PH Lecture 12 Flashcards
(8 cards)
Methods to increase statistical power
Precise classification of disease and exposure status
Adequate follow up
Steps for judging studies/articles
Association Chance Hypothesis a prior Bias Confounding Causal criteria
Sensitivity analysis (see if follow up is adequate)
If had 100 subjects
84 followed up
4 died
Best case scenario, only 4 died
Worst case scenario, 16+4=20 died
2 broad types of bias
Selection bias - how recruit subjects
Information bias - how collect info from subjects
Healthy worker effect
- what type of bias
- what is it
Selection bias
Workers are more likely to be healthy, so don’t compare working population with general population
Volunteer bias
- what type of bias
- what is it
Selection bias
People who volunteer are diff than general population
Berkson’s bias
- what type of bias
- what is it
Selection bias
Amount of exposure is systematically diff between case population and control population in a case control study
Having both exposure and disease makes a person more likely to be in hospital
3 types of selection bias
Healthy worker effect
Volunteer bias
Berkson’s bias
