Frequency and Association

Question 1

Attributable risk

Answer 1

= Incidence in exposed - Incidence in unexposed

Question 2

Percentage of risk attributed to x

Answer 2

= ( AR / Incidence in exposed ) x 100%

Question 3

Population attributable risk

Answer 3

What percentage of the total risk in a population is due to x?

PAR = Incidence in population - Incidence in unexposed

Question 4

Condyloma acuminata

Answer 4

Fancy name for anogenital warts. Caused by HPV.

Small, skin-colored or pink growths on the labia, at the opening of the vagina, or around or inside the anus

Note that strains 6 and 11, the wart-producing strains of HPV, do not typically cause cancer. However, patients may be coinfected.

Question 5

Cumulative incidence

Answer 5

of new events / total # persons at risk over a specified time period. (It is a unitless fraction, but the observation period is described (e.g. “the annual incidence of pancreatic cancer among men 40-50 years old”).

Question 6

Incidence rate

Answer 6

aka Incidence density

of new events/ sum of total person-time at risk (e.g. 20 cases per 1000 person-years)

Question 7

person-years

Answer 7

A person-year is one person followed for one year, or one person followed for 3 months together with another followed for 9 months, or 52 people each followed for a week, etc…

Question 8

Risk ratio (RR)

Answer 8

Risk of disease in exposed group

______________________________________

Risk of disease in unexposed group

Question 9

Odds ratio (OR)

Answer 9

Odds of disease in the exposed group

_______________________________________

Odds of disease in the unexposed group

Question 10

Regression discontinuity design

Answer 10

Utilizes arbitrary administrative cutoff dates to act as faux-randomization. Focuses on the population just proximal to the cutoff date, assuming that as you go farther and farther away the confounds increase. Birth date used to determine school year is a common “forcing variable” in this case, however physical differences between those born at the very end of August and at the very beginning of September can be assumed to be minimal just proximal to the September 1st cutoff.

Question 11

number needed to treat

Answer 11

The number of additional patients you must treat in order to prevent one additional (operationally defined) bad outcome

Question 12

OR vs the RR

Answer 12

The OR is always further from 1 compared with the RR, but when the outcome is rare, the RR and the OR are similar.

Question 13

Which measure of risk would you use when discussing with your patient whether or not she should take NSAIDS?

Answer 13

Attributable risk/Risk difference