Biostatistics

Question 1

Observing a difference that did NOT exist. if p<0.05, then there is <5% chance the data will show something that is not really there

Answer 1

type 1 error (alpha) - false-positive error

Question 2

Stating there is not an effect or difference when one exists. (1-beta= power)

Answer 2

type 2 error (beta) - false-negative error

Power = prob of rejecting null when it is truly false

Question 3

methods to avoid consequences of unnecessary or excessiev interventions in the health system (eg limit use of abx in kids with viral illness)

Answer 3

quarternary prevention

Question 4

prevent measures so a person doesnt get the disease

Answer 4

primary prevention

Question 5

Reducing negative impact of symptomatic disease via treatment

Answer 5

tertiary prevention

Question 6

Used after a disease has occurred but before the person has symptoms (eg HIV screening in high risk population)

Answer 6

secondary prevention

Question 7

APACHE II scoring system measure what?

Answer 7

Severity of disease and prognosis of patients in the ICU in the first 24 h

Question 8

If 2 studies have the same sample size, how can you determine which has a larger statistical power?

Answer 8

Larger power - larger expected EFFECT size. A larger difference is easier to detect.
Power is the capacity to detect a difference

Question 9

What is the most important feature of screening tests?

Answer 9

High sensitivity - neg test rules out diagnosis (SnNout)

Question 10

Differentiate effect modification from confounding

Answer 10

Effect modification - stratification will result in different measures of association. Effect of main exposure on outcome is modified by presence of another variable
Confounding - stratification will NOT reveal significant difference between strata

Question 11

Test that is reproducible (gives similar results on repeat measurements) is considered what?

Answer 11

Reliable

Maximal when random error is minimal

Question 12

Test’s ability to measure what it is supposed to measure

Answer 12

Validity/Accuracy

Need to compare to gold standard to test this

Question 13

Chi-square tests for what?

Answer 13

assoc bw 2 CATEGORICAL variables

Question 14

two-sampel z and t test are used for what?

Answer 14

Compare 2 group MEANS

Question 15

ANOVA tests what?

Answer 15

Compare the MEANS of 2+ groups

Question 16

Correlation coefficient closer to 1 or -1 indicates what?

Answer 16

0 = no association
1 = positive association
-1 = negative association
closer to |1| means points are less variable along the line of best fit

Question 17

Why do studies with subjective outcomes being evaluated (eg feeling better) raise concern about the validity of the conclusion?

Answer 17

There is variability in outcome assessment

Question 18

Positive predictive value is dependent on prevalence. What happens to PPV if prevalence is high

Answer 18

If prevalence is HIGH, PPV will be high as well

Question 19

Procedure A has a lower rate of wound infection compared with Procedure B (RR of 0.66 with a 95% CI of 0.30-1.45). What can be said about the two procedures.

Answer 19

If a 95% CI for the relative risk includes the null value of 1, then there is insufficient evidence to conclude that the groups are statistically significantly different.