Indian J Pediatr (November 2012) 79(11):14821488
DOI 10.1007/s12098-012-0763-3
SPECIAL ARTICLE
Sample Size Estimation in Prevalence Studies
Ravindra Arya & Belavendra Antonisamy &
Sushil Kumar
Received: 8 August 2011 / Accepted: 5 April 2012 / Published online: 6 May 2012
# Dr. K C Chaudhuri Foundation 2012
Abstract Estimation of appropriate sample size for preva-
lence surveys presents many challenges, particularly when
the condition is very rare or has a tendency for geographical
clustering. Sample size estimate for prevalence studies is a
function of expected prevalence and precision for a given
level of confidence expressed by the z statistic. Choice of
the appropriate values for these variables is sometimes not
straight-forward. Certain other situations do not fulfil the
assumptions made in the conventional equation and present
a special challenge. These situations include, but are not
limited to, smaller population size in relation to sample size,
sampling technique or missing data. This paper discusses
practical issues in sample size estimation for prevalence
studies with an objective to help clinicians and healthcare
researchers make more informed decisions whether review-
ing or conducting such a study.
R. Arya
Comprehensive Epilepsy Centre, Division of Neurology,
Cincinnati Childrens Hospital Medical Centre,
Cincinnati, OH, USA
B. Antonisamy
Department of Biostatistics, Christian Medical College,
Vellore, Tamil Nadu, India
S. Kumar
Division of Basic and Translational Research,
Department of Surgery, University of Minnesota,
Minneapolis, MN, USA
R. Arya (*)
MLC 2015, Cincinnati Childrens Hospital Medical Centre,
3333 Burnet avenue,
Cincinnati, OH 45229, USA
e-mail: Ravindra.Arya@cchmc.org
Keywords Sample size estimation . Prevalence . Precision
Introduction
Sample size estimation is a key issue in design of most
studies. In a study conducted to estimate prevalence of a
given condition in a geographic area, the objective is to
sample sufficient population to get adequate number of
subjects correctly classified as having the condition of in-
terest or not, with a given confidence about the amount to
which this estimate might be affected by sampling error.
Several factors may potentially contribute to difficulty in
this regard, including misclassification of diseased and
healthy subjects due to limitations of diagnostic test, the
given condition being very common or uncommon, or,
limitations of sampling technique. This paper discusses the
issues involved in sample size estimation for prevalence
studies with an objective to provide practically useful infor-
mation to clinicians and healthcare researchers.
Let us say our objective is to determine the prevalence of
a disease condition in the population living in an isolated
geographic area. Here, one simply cannot perform the diag-
nostic test on the whole population. Such an approach
would be impractical due to logistic or financial reasons,
and, ethically unacceptable. Moreover, there would still be
misclassification because the predictive values of the test
would not be 100 %. So our target is to find out sufficient
number of people sampled at random from this population,
which if subjected to diagnostic testing, would yield an
estimate of the prevalence of the condition in the given
population with some confidence. It can be argued that if
the condition is very rare in the given population, a larger
sample would be required to yield sufficient number of
Indian J Pediatr (November 2012) 79(11):14821488
cases and non-cases; if it is relatively common we might
need a rather small number.1 Hence, sample size n is directly
proportional to the prevalence of the disease P in the popu-
lation. Secondly, we have to decide how precisely we want
to estimate this prevalence. This is usually measured as the
amount of acceptable (or allowable) margin of error in the
prevalence estimate. This is a random sampling error which
decreases on repeated trials. Hence, n is inversely propor-
tional to allowable error d, which is a surrogate measure of
precision. These relationships can be expressed by a simple
equation.
z2 P1  P
n
d2
Where n0sample size, z0z statistic for the level of
confidence, P 0expected prevalence and d 0allowable
error. This formula assumes that P and d are decimal
values, but would hold correct also if they are percen-
tages, except that the term (1-P) in numerator would
become (100-P). In this straight forward equation
(Box 1), several practical issues arise in the choice of
values for z, P and d.
Precision
Precision of any measurement system is the degree to
which repeated measurements under unchanged condi-
tions generate similar results. It captures the extent of
reproducibility or repeatability of measurements under
identical measuring conditions. There are different ways
of expressing precision e.g., in arithmetic it is usually
expressed by last significant place. To illustrate this,
consider a value expressed as 8. It would mean that
the measurement has been made with a precision of 1
(the measuring instrument was able to measure only
down to units place) whereas a value of 8.0, though
numerically equal to 8, would mean that the value at
the first decimal place was measured and was found to
be zero. The second value is more precise. When meas-
urements are repeated and averaged, the standard error
of mean serves as an indicator of precision. In preva-
lence studies, results might be stated as the prevalence
of this condition was found in this sample to be 30 %
[95 % confidence interval (CI) 20 to 40 %]. This
statement can be interpreted that, if we were to repeat
our study 100 times, 95 times our estimated prevalence
is likely to fall between 2040 % (Frequentist way),
and, in this particular study we are 95 % confident that the
1
If a trait is too common, we would again run into problems, as we
would now be short of non-cases. We will address this in a while.
1483
true prevalence lies between 20 to 40 % (Bayesian way).
Sample size estimation for prevalence surveys has 2 inter-
related measures of precision, which are discussed subse-
quently: CI and the allowable margin of error d.
Z Statistic
Whenever we make inferences about population from a
sample, an element of uncertainty is introduced. One way
to quantify this uncertainty is the use of CI. Z statistic
indicates the number of standard deviations an observation
is above or below the population mean. It captures the
level of confidence, assuming a normal distribution. For
conventional 95 % confidence level, the z value is 1.96,
since 95 % of a normal distribution would lie within 1.96
standard deviations on either side of the mean. The value of
z is chosen by the investigator according to the desired level
of confidence.
Allowable Margin of Error (d)
Allowable error is the maximum risk in the sample size
estimate acceptable to the clinician and should be de-
cided a priori. In any protocol or manuscript, it should
be stated explicitly along with the basis for its choice.
Conventionally, an absolute allowable error margin d
of 5 % is chosen, but, as is common in clinical
practice, if expected prevalence P is <10 %, the 95 %
confidence boundaries may cross 0, which is impracti-
cal. Hence, for an expected prevalence between 10 to
90 % (P00.1 to 0.9), d of 5 % might be a reasonable
choice. But for rare (P<0.1) or very common (P>0.9)
conditions, d should be chosen as a relative value with
respect to expected prevalence P. A common recommen-
dation is to set d 0 P/2 for rare and d 0 (1-P)/2 for very
common conditions [1]. The choice of relative allow-
able error as opposed to an absolute value, is indepen-
dent of expected prevalence and one might choose it for
mid-range values of P which is a valid approach. Inves-
tigators may sometimes wish to choose a smaller d,
which would increase the sample size considerably,
since n is inversely proportional to d2. Conversely, if
the sample size is impractical, investigator may be
tempted to increase the allowable error and get a
smaller n. Instead, the researcher could state that due
to resource constraints, the ideal sample size cannot be
met and hence, a smaller and pragmatic sample is
investigated. Also, a larger d might invalidate the as-
sumption of normal approximation. It should be appre-
ciated that allowable margin of error in the sample size
estimate is a conscious choice of investigator, whereas,
1484
Box 1 The missing
the element of chance is captured by the chosen CI or Z
statistic.
Expected Prevalence (P)
This might seem like a paradox of sorts. Our objective in
conducting the study is to determine the prevalence of a
particular condition in a particular population i.e., it is this P
which we are out to find. But to do so, we need to have a
prior idea of it! Mostly, this idea can be had from prior
studies. Usually, the previous studies will give a range of
P and not a single number. It has been suggested that one
should err towards P00.5 in these situations [2]. That is, if
the range provided by previous studies is 3040 %, then P
should be set at 0.4, whereas if the range is 7080 %, it is
better to take P00.7, i.e., the value nearer to 0.5 or 50 %.
This recommendation is based on the fact that choosing a
value nearer to 0.5 leads to the largest n, within certain
limitations (Fig. 1).
Indian J Pediatr (November 2012) 79(11):14821488
Suppose one is doing the first ever prevalence study for a
particular condition in a given population, then there is no
previous study to help estimate P. In such a situation, some
authors [3] recommend that n may be calculated using
P00.5. As ascertained from Fig. 1, this contention is valid
if P lies between 10 to 90 %, as it will give the largest
estimate for n. However, for rare (P<0.1) and very common
(P>0.9) conditions, the sample size estimated with an as-
sumption of P00.5 is likely to be unsuitable. Suppose we
plug values of P00.5, d00.05 in our formula and obtain n0
385 (Fig. 1). This will be the empirical sample size estimate
for all purposes, when we make a blind assumption of P0
0.5. Assume that we are dealing with a rare condition having
a true prevalence of only 1 %, that is to say P00.01. In such
a case, our sample is likely to capture only 3 or 4 cases.
There is a finite probability that it may not capture even a
single case! On the contrary if we are dealing with a very
common condition, say whose prevalence is 99 % (P00.99),
then we will not be able to capture sufficient non-cases with
our sample size and as above, we may indeed capture none.
Indian J Pediatr (November 2012) 79(11):14821488 1485

Fig. 1 Sample size n (y-axis) as

a function of expected
prevalence P (x-axis) for
Z=1.96 and d=0.05, has a
maxima at n=385 for P=0.5.
For P=1 the equation returns an
indeterminate value

In either case, the assumption of normal approximation is the best strategy is to conduct a pilot study, estimate P,
invalidated. and use it for calculating a sample size.
At a minimum this requires an estimate as to whether
P is <0.1, falls between 0.1 and 0.9, or is >0.9. If we
can estimate this, our value for allowable error (d) will Assumption of Normal Approximation
change, as discussed above, and we will get a more
meaningful calculation for sample size (n). Table 1 The formula for sample size estimation is based on the
illustrates this calculation for few representative values assumption that the sample will capture and correctly clas-
of P<0.05 and P>0.95. Notice how n is different for sify certain minimum number (05) of cases and non-cases.
these extremes of P from the empirical figure of 385, This means that nP and n(1-P) must be Q5. We know that
increasing symmetrically as we approach P00 or P01. allowable error is directly related to standard error of pro-
It is obvious intuitively that at P 00, n will become portion (SEp) by the equation:
infinite. If there are no cases of the given condition in
a given population, no matter how large a sample we d z  SEp
take, we will never find one! Similar argument holds
true for P01. So, our problem boils down to finding Also:
whether P is <0.1, between 0.1 and 0.9 or >0.9. This r
can be easily estimated by even a crude pilot study. To P 1  P
SEp
summarize, if we have an estimate for P from prior n
studies, we can use it erring towards P00.5; otherwise,
Using above 2 equations and re-arranging, we can get the
formula for n, discussed above. For z01.96 and d00.05, the
equation would be:
Table 1 Calculating n 0:05 1:96  SEp
for very rare (P<0.05) z d P n
and very common
(P>0.95) conditions 1.96 0.005 0.01 1521.27 This means that 1.96 standard errors of our estimate
1.96 0.01 0.02 752.95 would be equal to 0.05. Stated otherwise, if 1.96*SEp is
1.96 0.015 0.03 496.85 equal to 0.05, then our sample size estimate has a 95 %
1.96 0.02 0.04 368.79 chance of being within 5 percentage points of the true
1.96 0.025 0.05 291.96 prevalence. This inherently assumes that n is sufficiently
1.96 0.025 0.95 291.96 large, so that all possible values of P will have a normal
1.96 0.02 0.96 368.79 distribution. It also means that if we were to repeat our
1.96 0.015 0.97 496.84
study, about 95 % of times our prevalence estimate will fall
1.96 0.01 0.98 752.95
within the value specified by P1.96*SEp. This indeed,
Precision is calculated 1.96 0.005 0.99 1521.27
shows that these values have a normal distribution with a
as discussed in text mean of P and standard deviation of d/Z, approximately
1486
(Box 2). This is nothing but approximating binomial distri-
bution with a normal distribution, justified by the central
limit theorem. In this context, P is the proportion of suc-
cesses in a Bernoulli trial process estimated from a sample
of size n with z(1-/2) as the (1-/2) percentile of standard
normal distribution and as the error percentile. The central
limit theorem is applicable to a binomial distribution when
the proportion is not very close to 0 or 1 i.e., the normal
approximation fails when sample proportion approaches
these values.
Finite Population Correction
The above sample size formula is valid only for samples
which are relatively small as compared to total population. If
total population is N, then n/N should be e0.05. When the
sampling fraction is larger than approximately 5 %, the
estimate of standard error, used in the sample size formula,
Box 2 Worked example
illustrating principles of sample
size estimation in prevalence
study
Indian J Pediatr (November 2012) 79(11):14821488
must be corrected by multiplying with a finite population
correction (FPC), to account for the added precision gained
by sampling close to a larger percentage of population.
p
FPC N  n=N  1
The effect of FPC is annulled when n 0 N. The formula
for sample size calculation including the effect of FPC is:
0 N z2 P1  P
n
d 2 N 1 z2 P1  P
Where n is the sample size with finite population
correction, N is the population size, Z is the statistic
for the level of confidence, P is expected proportion and
d is allowable error. The need for FPC arises because
typically the sampling method in prevalence studies is
without replacement (hypergeometric sampling), but the
sample size estimation method is based on sampling
with replacement (binomial distribution). When the tar-
,
Indian J Pediatr (November 2012) 79(11):14821488
get population is very large (theoretically infinite),
hypergeometric distribution can be well approximated
by a binomial one, and thus, formula based on latter
is acceptable. Any investigator using FPC in sample
size estimate should also incorporate adjustment for
hypergeometric sampling in analysis.
Design Effect
Above calculation for sample size assume independence
among sampling units i.e., it is only valid for simple or
systematic random sampling. Lack of independence intro-
duced due to other sampling methods should be adjusted by
changing the variance estimate e.g., for clustered sampling,
the design effect (DE) or variance inflation factor is defined
as:
DE 1 m  1
Where is the intra-class correlation and m is the sample
size within each cluster. The intra-class correlation is a
relative measure of the homogeneity of sampling units with-
in each cluster compared with a randomly selected unit,
defined as the proportion of total variation within sampling
units that can be accounted by variation among clusters. It is
generally estimated from prior literature or pilot data. If the
number of clusters is fixed by design and the sample size per
cluster (m) is unknown then that has to be estimated first as:
ES 1 
m comparative estimates of the frequencies of rare conditions
k  ES
Where ES is the effective sample size assuming indepen-
dence and k is the number of clusters. Finally, the sample
size (n) for cluster or multi-stage sampling is calculated
using:
0
n n  DE
Where n is the sample size estimated assuming a simple
random sampling. DE is actually a ratio measure that
describes how much precision is gained or lost if a more
complex sampling strategy is used instead of simple random
sampling. Usually, complex sampling techniques lead to a
decrease of precision, resulting in DE >1. This also implies
that for the same precision, cluster or multistage sampling
technique would require a larger sample size than simple or
systematic random sampling. A classic example is cluster
surveys for immunization coverage, where DE has shown to
be approximately 1.9 [2].
Sample size calculations are based on large sample ap-
proximation methods. Together FPC, DE and continuity
correction are adjustment factors which help improve spe-
cific sample size approximation to exact distributions.
1487
Continuity correction factor is employed to compare two
population proportions, usually in a diagnostic evaluation,
and is not discussed further here [4].
Non-Response Correction
Formula for sample size estimation inherently assumes that
for a chosen CI or d, there will be no non-response or
missing data, a condition rarely met. Suppose we wish to
estimate the prevalence of depression in a given population
using a questionnaire. In real life situation, it is improbable
that every one of our subjects will meticulously fill the
survey form and return it, resulting in some loss to our
estimated sample size. This loss is difficult to estimate
upfront, and, is usually a clinical decision. If n is the desired
sample size, then for an expected loss r we would have to
oversample to n such that:
0 0
n  rn n
Solving for n:
0 n
n
1r
Zero Patient Method
A very interesting method has been proposed to provide
even when the true prevalences remain unknown [5]. Let us
say we take a sample of n people from a population such
that all people in our sample are free of disease. Then the
95 % upper bound CI for the prevalence P of the disease can
be computed from the equation:
a 1  Pn
This corresponds to rejection of the null hypothesis that
the true prevalence is P at a 1-sided significance level of .
Taking the natural logarithm, setting 00.05 and re-
arranging, we get a simple relationship:
ln0:05 3
n
P P
This simplification makes an assumption that ln
1  P P , which holds true only for P < 0.02.
Hence, there are two important things in using this
formula for sample size estimation. First, a previous
estimate of the prevalence in a region with relatively
high disease frequency, determined by conventional
methods, will be helpful in determining the size of
population to be screened in a region thought to be
1488
having lesser prevalence. Secondly, this formula will be
suitable only for conditions with prevalence less than
2 % in the target population undergoing screening.
For example, Gaucher disease type 3 is relatively com-
mon in the population of Northern Swedish region of Nor-
botten having a prevalence of approximately 0.0006. So, if
one screens about 3/0.000605000 people from any popula-
tion and does not find a case of this disease in the sample,
then one could say with 95 % confidence that the prevalence
of Gaucher disease type 3 is less than 0.0006 in that popu-
lation. Extending this argument, if 10000 people from the
same population are screened for Gaucher disease type 3
and not a single case is found, then a conclusion can be
made with 95 % confidence that the prevalence of this
disease is less than 0.0003 in the given population.
If we introduce the concept of statistical power in this
formula, it will represent the probability of all persons
within the sample of size n being disease free, given the
true prevalence of disease in the population is p (alternative
hypothesis) instead of P (null hypothesis), such that p<<P.
In such a case, power () will be:
 
0 n
1P
Taking natural logarithms and substituting n03/P and
00.8 yields P00.074*P. In the example of Gaucher dis-
ease type 3, this means that assuming the null prevalence in
Northern Swedish region of 0.0006, the alternative preva-
lence in our hypothetical population should be less than
(0.074*0.00060) 4.4*105 for this method to have at least
80 % power. This power will increase for yet smaller
prevalence.
This method is called zero patient design and can only
be used for estimating relative prevalence of disease in a
population, provided an estimate of its prevalence is
Indian J Pediatr (November 2012) 79(11):14821488
available in some other population having relatively higher
prevalence of it and the disease has an estimated prevalence
of less than 2 % in the population in which we want to have
the estimate. This method cannot provide an estimate of true
prevalence but can generate a hypothesis for the same.
Conclusions
Sample size estimation for prevalence studies is straight-
forward for the most part. Certain challenging situations like
extremes of prevalence, small target population, sampling
technique, expected misclassification or missing data re-
quire diligent choice of calculating formula and plug in
values.
Conflict of Interest None.
Role of Funding Source None.
References
1. Antonisamy B, Christopher S, Samuel PP. Biostatistics: principles
and practice. New Delhi: Tata McGraw-Hill; 2010.
2. Macfarlane SB. Conducting a descriptive survey: 2. Choosing a
sampling strategy. Trop Doct. 1997;27:1421.
3. Lwanga SK, Lemeshow S. Sample size determination in health
studies: a practical manual. Geneva: World Health Organization;
1991.
4. Fosgate GT. Practical sample size calculations for surveillance and
diagnostic investigations. J Vet Diagn Invest. 2009;21:314.
5. Yazici H, Biyikli M, van der Linden S, Schouten HJ. The zero
patient design to compare the prevalences of rare diseases. Rheu-
matology (Oxford). 2001;40:1212.