Key areas
• Analytical case control studies • Population base of case control studies • Case and control recruitment • Measurement of exposure
Case control study
• Recruitment of: Case-patients affected with a disease Unaffected control-subjects
• Comparison of exposure status• Observation of the in past presence of
one or more potential risk factors
Analytical case control studies
Objective of a case control study
• Case control studies are analytical in nature
• Case control studies compare in terms of exposure status : Case-patients affected by a disease Unaffected control-subjects
Analytical case control studies
Exploratory case control studies
• Risk factors for a disease are unknown • The case control study explores
potential risk factors for further investigation
• Risk of multiple comparisons! If the p value is set at 0.05, every 20 tables
will generate a “significant” association
Analytical case control studies
Case control studies designed to test a limited number of hypotheses
• Risk factors for a disease are better framed Established risk factors Unclear risk factors Possible risk factors
• A limited number of hypotheses are examined Unclear risk factors Possible risk factors
• Ideal situation from a methodological point of view
Analytical case control studies
Elements defining a case control study
• Study participants Selected on the basis of their disease status
• Logic Retrospective examination of potential
exposures
• Logistic Prospective Retrospective
Analytical case control studies
Cases Controls Total
Exposed a b -
Non-exposed c d -
Total a+c b+d -
Presentation of the data of a case control study in a 2 x 2 table
Analytical case control studies
All case control studies come from a theoretical cohort
• Cohorts follow exposed and unexposed subjects for the development of illness
• Cases and controls can be thought of as extracted from a theoretical cohort in which they are “nested”
• Thinking of case control studies as nested in a theoretical cohort help investigators in designing them appropriately
Population base
Representation of a cohort study
One year
Study subject developing the illness
Study subjectcensored
Study subject followed up
Observation
Legend
Nesting a case control study in a cohort for
a disease with a one year referent exposure period
Case
Control
Legend
Time comparability of cases and controls
• Cases have an onset date• Controls have no onset dates
The need to identify a referent exposure period is less obvious
Controls must nevertheless be observed over a referent exposure period that needs to be clarified
Population base
Characteristics of the two cases and three controls
• Come from the same population
• Can be exposed the same way
• Can develop the disease the same way
• Could have been identified as cases the same way
• Have identical exposure windows
Population base
Case Control Total
Exposed a.f1 b.f2 N/A
Non exposed c.f1 d.f2 N/A
Total C1.f1 C0.f2 N/A
Impossibility to calculate a relative risk in a case control study
Cases are sampled from all cases (sampling fraction: f1)
Controls are sampled from all controls (sampling fraction: f2)
f1 and f2 are unknown, risks cannot be calculated Population base
Unexposed and exposed study subjects in the cohort Legend
Subject exposed
Subject unexposed
a= ?
b= ?
c= ?
d= ?
ill Non-ill Total
Exposed 1 2 3
Non-exposed 1 4 5
Total 2 6 8
Presentation of the data of the analytical cohort study in a 2 x 2
table
Relative risk = (1/3) / (1/5) = 33% / 20% = 1.7
Population base
Sampling fraction for cases and controls in the example Legend
Subject exposed
Subject unexposed
f1 =
f0 =
Population base
Case Control Total
Exposed 1x1=1 2x0.5=1 N/A
Non exposed 1x1=1 4x0.5=2 N/A
Total 2x1=2 6x0.5=3 N/A
Calculation of the odds ratio in a case control study
Odds ratio = (1x2) / (1x1) = 50 / 100 = 2Higher than the relative risk? Why?
The disease is not rare! Attack rate: 25%Population base
Prospective case control studies
• Identification of cases prospectively Surveillance Recruitment in a health care facility Recruitment with health care providers
• Recruitment of controls prospectively • Investigators look at exposure
retrospectively
Population base
Retrospective case control studies
• Identification of cases retrospectively Surveillance data Health care facility registers Case records
• Recruitment of controls retrospectively • Investigators look at exposure
retrospectively
Population base
Specific case control study designs
• Truly “nested” case control studies Set of cases and controls identified
prospectively from a cohort to obtain intermediate results
• Case-cohort studies Cases obtained from surveillance data Controls selected from a cohort study
Population base
Case definitions in case control studies
• Person Signs and symptoms Biological criteria Demographic characteristics
• Time Time of onset
• Place Place of residence
Cases and controls
Control recruitment strategies in case control studies
• Population based Sampling of the general population Random digit sampling
• Health care facility based Hospital based Patients with other diseases
• Case-based (Beware of matching) Friends Neighbourhood
Cases and controls
Defining the recruitment strategy in case control studies
• Person Signs and symptoms (or absence of…) Biological criteria
(e.g., susceptibility) Demographic characteristics
• Time Referent exposure period
• Place Place of residence
Cases and controls
Collecting good data on exposure
• Objectively Reproducibility of exposure measurement
• Accurately Information reflecting as closely as possible
the effect of exposure
• Precisely Quality management in exposure
measurement
Exposure
Measuring the dose of exposure
• Dichotomous exposure measurement Exposed / unexposed
• Measurement of the dose of exposure Accurate measurement of the dose of
exposure(e.g., Cumulated number of cigarettes smoked)
Exposure categories Dose / response effect
Exposure
Understanding the basic relation between exposure, time and
outcome when designing a case control study
Exposure
Outcomes(e.g., Disease)
Time
Referent exposure
period(Time during which
exposure occurs)
Time at risk for exposure effects
Exposure