Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Case-Control Studies

No description

on 1 December 2014

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Case-Control Studies

Selecting Cases
Establish a definition for "case"

Sources used to identify cases efficiently and accurately

Incident cases over prevalent cases for causal research
Selecting Controls
Control and cases should be from same source population

Sources used to identify controls

Ratio of controls to cases

Methods for sampling controls

Sources of exposure information
Case control Studies
What is a Case-control Study?
Case-control studies, like other epidemiological studies, are meant to measure the relationship between an exposure and a disease.

In case-control studies, subjects are selected on the basis of whether or not they have the disease or condition of interest.

Those who have the disease/condition are CASES

Those who do not have the disease/condition are CONTROLS

Once cases and controls have been identified and enrolled, investigators obtain exposure information from both cases and controls, and then that information is compared.

Case-Control Study
2x2 Table
Traditional vs. Modern view
Are case-control studies "backward"?
The traditional view is that case-control studies are "backwards" because they move from effect to cause (subjects are initially divided based on whether or not they have the disease, after which exposure histories are obtained and compared between groups) and cohort/experimental studies are preferable because they move from cause to effect (subjects are initially divided based on whether or not they are exposed, and then followed prospectively to identify occurrence of disease). TROHOC
When is a case-control study appropriate?
There are six types of situations where case-control studies are most suitable.
OR = 25*225
= 0.6
OR = (a/b)/(c/d) = ad/bc
When are they most efficient? Rare Diseases. Why?

Less Time and Money

Ability to analyze multiple exposures related to a disease

Good for diseases with long induction and latent periods. Why?

Another example:
Clear and precise criteria
Leads to accurate classification of diseased and non-diseased

Criteria based on:
Signs and symptoms
Physical and pathological exams
Results of diagnostic tests

Reducing classification errors
Subdivisions of severity
Likelihood of diagnosis accuracy
NOTE: exercise caution in comparing studies with different definitions
Establishing a case definition
Incident vs. prevalent cases
Typical sources for identifying:
Special surveys
Special reporting systems
Death Certificates

Identify as many true cases as
quickly and cheaply as possible
Incident cases are preferred over prevalent

Interested in facts that lead to developing the disease (incidence)
NOT in factors that affect the duration of the disease (prevalence)
When exposure data are difficult or expensive to obtain

When the disease is rare, frequency of 20% or less
When the disease has a long induction and latent period
When not much is know about the disease (ie: etiology)
When the population of interest is dynamic
When investigators need to perform a nested case-control study within a cohort study

Prone To Bias. Which Biases?

Inefficient for rare exposures

Difficult to infer a temporal relationship between exposure and outcome. However, you can say there is a meaningful association

Poor information on exposures. Why is this?
What are some factors that make it more favorable to use a case-control study design?
What are some factors that make it less favorable to use a case-control study?
However, the modern view - based on work by Miettinen in the 1980s - asserts that case-control studies are an effective and efficient means of learning about the relationship between exposures and diseases. Specifically, because the control group is meant to reflect the exposure distribution in the population that produced the cases, the key comparison - a comparison between exposed and unexposed groups - is the same. TROHOC fallacy.
Exposure data is collected on a select few instead of entire cohort
Low incidence means "low yield"
in a cohort study. This lowers power.
A cohort study of diseases with
long latency periods would require
very lengthy follow-up.
Multiple exposures (ie: more than one hypothesis) can be considered.
Because lengthy follow-up can be avoided in case-control studies, it is an ideal design for changing populations.
Special Case: Nested Case Control
-Use the participants of a cohort study to run a separate case-control study
-Efficient and cost-effective
-Comparable relative risk for entire cohort if appropriate controls selected

Population control

Hospital/clinic control

Dead control

Friends/Relatives Control

REMEMBER: The goal is to draw from same source population from which the cases were drawn
Sources used to identify controls
Can obtain information easily
Possible to mask interviewers to outcome status
Must be carefully designed to avoid misclassification

Pre-existing records
Avoids bias
May lack details
Low cost

Can be expensive
If disease already occurred, accurary effected
Sources of Exposure Information
Survivor sampling
Controls sampled from survivors

Base sampling
Controls sampled from population at risk when follow up begins

Risk-set sampling
Controls sampled from population at risk as cases are diagnosed
Methods for sampling controls
(Nested case-control study)
Control to case ratios of up to 4:1 help increase power
Can have a higher ratio if cost of additional controls is small
Ratio of controls to cases
Control group can also be called the
referent group

to exposure distribution in the source population

Use the "would criterion"
a control got the disease, they
end up as a case
Implies same criterion except for having the disease

Be sure to sample controls independent of exposure status
Same probability of selecting exposed versus unexposed
Basics of Selecting Controls
"In the nested case-control study, cases of a disease that occur in a defined cohort are identified and, for each, a specified number of matched controls is selected from among those in the cohort who have not developed the disease by the time of disease occurrence in the case. For many research questions, the nested case-control design potentially offers impressive reductions in costs and efforts of data collection and analysis compared with the full cohort approach, with relatively minor loss in statistical efficiency. The nested case-control design is particularly advantageous for studies of biologic precursors of disease. (Ernster, 1994, p.587)"

Ernster, V.L. (1994). Nested case-control studies. Preventive Medicine, 23(5), 587-590.

Selecting cases
& controls

Real world examples
Strengths &

The Basics
Special Case: Case-Crossover Study
-Studies the effects of brief exposures on the risk of acute events
-Hazard Period
-Case serves as controls-> Advantages to this?
-exposure frequency compared between control period and hazard period

Some more about Risk Set sampling
Nested case-control study

Real life example of a case control study
Indoor Tanning and Risk of Melanoma

Questions showing Key concepts/Strengths
1. Was the study design appropriate?
2. How were the cases identified; incident or prevalent?
3. What type of control was used? Why?
4. How was selection bias minimized?
5. What were the sources of exposure information?
6. How was recall bias minimized?
7. What was done to ensure the validity of results?
8. What was the measure of effect estimate used?

1. Yes, because melanoma is a rare disease
2. Eligibility criteria; incident cases
3. Population control; most representative
4. Same eligibility criteria (except outcome), frequency matching, recontact

5. Self-administered questionnaires, Computer assisted telephone interview
6. Pre-tested and simplified questionnaires with appropriate response options
7. Adjustment of confounders
8. Odds ratio

Any identified or potential limitations?

Unequal participation rates
Time consuming and expensive to identify controls
Opportunity of exposure in control and cases may not be the same
Cannot assess effect of matched variables (age and gender)
Full transcript