1. NAPLEX preparation: Biostatistics
Ron Lyon, MS, RPH

2. Learning objectives
At the completion of this lecture, you should be correctly answer multiple-choice questions related to:
Descriptive statistics
Describe the difference between relative and absolute risk and how to calculate
Describe number needed to treat (NNT) or harm (NNH) and how to calculate
Understand how to interpret a confidence interval
Distinguish between the major types of cost-____ analyses

3. Mean, median and mode

4. Normal distribution and skew
In a normal distribution, the mean, mode and median are all the same.
Skew

5. Skew and standard deviation

6. Statistics that suggest whether you can trust the results
p value (α [alpha])
Example: Probability that the observed differences between two drugs was actually due to chance
Type 1 error: False positive
First thing you check
Power (1-β [beta])
Do you have enough patients or measurements to be able to distinguish actual results from pure chance?
Example: The study didn’t show a difference between two drugs but would have if it included more patients.
Type II error: False negative

7. No more bacon? Let‘s discuss statistics with clinical relevance

8. Relative difference
The difference could be improvement or increased risk (e.g., relative risk)
“eating 50 grams of processed meat each day -- the equivalent of two slices of ham -- can increase the risk of such cancer by 18%”
Know how to calculate Relative Difference:
([incidence % group A]/[incidence % group B]) – 1
Usually reported as a percentage (e.g., 18%) sometimes as a number (e.g., 0.18)
Related terms: odds ratio, hazard ratio

9. Absolute difference
“eating 50 grams of processed meat each day -- the equivalent of two slices of ham -- can increase the risk of such cancer by 18%”
Baseline incidence: ~1 in 200,000 = = %
With bacon: ~1.18 in 200,000 = = %
Know how to calculate Absolute Difference:
([incidence % group A]-[incidence % group B])
% % = % increase

10. Number needed to (treat/harm)
NNT: number needed to treat
NNH: number needed to harm
Know how to calculate Number needed to ____:
1/absolute risk
1/ % = 1,111,111 people need to start eating 50 gm of processed meat each day in order to find one more case of colon cancer

11. Confidence Intervals Confidence interval of WHO assessment:
(95% CI 1.10 to 1.28)
95% certain that the ‘true’ number is between a 10% and a 28% increase in the incidence of colon cancer
When evaluating a study look to determine if the confidence interval suggests the ‘true’ number could be an increase OR a decrease.
(95% CI to +0.20) = (95% CI -20% to +20%)

12. Cost ______ analysis
Cost effectiveness analysis ($ per successful outcome)
What is the cost to achieve a successful outcome?
Cost benefit analysis ($ only)
Compares the benefits (total costs savings) of two or more different outcomes in terms of monetary units, including time value of money (net present value [NPV])
Cost utility analysis
Adds a qualitative measure, usually QALY or DALY, to cost effectiveness analysis
Cost minimization analysis
If both treatments are equal, which one is less expensive

13. Cost effectiveness analysis
What if we want to know how much it costs to achieve an outcome?
Incremental cost to prevent a death
ICER: incremental cost effectiveness ratio
PCSK9-inhibitor example
Death
Absolute risk reduction: 0.22%
Number needed to treat (NNT) to save one life: 455
Calculation: 1/0.22%
Incremental cost to save one life? (ICER)
455 (NNT) x $14,144 (incremental medication cost) =
$6.4 million
What primary endpoint should be our focus?
Secondary endpoints?
How does this compare to current standard of care?
NNT to achieve primary endpoint
What’s the incremental cost of therapy per patient to achieve that endpoint?
NNT x therapy cost comparison
How does that cost/benefit (e.g., saved life, QALY) compare to other unrelated treatments? Pick medical treatment that is broadly accepted as cost beneficial.
High risk patient statin example: The pooled estimates show values of $21571 per life year saved for a 10 year coronary heart disease risk of 20% and $16862 per life year saved for 10 year risk of 30%.
Did you report absolute or relative risk reduction? Answer should almost always be absolute risk.
If price was the entire plan budget would you pay it?
Average life insurance policy is:

14. Cost-effectiveness analysis
How much does it cost to achieve a defined outcome, such as quitting smoking or achieve normal body weight.
Which one do you choose?
Program Cost
# of enrollees
Success rate
# of successes
Cost per success (effectiveness)
Smoking cessation
$50,000
100
10% 10
$5,000
Weight loss
$100,000
1,000 2% 20

15. Cost-benefit analysis
Logically, if savings exceed the costs then you should do both programs.
But you have limited start-up costs. Which one do you choose?
Total Savings
Program Cost
# of enrollees
Net savings per enrollee
Net program savings
Smoking cessation
$150,000
$50,000
100
$1,000
$100,000
Weight loss
$250,000
1,000
$150

16. Cost-utility analysis
Adjusts cost-effectiveness to account for patient’s perceived value of their quality of life.
QALY: quality-adjusted life year
DALY: disability-adjusted life year
Cost per success (effectiveness)
QALY
Cost per adjusted effectiveness
Cost-utility
Smoking cessation
$5,000
0.5
$10,000
Weight loss
1.0

17. Cost ______ analysis
Cost effectiveness analysis ($ per successful outcome)
Effectiveness success
Cost benefit analysis ($ only)
Benefit overall savings
Cost utility analysis
Utility quality
Cost minimization analysis
Minimization lowest program cost