RESEARCH AND STATISTICS

Question 1

Study best suited to evaluate the risks and benefits of a new treatment, establishes a direct causal relationship between treatment and outcomes

Answer 1

Randomized control trial

Question 2

Study in which 2 groups are followed prospectively over time to see which exposures/ risks cause disease and to provide information about prognosis

Answer 2

Cohort study (type of observational study)

Question 3

Retrospective study in which people with a disease are compared to those without the disease to evaluate risk factors, useful for rare outcomes

Answer 3

Case-control study (type of observational study)

Question 4

Type of study used to compare a new diagnostic test to the current gold standard diagnostic test in a given population

Answer 4

Cross-sectional study

Question 5

Retrospective statistical analysis of several studies on the same topic

Answer 5

Meta-analysis

Question 6

What is the best way to eliminate confounding variables?

Answer 6

Randomization of study subjects

Question 7

Type of bias that occurs when an association between an exposure and an outcome is distorted by another variable

Answer 7

Confounding bias

Question 8

The percent of people with the disease in a given population being studied

Answer 8

Prevalence

Question 9

The number of new cases of a given disease in a specific period of time

Answer 9

Incidence

Question 10

The probability of correctly identifying those who truly have the disease

Answer 10

Sensitivity = TP/(TP+FN) SnOUT sensitivity → rules out disease when they’re negative

Question 11

The probability of correctly identifying those who truly do not have the disease

Answer 11

Specificity = TN/(TN+FP) SPIN specificity → rules in disease when they’re positive

Question 12

The probability of correctly identifying those who truly have the disease among those whose tests are positive

Answer 12

Positive predictive value = TP/(TP+FP)

Question 13

The probability of correctly identifying those not having a disease among those whose tests are negative

Answer 13

Negative predictive value = TN/(TN+FN)

Question 14

The hypothesis of no difference, i.e., daily exercising does not reduce the risk of heart disease

Answer 14

Null hypothesis

Question 15

The hypothesis of difference, i.e., daily exercising does reduce the risk of heart disease

Answer 15

Alternative hypothesis

Question 16

Rejecting the null hypothesis when in fact it is true—a difference that was seen when one does not exist

Answer 16

Type 1 error

Question 17

Failure to reject the null hypothesis when the null hypothesis is false—a difference that was not seen when one does exist

Answer 17

Type 2 error

Question 18

The probability that a study can detect a treatment effect

Answer 18

Power of a study

Question 19

A range of values with a specified probability that a given parameter falls in that range

Answer 19

Confidence interval

Question 20

What is the interpretation of a confidence interval that includes 0 or 1

Answer 20

There is no statistical significance

Question 21

The total number needed to treat to prevent one bad outcome or adverse events

Answer 21

The number needed to treat NNT = 1/absolute risk reduction

Question 22

The total number of patients receiving intervention for each patient that is harmed

Answer 22

The number needed to harm NNH = 1/attributable risk