IN.gov - Skip Navigation

Note: This message is displayed if (1) your browser is not standards-compliant or (2) you have you disabled CSS. Read our Policies for more information.

Indiana State Department of Health

Epidemiology Resource Center Home > Surveillance and Investigation > Surveillance and Investigation Division > Newsletters > Indiana Epidemiology Archived Newsletters > Epi_Newsletter_June_2008 E3 Easy Epidemiology for Everyone

Last month, we learned about cohort studies.  Case-control studies are another method epidemiologists use to measure associations between exposure factors and disease/condition. Case-control studies compare two groups, one group with the condition of interest (cases) and a group similar to the first but without the disease or condition (controls). A detailed history of each group is collected to determine what exposures or factors might be associated with the disease being studied.

 

A case-control study is a great method to use when the disease of interest is rare, has a long development period, or if you want to study multiple causes for a single disease. Case-control studies typically are less expensive and faster than cohort studies. Disadvantages of using case-control studies are they are subject to recall bias, since most often the information regarding exposure history is taken after diagnosis; selection of proper controls can be difficult, leading to the possibility of selection bias; and they cannot directly provide absolute risk or rates. It is important to remember that when selecting controls for your study that you pick people as similar to your case population as possible. A good rule to apply is to ask, “If this person had become ill, would he be a case?”  If the answer is yes, then he will make a good control for your study.

 

Case-control studies measure the frequency of exposures associated with the cases, those having the illness, and the controls, those without the illness. From the difference in frequency between the two groups, the Odds Ratio (OR) is calculated. An OR is the ratio of the odds of exposure in diseased subjects to the odds of exposure in the non-diseased. A 2x2 table illustrates the relationship:

Exposure

Disease

Yes

 (cases)

No

(controls)

Yes

A

B

No

C

D

Odds of exposure

A/C

B/D

 

Odds Ratio  = (A/C)/(B/D) or simply AD/BC

 

 

 

 

Example Odds Ratio:

 

 

Ill

Not Ill

Total

Ate Ice Cream

22

16

38

Did Not Eat Ice Cream

7

24

31

Total

30

90

 

 

Odds ratio =     AD    =   22 (24)   =   528   =   4.71

          BC          16 (7)          112

 

The above example shows that those who ate ice cream were 4.71 times more likely to get ill than those people who did not eat ice cream.

 

An OR can sometimes be a good estimate of relative risk but only if the cases and controls you selected are close to the true target population at large. Keep in mind when interpreting an OR, it is only the odds that the exposure is associated with the condition, not the absolute risk associated with the exposure.

 

Reference

1. Friis, Robert. Sellers, Thomas. Epidemiology for Public Health Practice. 3rd Edition. 2004.