and Scd Mhs. Gordis Epidemiology. Elsevier, 2018.
• Hulley, Stephen B., ed. Designing clinical research. Lippincott Williams & Wilkins,
2007.
• Haynes, R. Brian. Clinical epidemiology: how to do clinical practice research.
Lippincott williams & wilkins, 2012.
• Carlson, Melissa DA, and R. Sean Morrison. "Study design, precision, and validity in
observational studies." Journal of palliative medicine 12.1 (2009): 77-82.
• The Centre for Evidence-Based Medicine develops, promotes and disseminates
better evidence for healthcare. Study Design. NA. Accessed September 13, 2019:
https://www.cebm.net/2014/04/study-designs/
• Alexander, Lorraine K., Brettania Lopes, Kristen Ricchetti-Masterson, and Karin B. Yeatts.
"ERIC notebook." 2014. Accessed September 27, 2019:
https://sph.unc.edu/files/2015/07/nciph_ERIC6.pdf
• Coggon, David, David Barker, and Geoffrey Rose. Epidemiology for the Uninitiated. John
Wiley & Sons, 2009.
• Hennekens, Charles H., and J. E. Buring. ”Cohort studies." Epidemiology in medicine (1987).
