Biostats

Question 1

Number Needed to Harm

Answer 1

NNH = 1/AR

AR (Attributable Risk) = risk of an outcome attributable to an exposure
AR = incidence in exposed group - incidence in unexposed group

Question 2

Sensitive or Specific

• Which make good screening tests?
• Which make good confirmatory tests?

Answer 2

High sensitivity = good for screening in a disease with low prevalence

High specificity = good for confirmation after a positive screening test

Question 3

Definition: Negative Predictive Value

Answer 3

NPV = probability of being disease free if test result is negative

• • NPV varies with pretest probability and disease prevalence:
• patient with high probability of having disease will have a low NPV with a negative test
• patient with low probability of having disease will have a high NPV with a negative test

Question 4

Define reliability in terms of a test

Answer 4

Reliability = reproducibility (similar or very close results on repeat measurements)

NOT the same as validity = accuracy

Question 5

Absolute Risk Reduction (ARR)

Answer 5

ARR = event rate in control group - event rate in treatment group

Question 6

Define Latent Period

Answer 6

• The initial steps of pathogenesis and/or exposure to a risk factor sometimes occur years before clinical manifestation of disease are evident
• Exposure to risk modifiers may need to be continuous over a certain period of time before influencing the outcome

Question 7

Specificity (DFN)

Answer 7

The ability of a test to exclude those without the disease

Good for confirmatory tests

Question 8

What does lowering the cut off point of a test accomplish?

Answer 8

Lower cutoff point:

• increased sensitivity
• increased true positives
• increased false positives (more than TP)
• decreased PPV
• decreased FN

Question 9

Sensitivity (DFN)

Answer 9

The ability to correctly identify those with the disease

Good for screening tests

Question 10

Power (Formula + DFN)

Answer 10

Power = 1 - Beta; the probability of seeing a difference when there is one

Beta = Type II error; the probability of concluding there is no difference between two groups when one exists

Question 11

Define Effect Modification

Answer 11

the main exposure has a different impact on the outcome in different circumstances (ie in the presence of a different variable)

UW 1279 [7812851]

Question 12

In positive skew….

Answer 12

There is an outlier making the positive tail long

Mean > Median > Mode