Sample Size Calculation for

Clinical Studies
Rizaldy Pinzon, MD, PhD
Email: drpinzon17@gmail.com
Objectives of this session:
Students able to

(1) Know the importance of sample size in a clinical studies.

(2) Understand the simple mathematics & assumptions involved in the
sample size calculations.
(3) Apply sample size methods appropriately in their research projects.
Why to calculate sample size?
• To show that under certain conditions, the hypothesis test has a good
chance of showing a desired difference (if it exists)
• To show to the IRB committee and funding agency that the study has
a reasonable chance to obtain a conclusive result
• To show that the necessary resources (human, monetary, time) will
be minimized and well utilized
What do I need to know to calculate sample
size?
• Most Important: sample size calculation is an educated guess
• It is more appropriate for studies involving hypothesis testing
• There is no magic involved; only statistical and mathematical logic
and some algebra
• Researchers need to know something about what they are
measuring and how it varies in the population of interest
Before We Can Determine Sample Size
We Need To Answer The Following:

1. What is the primary objective of the study?

2. What is the main outcome measure?

Is it a continuous or dichotomous outcome?

3. How will the data be analyzed to detect a group

difference?

4. How small a difference is clinically important to detect?

Before We Can Determine Sample Size
We Need To Answer The Following:

5. How much variability is in our target population?

6. What is the desired a and b?

7. What is the anticipated drop out and non-response % ?

Where do we get this knowledge?
• Previous published studies

• Pilot studies

• If information is lacking, there is no good way to calculate the sample

size
 Estimation of Sample Size by
Three ways:
By using
(1) Formulae (manual calculations)
(2) Sample size tables or Nomogram
(3) Software
Scenario 1 All studies Scenario 2
Precision Power

Descriptive Hypothesis testing

Sample Simple - 2 groups

surveys
More than 2
groups &
Complex studies
SAMPLE SIZE FOR ADEQUATE PRECISION

• In a descriptive study,
• Summary statistics (mean, proportion)
• Reliability (or) precision
• By giving “confidence interval”
• Wider the C.I – sample statistic is not reliable and it may not give an
accurate estimate of the true value of the population parameter
 Sample size formulae for reporting
precision
For single mean : n = Z2α S2 /d2
where S=sd (s )

For a single proportion : n = Z2αP(1-P)/d2

Where , Zα =1.96 for 95% confidence level
Zα = 2.58 for 99% confidence level
 Problem 1 (Single mean)
A study is to be performed to determine a certain
parameter(BMI) in a community. From a previous
study a sd of 46 was obtained.
If a sample error of up to 4 is to be accepted. How
many subjects should be included in this study at 99%
level of confidence?
Answer
n = (Za/2)2 s2 / d2
s: standard deviation = 46

d: the accuracy of estimate (how close to the

true mean)= given sample error =4

Za/2: A Normal deviate reflects the type I error.

For 99% the critical value =2.58
Problem 2 (Single proportion)

It was desired to estimate proportion of anemic

children in a certain preparatory school. In a similar
study at another school a proportion of 30 % was
detected.
Compute the minimal sample size required at a
confidence limit of 95% and accepting a difference
of up to 4% of the true population.
Answer
n = (Za/2)2 p(1-p) / d2
p: proportion to be estimated = 30% (0.30)
d: the accuracy of estimate (how close to the true
proportion) = 4% (0.04)
Za/2: A Normal deviate reflects the type I error
For 95% the critical value =1.96
 Scenario 2
Three bits of information required to
determine the sample size

Type I & II Variation

errors Clinical
effect
Quantities related to the research question (defined by
the researcher)

v a = Probability of rejecting H0 when H0 is true

v a is called significance level of the test

v b = Probability of not rejecting H0 when H0 is

false

v 1-b is called statistical power of the test

• Researcher fixes probabilities of type I and II errors
• Prob (type I error) = Prob (reject H0 when H0 is true) = a
• Smaller error Þ greater precision Þ need more information
Þ need larger sample size

• Prob (type II error) = Prob (don’t reject H0 when H0 is false) = b

• Power =1- b
• More power Þ smaller error Þ need larger sample size
Sample size and Hypothesis testing

• Types of Errors in Hypothesis Testing

– Ho: null hypothesis: no difference between groups

– Ha: alternative hypothesis: difference between groups

Decision In the Population

H0 true Ha true

Do not reject H0 Correct decision Type II error

(Probability 1 – α) (Probability β)

Reject H0 Type I error Correct decision

(Probability α) (Probability 1 – β)
Sample size and Hypothesis testing

• Type I Error: decided to reject the null hypothesis Ho when the null

hypothesis is true

• Probability of making a Type I error is denoted by α (alpha)

• Typically set at 0.05 (95% confidence/probability of correct decision)

Decision In the Population

H0 true Ha true

Do not reject H0 Correct decision Type II error

(Probability 1 – α) (Probability β)

Reject H0 Type I error Correct decision

(Probability α) (Probability 1 – β)

alpha = 0.05
Sample size and Hypothesis testing

• Type II Error: we not reject the null hypothesis Ho when the null

hypothesis is false (Ha: alternative hypothesis is true)

• Probability of making a Type II error is denoted by β (beta)

• Typically set at 0.1 or 0.2 (80% or 90% power)

Decision In the Population

H0 true Ha true
Beta = 0.2
Do not reject H0 Correct decision Type II error
(Probability 1 – α) (Probability β)

Reject H0 Type I error Correct decision

(Probability α) (Probability 1 – β)
Sample size and Hypothesis testing

• Decision to reject or accept null hypothesis measured by a p-value

(probability that the difference/change is a result of chance alone)

• P-value of <0.05 typically used to reject null hypothesis (α=0.05), or

less than 1 in 20 chance that difference due to chance alone

Decision In the Population

H0 true Ha true

Do not reject H0 Correct decision Type II error

(Probability 1 – α) (Probability β)

Reject H0 Type I error Correct decision

(Probability α) (Probability 1 – β)
Sample size and Hypothesis testing – p-values

Source: Sterne J,

et al. BMJ, Jul

2001; 323:

101 - 105 ;
Statistical versus Clinical Significance

• If sample size very large:

– n=1000 in each group

– difference of 5% in each group (i.e. 45% vs 50%)

– p-value = 0.025 (i.e. < 0.05, thus reject H0)

• If sample size quite small:

– n=50 in each group

– difference of 5% in each group (i.e. 45% vs 50%)

– p-value = 0.617 (i.e. > 0.05, thus do not reject H0)

Sample size (proportions) - estimated margin of error

• for question with expected/estimated proportion of 50% and with

95% confidence interval and unknown/large population

1.0 Sample Margin of

0.9
Size Error
0.8
10 0.30
Sample Proportion

0.7
0.6 20 0.22
0.5
30 0.18
0.4
0.3 50 0.14
0.2 75 0.115
0.1
100 0.10
0.0
0 50 100 150 200 250 300 350 400 200 0.07
Sample Size 300 0.06
400 0.05
Sample size (proportions) - effect of proportion

• assuming unknown/large population, margin of error of ±5%

Sample Size
and 95% confidence interval Proportion Required
0 0
0.05 73
450
0.1 139
0.15 196
400
0.2 246
350
Sample size required

0.25 289
300 0.3 323
250 0.35 350
0.4 369
200
0.45 381
150
0.5 385
100 0.55 381
50 0.6 369
0 0.65 350
0.7 323
0.1
0

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1
0.75 289
Expected proportion meeting criteria for question 0.8 246
0.85 196
0.9 139
0.95 73
1 0
Sample size (proportions) - effect of population

• assuming proportion of 0.50, margin of error of ±5% and

95% confidence interval

450 Sample Size

400
Population Required
350
Unknown (>100,000) 385
Sample size required

10,000 370
300
1,000 278
250
500 218
200
200 132
150
100 80
100
50 45
50
30 28
0
25 24

30
50

25

20
(>100,000)

10,000

1,000

500

200

100
Unknown

20 20
Estimated population size
Sample size (means) - estimated precision in mean

• for variable with mean of 100, sample standard deviation of 50

and with 95% confidence interval and unknown/large population

140 Sample
130 Size Precision
120 10 31
110 20 22
Sample Mean

100 30 18
90 50 14
80 75 11.5
70 100 10
60
200 7
0 50 100 150 200 250 300 350 400 450 300 5.75
Sample Size
385 5
Sample size (means) – effect of sample standard deviation

• Assuming mean of 100, precision of ± 5 in the mean and with

95% confidence interval and unknown/large population

1800 Sample Sample

1600 standard Size
1400 deviation Required
Sample size required

1200
5 4
1000
10 16
800
20 62
600
30 139
400
40 246
200
0
50 385
50
75 865
10

20

30

40

75
5

100
100 1537
Estimated population size
 Sample size (means) – effect of population

• Assuming mean of 100, precision of ± 5 in the mean, sample

standard deviation of 50 and with 95% confidence interval

450 Sample Size

400 Population Required
350 Unknown (>100,000) 385
Sample size required

300 10,000 370

250
1,000 278
200
500 218
150
200 132
100
100 80
50
50 45
0
30 28
50

30

25

20
500

200

100
10,000
(>100,000)

1,000
Unknown

25 24
Estimated population size 20 20
Quantities related to the research question (defined by
the researcher)

v Size of the measure of interest to be detected

v Difference between two or more means
v Difference between two or more proportions

v Odds ratio, Relative risk, Correlation, Regression

coefficients
2
v Change in R , etc

v The magnitude of these values depend on

the research question and objective of the
study (for example, clinical relevance)
 Clinical Effect Size

“What is a meaningful difference between the groups”

• It is truly an estimate and often the most challenging
aspect of sample size planning
• Large difference – small sample size
• Small differences – large sample size
- Cost/benefit
Example 1: Does the consumption of large doses of
vitamin A in tablet form prevent breast cancer?

• Suppose we know from our tumor-registry data that

incidence rate of breast cancer over a 1-year period
for women aged 45 – 49 is 150 cases per 100,000

• Women randomized to Vitamin A vs. placebo

Example 1 continued

• Group 1: Control group given placebo pills. Expected to

have same disease rate as registry (150 cases per
100,000)

• Group 2: Intervention group given vitamin A tablets.

Expected to have 20% reduction in risk (120 cases per
100,000)

• Want to compare incidence of breast cancer over 1-year

• Planned statistical analysis: Chi-square test to compare

two proportions from independent samples

H0: p1 = p2 vs. HA: p1 ¹ p2

Example 1: Does ingestion of large doses of vitamin A prevent
breast cancer?
• Test H0: p1 = p2 vs. HA p1 ¹ p2

• Assume 2-sided test with a=0.05 and 80% power

• p1 = 150 per 100,000 = .0015

• p2 = 120 per 100,000 = .0012 (20% rate reduction)
• D = p1 – p2 = .0003
• z1-a/2 = 1.96 z1-b = .84

• n per group = 234,882

• Too many to recruit in one year!

 Sample Size:X-Sectional, Cohort, & Randomized Clinical Trials
Two-sided significance level(1-alpha): 95
Power(1-beta, % chance of detecting): 80
Ratio of sample size, Unexposed/Exposed: 1
Percent of Unexposed with Outcome: 0.0015
Percent of Exposed with Outcome: 0.0012
Odds Ratio: 0.8
Risk/Prevalence Ratio: 0.8
Risk/Prevalence difference: 0.0

Kelsey Fleiss Fleiss with CC

Sample Size - Exposed 23546695 23546694 24208779

Sample Size-Nonexposed 23546695 23546694 24208779

Total sample size: 47093390 47093388 48417558

Example 2: Does a special diet help to reduce
cholesterol levels?

• Suppose an investigator wishes to determine sample

size to detect a 10 mg/dl difference in cholesterol
level in a diet intervention group compared to a
control (no diet) group

• Subjects with baseline total cholesterol of at least

300 mg/dl randomized
Example 2 continued

• Group 1: A six week diet intervention

• Group 2: No changes in diet

• Investigator wants to compare total cholesterol at the

end of the six week study

• Planned statistical analysis: two sample t-test (for

independent samples)(comparison of two means)
H0: µ1 = µ2 vs. HA: µ1 ¹ µ2
Sample Size Formula

To Compare Two Means From Independent Samples: H0:

µ1 = µ2

1. a level
2. b level (1 – power)
3. Expected population difference (D= |µ1 - µ2|)
4. Expected population standard deviation (s1 , s2)
Example 2: Does a special diet help to reduce cholesterol
levels?

• Test H0: µ1=µ2 vs. HA : µ1¹µ2

• Assume 2-sided test with a=0.05 and 90% power

• d = µ1 - µ2 = 10 mg/dl
• s1= s2 = (50 mg/dl)
• za = 1.96 zb = 1.28

• n per group = 525

• Suppose 10% loss to follow-up expected,

adjust n = 525 / 0.9 = 584 per group
Problem (comparison of two means)

• A study is to be done to determine effect of 2 drugs

(A and B) on blood glucose level. From previous
studies using those drugs, Sd of BGL of 8 and 12 g/dl
were obtained respectively.
• A significant level of 95% and a power of 90% is
required to detect a mean difference between the
two groups of 3 g/dl. How many subjects should be
include in each group?
The following steps constitute a pragmatic approach to decision
taking on Sample size:

(1) Remember that there is no stock answer.

(2) Initiate early discussion among research team members.
(3) Use correct assumptions – consider various possibilities.
(4) Consider other factors also– eg., availability of cases, cost,
time.
(5) Make a balanced choice
(6) Ask if this number gives you a reasonable prospect of coming
to useful conclusion.
(7) If yes, proceed if no, reformulate your problem for study.
Summary

• Define research question well

• Consider study design, type of response variable, and
type of data analysis
• Decide on the type of difference or change you want
to detect (make sure it answers your research
question)
• Choose a and b
• Use appropriate equation for sample size calculation
or sample size tables/ nomogram or software.