Answer Happy

Case (MI+, D+) Control (MI-, D-) Diabetes Yes 122 31 Diabetes No 1490 1581 In this case-control study we want to see if diabetes is a risk factor (exposure) for myocardial infarction (outcome) 1. Calculate the odds ratio. 2. To see if smoking is a confounder, we can check the three properties of a confounder in our data. PROPERTY 1: Is the prevalence of smoking different between our exposed and unexposed groups? We must look at this association in the ABSENCE of disease (i.e. in our control group) because we know that smoking may be a cause of MI (the outcome). Hence the question we ask is: Is smoking associated with diabetes in the controls? Here are the data for controls only in a 2x2 table with smoking (counfounder) CONTROLS ONLY (D-Smokers (C+) Non-smokers (C) Total Diabetes + (E+) 15 16 31 Diabetes - (E- 155 1426 1581 a) What is the prevalence of smoking in controls with diabetes? b) What is the prevalence of smoking in controls without diabetes?