Professional Documents
Culture Documents
Observational Studies: Cohort Studies: PBHE 6200: Epidemiology 1 October 29, 2019
Observational Studies: Cohort Studies: PBHE 6200: Epidemiology 1 October 29, 2019
Observational Studies: Cohort Studies: PBHE 6200: Epidemiology 1 October 29, 2019
Common characteristic
Defined Study
Population
Exposed Unexposed
Differences
Why not always use an experimental study?
Practicality of randomization
1. Formulate a hypothesis
Based on exposure
Exposed Unexposed
Defined Study
Population
Exposed Unexposed
• Incidence Proportion
• Incidence Rate
Incidence Proportion
Number of Events
Number at Risk
Incidence Rate
Number of Events
Total Person Time contributed among those at Risk
Consider a cohort study that wants to investigate the association
between HRT (hormone replacement therapy) use and
endometrial (uterine) cancer
Men (?)
Unexposed Incidence
Types of Cohort Studies
1. Prospective (Concurrent)
2. Retrospective (Historical)
Prospective Cohort
Limitations
Retrospective Cohort
• Identify exposed and unexposed individuals from
historical records and follow them to the present to
determine the incidence of disease
example: locate hospital records from maternity wards
from 2006. Identify infants who were breast-fed and
infants who were not. Follow these infants to the
present (i.e. 2 years later) to determine the incidence
of sudden infant death syndrome
Issues with Retrospective Studies
Strengths
Limitations
An epidemiologist is interested in exploring the
association between infant breastfeeding and
sudden infant death syndrome
Prospective
Identify a sample of new mothers in the maternity
ward today. Determine which mothers breastfeed
and which do not. Follow these women for 2 years
to determine the incidence of sudden infant death
syndrome.
Retrospective Study
Exposed Unexposed
Information
When the information gathered in study cohorts
leads to a misclassification of exposure and/or
outcome
Common bias in cohort studies
Selection
‘Healthy Worker Effect’
Attrition
Information
Interviewer bias
Strengths of cohort studies
Cost
Attrition
Duration of Study
Some important cohort studies
Men battling prostate cancer are more than twice as likely to die
from the disease if they are obese, according to a new study that
suggests soaring insulin levels related to excess weight are to blame.
… The study found that obese men with prostate cancer are 2.5
times more likely to die from the disease than patients of a healthy
weight….
Men who were overweight (a BMI of 25 to 29.9) or obese (a BMI of
greater than 30) before diagnosis were significantly more likely to
die from their prostate cancer than men of normal weight, the
researchers found.
Smoking Boosts Aneurysm Risk for Women
http://www.healthzone.ca/health/article/517362