Sample Size Calculator provided by United Bristol Healthcare NHS Trust

Expected incidence Degree of

Population - N Confidence level - c Sample size
('best guess') - p accuracy - A
500 50% -0.05 95% 217
500 50% -0.1 90% 176
500 40% -0.05 95% 213
500 20% -0.05 95% 165
500 5% -0.05 95% 64
500 50% -0.025 95% 378
500 5% -0.025 95% 185

Created by Eleanor Ferris, January 2005

 Sample Size Calculator provided by United Bristol Healthcare NHS Trust

to calculate sample size,

amend variables in bold
below

N 400 population
p 70% expected incidence
A 0.05 accuracy
c 1.96 c = 1.96 for 95% confidence, or 1.645 for 90% confidence

formula
result = 178.61

therefore
n= 179

Created by Eleanor Ferris, January 2005

 Sample Size Calculator provided by United Bristol Healthcare NHS Trust

n= c2Np(1-p)
(A2N) + (c2p[1-p])

n is the sample size required

N is the whole target population in question
p is the average proportion of records expected to meet the various criteria (1-p) is the average proportion of records not expected to meet the criteria
A is the margin of error deemed to be acceptable (calculated as a proportion) e.g. for 5% error either way A = 0.05
c is a mathematical constant defined by the Confidence Interval chosen ie (how sure we need to be of the result)

Values for c
To be 95% sure of the result the constant c = 1.96
To be 90% sure of the result the constant c = 1.645
To be 80% sure of the result the constant c = 1.28

For most audits, the use of 90% Confidence levels with a margin of error of +/- 5% is reasonable

with acknowledgement to Cheshire MAAG for description of formula

www.postgradgpliv.com/Vocational_Training/summ_assessment/Audit/audit_concise_guide.htm#Appendix%201%20-%20Sample%20Sizes

Created by Eleanor Ferris, January 2005