Public Health Science: Epi and Biostats

Question 1

Cross sectional study

Answer 1

tells you frequency of disease and risk factors, but doesn’t tell you causality, tells you prevalence

Question 2

Case control

Answer 2

looks at two populations and tells you the odds, like people with breast cancer were more likely to have smoked

Question 3

cohort study

Answer 3

can tell you relative risk- who will develop this disease and who did develop this disease? looking at group with risk factor and group without and seeing how many developed it

Question 4

What is the likelihood ratio?

Answer 4

The likelihood that a test result will occur in someone with the disease compared to someone without the disease, LR >10 is good test= sensitivity/1-specificity

Question 5

What is the number needed to treat?

Answer 5

1/absolute risk reduction (absolute risk reduction the incidence in group that didn’t receive tx - the incidence in those that did ie 8% -2% = 0.06 so 1/0.06)

Question 6

What is incidence vs prevalence?

Answer 6

Incidence is the number of new cases in the group at risk and the prevalence is the total number of cases in the population.

Question 7

What are the types of bias?

Answer 7

1. Selection bias (Berkson, nonresponse)
2. Recall bias
3. Measurement bias (Hawthorne effect)
4. Procedure bias
5. Observer-expectancy bias
6. Confounding bias (crossover design or matching studies where you have patients with similar characteristics in the tx and control groups)
7. Lead-time bias
8. Length-time bias (a longer term cancer is more likely caught by a screening test then a rapidly progressing one)

Question 8

What is effect modification?

Answer 8

when the effect of the exposure on the outcome is affected by another variable (ie. age and shoe size and smoking and DVT development in women taking estrogen receptor agonist)

Question 9

What is the formula for FN and TP?

Answer 9

FN= (1-sensitivity) x number of patients with disease
TP= sensitivity x number of patients with disease