statisticsforu.com
Online statistical service
Sample size estimation in
Medical Research
Dr. I. Kannan Ph.D
Associate Professor of Microbiology
Tagore Medical College and Hospital
Chennai – 600127
dr.ikannan@tagoremch.com
statisticsforu.com
Online statistical service
What is sample size?
•It is the total representative samples from
the given population used for the research
study.
statisticsforu.com
Online statistical service
Why only representative samples?
•Difficult to subject the entire
population for the research because
of
➢Economical reason
➢Ethical reason
➢Time constraints
statisticsforu.com
Online statistical service
How the sample size should be?
•Should be optimal, neither high nor less.
•High
➢Study will become costly and time consuming
➢Unethical to include participants than the
required numbers.
•Less
➢Validity of the study is lost.
➢Low power of the study and the research
outcome is not trustable.
statisticsforu.com
Online statistical service
What is the power of study?
•The power is the probability to reject the null
hypothesis (Ho), given that the null hypothesis is
false.
•Minimum expected probability is 80%.
•It is applicable to those studies wherein there is null
and alternative hypothesis.
•The prevalence studies do not have any hypotheses
and thus power of the study is not applicable.
statisticsforu.com
Online statistical service
What is power?
Reality
Decision
based on
research
H0 is true H0 is false
Reject H0 Type I error ()
Correct Decision
(1 – β)
Power
Accept H0
Correct Decision
(1 – )
Type II error (β)
1 minus Type II error (β) is the power of the study
statisticsforu.com
Online statistical service
Power is determined by?
•The power of the study depends on
•Sample size
•Alpha level set in the study (it is
optimal to set at 5% [0.05] in medical
research)
•Effect size
statisticsforu.com
Online statistical service
Sample size and power
•Increase in sample size increases the power
of the study.
•80% and above of power is considered to be
enough in medical research.
•During sample size calculation, it is
mandatory to calculate the expected power
of the study to justify your sample size.
statisticsforu.com
Online statistical service
Strategy to determine the sample size
•Different formulae based on the study design
•Data obtained from previous similar studies are
used in the formula to determine the sample
size.
•If there is no similar studies, the researcher has
to perform a pilot study to obtain the values that
can be used for sample size calculation.
statisticsforu.com
Online statistical service
Prevalence studies
statisticsforu.com
Online statistical service
Sample size estimation WITHOUT finite
population size*
For the level of confidence of 95%, which is conventional, Z value is 1.96.
*Daniel WW (1999). Biostatistics: A Foundation for Analysis in the Health Sciences.
7thedition. New York: John Wiley & Sons.
statisticsforu.com
Online statistical service
Sample size estimation WITHOUT finite
population size (Caution)
•This sample size formula is valid if the calculated sample
size is smaller than or equal to 5% of the population size
•Sample size calculation without considering total
population size should be avoided because:
➢It may overshoot the 5% of total population size.
➢Sometimes it may even overshoot the total
population size.
statisticsforu.com
Online statistical service
Sample size estimation WITH finite population
size*
*Daniel WW (1999). Biostatistics: A Foundation for Analysis in the Health Sciences.
7thedition. New York: John Wiley & Sons.
statisticsforu.com
Online statistical service
Binary outcome
statisticsforu.com
Online statistical service
Formula for a binary outcome
Formula for a binary outcome and equal sample sizes in both groups, assuming: alpha = 0.05 and
power =0.80 (beta = 0.20).
statisticsforu.com
Online statistical service
Formula for a binary outcome
• When the significance level alpha is chosen at 0.05 one should enter
the value 1.96 for a in the formula.
• When beta is chosen at 0.20, the value 0.842 should be filled in for b
in the formula.
statisticsforu.com
Online statistical service
Formula for a binary outcome
• X in the formula is the minimal difference between
the groups that the investigator considers biologically
plausible and clinically relevant.
statisticsforu.com
Online statistical service
Continuous outcome
statisticsforu.com
Online statistical service
Formula for a continuous outcome
Formula for a continuous outcome and equal sample sizes in both groups, assuming: alpha
= 0.05 and power = 0.80 (beta = 0.20)
statisticsforu.com
Online statistical service
Final comments
• There are many books that gives the sample size
estimation formulae.
• Many software are available to determine the sample
size which has the advantage of calculating the
expected power as well.

Sample size calculation in medical research