Exam Questions

Question 1

How do you calculate an odds ratio?

Answer 1

A/C / B/D or AD/BC

Question 2

Calculate the odds ratio, and interpret what it means

Answer 2

Question 3

OR = 1

Answer 3

exposure does not affect the odds of outcome

Question 4

OR > 1

Answer 4

exposure associated with higher odds of outcome

Question 5

OR < 1

Answer 5

exposure associated with lower odds of outcome

Question 6

How do you write an odds ratio statement?

Answer 6

The odds of (outcome) are ____ (higher/lower) given the exposed group vs. the non-exposed group.

Question 7

probability that a person with a disease will have a positive test for the disease

Answer 7

sensitivity

Question 8

high sensitivity means ___ false negatives

Answer 8

few

Question 9

probability that a person without the disease will have a negative test for the disease

Answer 9

specificity

Question 10

high specificity means ___ false positives

Answer 10

few

Question 11

probability that a person who tests positive for disease actually has the disease

Answer 11

positive predictive value

Question 12

probability that a person who tests negative for disease actually is without the disease

Answer 12

negative predictive value

Question 13

proportion of individuals in the population with the disease

Answer 13

prevalence

Question 14

How do you calculate sensitivity?

Answer 14

A/(A+C)

or TP/ (TP+FN)

Question 15

How do you calculate specificity?

Answer 15

D/(D+B)

or TN / (TN + FP)

Question 16

How do you calculate PPV?

Answer 16

A/(A+B)

or TP / (TP + FP)

Question 17

How do you calculate NPV?

Answer 17

D/(C+D)

or TN / (TN + FN)

Question 18

How do you calculate prevalence?

Answer 18

(A+C)/(A+B+C+D)

Question 19

Make a 2x2 table labelling true positive, false negative, false positive, and true negative

Answer 19

Question 20

How do you interpret a confidence interval?

Answer 20

“We are 95% sure that the average days to recovery for medicine A is 2-4 days. We are 95% sure that the average days to recovery for medicine B is 5-8 days”

Question 21

Confidence interval overlaps

Answer 21

not significant

*population means not different

Question 22

What 3 things impact a confidence interval?

Answer 22

1. variation in the population (standard deviation)
   - high variation (SD) = wider
2. size of the sample
   - smaller sample = wider
3. width of confidence interval
   - higher confidence = wider

Question 23

Confidence interval does not overlap

Answer 23

significant

*population means different

Question 24

How do you calculate ESS?

Answer 24

Question 25

Calculate the ESS

Answer 25

As rho increases, the sample size required to detect a significant difference for the variable increases

Question 26

What is clustering?

Answer 26

Clustering is used when its not possible to randomize to the measured unit. For example, if you were conducting a study using a dietary intervention, it would be impractical to expect all family members to have 4 different randomly selected diets. In clustering, the intervention is applied at the level of the cluster and the outcomes are measured at the level of the patient. However since homogeneity between subjects in clusters reduces variability compared to a random sample, an increase in sample size is required to maintain statistical power. This is determined using the estimated sample size calculation.

Question 27

What is the difference between prevalence and incidence?

Answer 27

Prevalence = proportion of a particular population found to be affected by a medical condition or outcome

prevalence = # of people in sample with X / total # of people in population

Incidence = probability of an outcome in a population during a specified period of time; only considers number of NEW cases in the ELIGIBLE population

Question 28

What is the difference between 1-tailed and 2-tailed tests?

Answer 28

A two tailed test the result could be higher or lower (Ex. mouthwash A will result in a different number of caries than mouthwash B)
alpha is 0.05 so there is 2.5 % chance of error claiming it is higher and a 2.5% chance of error claiming it is lower.

A one-tailed test decides that the effect will either increase or decrease the outcome. (Ex. Dat bootcamp will increase DAT scores). Alpha is 0.05 so there is 5% chance of claiming error is higher OR claiming error is lower.

A one-tailed test should only be chosen if the consequences of missing an effect in the untested direction is negligible and in no way irresponsible or unethical.

Question 29

What are the 5 uses for diagnostic tests?

Answer 29

Question 30

When would you want to use a high sensitivity vs. specificity?

Answer 30

few false negatives –> MEANING: “If someone has the disease, this test is very likely to detect if the person has the disease”

Highly contagious diseases (AIDS, covid, EBOLA). You want to make sure that everyone who has the disease you will catch it.

Question 31

When would you want to use a high specificity vs. sensitivity?

Answer 31

few false positives –> MEANING: “If someone doesn’t have the disease, this test is very likely to indicate they do not have the disease”

Giving a medicine or treatments. If the treatment is really extreme (i.e., cut off your left leg) you want to be really sure that if you don’t have it you’re not going to chop off your leg.

Question 32

If you double the prevalence, what happens to sensitivity?

Answer 32

Nothing, because sensitivity does not depend on prevalence.

Question 33

What does the green number mean?

Answer 33

The chance of the crown failing in year 3 if it has survived through year 2 is 0.111.

Question 34

What does the red number mean?

Answer 34

A crown that has survived 3 years has a 0.667 chance that it will survive to the end of year 4.

Question 35

What does the blue number mean?

Answer 35

The chance of the crown surviving to year 5 from the start is 0.328.

Question 36

What does the pink number mean?

Answer 36

The chance of a crown failing in year 2 from the start is 0.08.