Selection of cases: Issues
-Incident cases: Disadvantage: Hard to study rare diseases.
Advantage: Can establish before and after relationship between exposure and disease.
-Prevalent Cases: Advantage: Rare diseases are easier to study.
-Disadvantage: May uncover association between living with a disease and exposure, and not between a disease itself and exposure.
Selection of Controls:
-Healthy controls, neighborhood, "best friend" (relatives help control genetic factors that may contribute to the disease)
-easy to get, don't have a particular disease
-problem: Hospital controls are not necessarily representative of the underlying population.
Control Matching to Cases:
-Group Matching: have cases, figure out characteristics you think are important to match on, calculate % cases with characteristics, pick controls so the % of people with different characteristics matches the cases.
-Enroll control matched fro significant characteristics.
(ex: case 1: 45yr. old divorced white female with the disease. Control 1: 45yr. old divorced white female without the disease. Problems: Traits that you match case and controls on cannot be studied.)
Problems with determining exposure with interview/survey:
-People do not accurately know information.
-Recall bias: People who have bad outcome are more likely to remember prior events than people without a bad outcome.
-Greater than of equal to 1 control per case, lots of subjects increases statistical power.
-Cannot increase the number of cases much, but can increase the number of controls to increase the subject number.
-look at overall population
Population-Surveys/data collection-People with the disease and exposure, Disease and without exposure, Without disease and with exposure, without disease and exposure.