Measures of Disease

Question 1

what is a p value?

Answer 1

1. the probability that difference as large or larger than observed in a sample would arise due to chance alone if there were no real difference (no association)
2. by convention, P of less than or equal to 0.05 indicates significant difference
   -the factor and the outcome are statistically associated

Question 2

what is a confidence interval?

Answer 2

1. confidence interval = estimate +/- margin of error
2. a range of values with associated probability or confidence level
3. quantifies the chance that the interval contains the true population parameter
4. tells the precision of the estimate
5. margin of error dependent upon:
   -data dispersion: standard devioation
   -sample size
   -confidence level: typically 95%, a = 0.05

coleman likes confidence interval much better than p value!!

Question 3

describe relative risk of disease/risk ratio

Answer 3

RR = risk (or incidence) in exposed/risk (or incidence) in non-exposed
-if given a 4x4 chart, (a/b)/(c/d)

RR = 1: no difference in risk
RR >1: exposure increases risk
RR <1: exposure is protective

stated like: red cars had a 2.2x greater risk of getting a speeding ticket than blue cars

Question 4

describe relative odds of disease/odds ratio

Answer 4

OR = odds of dz in exposed/odds of dz in nonexposed
-if given a 4x4 chart:
(axd)/(cxb)
-SEE PPT FOR MATH DIFFERENCE

OR: odds exposed in diseased/odds exposed in non-diseased
-if given a 4x4 chart:
(axd)/(cxb)

OR = 1: no difference in odds
OR >1: exposure increases odds
OR <1: exposure is protective

stated like: red cars had 6.1x greater odds of getting a speeding ticket as compared to blue cars

-often overestimates risk!
-but approximates RR when the outcome is rare in a population

Question 5

describe the attributable proportion or fraction

Answer 5

attributable proportion = (incidence in exposed - incidence in non-exposed) / incidence in exposed

also called the attributable fraction = (RR-1)/RR

tells the proportion of the outcome that can be attributed to exposure

Question 6

what is a biologically relevant association or effect?

Answer 6

a biologically relevant effect is one that is judged to be important or meaningful in the given context

-ex. expert opinion non necessarily statistically significant, but still important

Question 7

is strength of association equivalent to statistical significance?

Answer 7

no, but it gives us some insight into causation and biological relevance