Biology 4a assignment

N O N - E X P E R I M E N T A L S T U D Y D E S I G N S P A R T 3 : C A S E - C O N T R O L S T U D Y D E S I G N

6310-WEEK 4

MAIN NON-EXPERIMENTAL STUDY DESIGNS

• Cross-sectional study design • Cohort study design • Case-control study design

CASE-CONTROL STUDIES

CASE-CONTROL RESEARCH DESIGN

• A non-experimental research design involving the comparison of a “case” (person with disease/condition of interest) and a “matched control” (similar person without the condition).

• Retrospective study design: A group of subjects with the outcome (cases) and another without the outcome (controls) are identified. The investigator then works backward to find differences in predictor variables that may be associated with the outcome.

• Advantages: • Inexpensive and efficient for studying rare diseases/conditions

CASE CONTROL STUDY DESIGN

Cases

Disease

Exposed Not Exposed

Controls

No Disease

Exposed Not Exposed

STEPS: CASE CONTROL STUDIES

• Develop a research question • Select a sample from a population of people with the

outcome of interest or disease (cases) • Select a sample from a population at risk without the

outcome of interest or disease (controls) • Measure predictor variables

• Note that the use of two samples in a case-control design is different from a double-cohort design • Double-cohort: two groups chosen based on level of predictor • Case-control: two groups chosen based on presence or

absence of an outcome

STATISTICAL MEASURES IN CASE-CONTROL DESIGNS

• Odds ratio (OR) is a measure of association between an exposure and an outcome. The OR represents the odds that an outcome will occur given a particular exposure, compared to the odds of the outcome occurring in the absence of that exposure.

• OR is used to determine whether an exposure is a risk factor for an outcome, and to compare the magnitude of various risk factors for that outcome. • OR=1 Exposure does not affect odds of outcome • OR>1 Exposure associated with higher odds of outcome • OR<1 Exposure associated with lower odds of outcome

Source: Szumilas M. Explaining Odds Ratios. Journal of the Canadian Academy of Child and Adolescent Psychiatry. 2010;19(3):227-229.

STRENGTHS OF CASE CONTROL STUDIES

• Efficient for rare diseases and those with long latent periods between exposure and disease

• Inexpensive • Small sample size • Ability to examine a large number of predictor

variables • Short duration

WEAKNESSES OF CASE-CONTROL STUDIES

• One outcome (presence or absence of the disease that is the criterion for drawing the two samples)

• No causal relations • Cannot yield estimates for the incidence or

prevalence of a disease • Susceptibility to bias

• Sampling bias • Measurement bias

SOURCES OF BIAS IN CASE-CONTROL STUDIES

• Sampling bias • Cases are sampled from patients in whom the disease has

already been diagnosed and who are available for study. This can lead to a potentially non-representative sample (sample does not include those who are undiagnosed, misdiagnosed, unavailable for study, or dead).

• Strategies for sampling controls to minimize sampling bias • Clinic- or hospital-based controls • Population-based samples of cases and controls • Two or more control groups • Matching

SOURCES OF BIAS IN CASE-CONTROL STUDIES

• Measurement bias • Due to retrospective approach to measuring the predictor

variables or exposure • Nondifferential misclassification of exposure (similar in cases and

controls)  difficulty in finding associations • Differential misclassification of exposure (recall bias: different

recollection of exposure among cases)  unpredictable effects on associations

• Strategies to minimize measurement bias • Use data recorded before the outcome occurred • Use blinding

EXAMPLES CASE-CONTROL STUDIES

1952 CASE-CONTROL STUDY: SMOKING AND LUNG CANCER

• Research question: Do patients with carcinoma of the lung differ from other persons, either in their smoking habits or in some way which might be related to the theory that atmospheric pollution is responsible for the development of the disease?

• Sample: 1,465 lung cancer cases & 1,465 controls with no cancer in 20 hospitals in England. • Cases matched in age, sex, hospital

• Results: There is a relationship between the number of cigarettes smoked daily and developing lung cancer in men.

Source: Doll R, Hill AB. A study of the aetiology of carcinoma of the lung. Br Med J. 1952 ;13;2(4797):1271-86.