EBM

Question 1

Positive Likelihood Ratio

Answer 1

LR+ = true positive / false positive
LR+ = sensitivity / (1-specificity)

Question 2

Positive LR:
approx. increase in post-test probability of disease

Answer 2

+LR of 2 ==> 15% weak evidence to rule in disease
+LR of 5 ==> 30% moderate evidence to rule in disease
+LR of >10 ==> 45% strong evidence to rule in disease

Question 3

Negative Likelihood Ratio

Answer 3

LR- = false negative / true negative
LR- = (1-sensitivity) / specificity

Question 4

Negative LR:
approx. decrease in post-test probability of disease

Answer 4

-LR of 0.5 ==> -15% weak evidence to rule out disease
-LR of 0.2 ==> -30% moderate evidence to rule out disease
-LR of <0.1 ==> -45% strong evidence to rule out disease

Question 5

LR: test or not?

Answer 5

If probability is between test and tx threshold, further testing is required.

If probability is below threshold, no testing is needed (worried about false positives).

If probability is above threshold, treat rather than continue testing (worried about false negatives).

Question 6

Likelihood Ratio interpretations

Answer 6

LR > 1.0:
- the particular test result is more likely to occur in those with disease than those without
- larger +LR = more informative test
- best test to rule in disease = largest positive LR (LR greater than or equal to 10)

LR = 1.0:
- there is no difference in the probability of the particular test result (positive
result for LR+ and negative result for LR−) between those with and without the disease
- aka, the test is useless

LR < 1.0:
- the particular test result is less likely to occur in those with disease than those without
- smaller -LR = more informative test
- best test to rule out disease = lowest negative LR (LR less than or equal to 0.1)

Question 7

number needed to harm (NNH)

Answer 7

number of individuals who need to be exposed to a certain risk factor before one person develops an outcome

NNH = 1/AR
AR = absolute risk

If NNH = 18, that means for every 18 patients there will be 1 who is harmed by the treatment

Question 8

number needed to treat (NNT)

Answer 8

number of individuals that must be treated, in a particular time period, for one person to benefit from treatment

NNT = 1/ARR
ARR = absolute risk reduction

Question 9

Absolute vs. Relative Risk

Answer 9

Absolute risk = actual risk of some event happening given the current exposure

Relative risk = ratio of [the probability of an event occurring in the exposed group] versus [the probability of the event occurring in the non-exposed group]

Question 10

Experimental Event Rate (EER)

Answer 10

Where:
a = number of those exposed with disease
b = number of those exposed without disease

EER = a/(a+b) = the absolute risk of getting the disease in the exposed population

Question 11

Control Event Rate (CER)

Answer 11

Where:
c = number of those NOT exposed with disease
d = number of those NOT exposed without disease

CER= c/(c+d) = the absolute risk of getting the disease in the unexposed population

Question 12

Relative risk (RR)

Answer 12

RR = [a/(a+b)] divided by [c/(c+d)] = EER/CER

Relative risk = the absolute risk of getting the disease in the exposed population over the absolute risk in the unexposed
population

Question 13

Absolute risk reduction (ARR)

Answer 13

ARR = |CER - EER|

ARR:
- is the number of percentage points actual risk goes down if you do something protective
- It is the ABSOLUTE difference in event rates between the experimental and control groups

Question 14

Relative risk reduction (RRR)

Answer 14

RRR = |(CER- EER)/CER|

RRR:
- is the difference in event rates in relative terms expressed in relation to control event rate
- tells you by how much the treatment reduced the risk of bad outcomes relative to the control group
- RRR is more impressive than ARR

Question 15

Confidence interval (CI)

Answer 15

• CI for a proportion that includes the value 1 is NOT statistically significant; for example: 95% (0.82, 1.35)