epiR (version 2.0.88)

epi.ssdxsesp: Sample size to estimate the sensitivity or specificity of a diagnostic test

Description

Sample size to estimate the sensitivity or specificity of a diagnostic test.

Usage

epi.ssdxsesp(se, sp, Py, epsilon, error = "relative", 
   nfractional = FALSE, conf.level = 0.95)

Value

A data frame listing: se, the prior estimate of diagnostic senstivity as entered by the user, sp, the prior estimate of diagnostic specificity as entered by the user, Py the estimate of the prevalence of the outcome in the population as entered by the user, se.n the sample size to estimate diagnostic sensitivity, sp.n the sample size to estimate diagnostic specificity and total.n the total sample size for the study (which is the maximum of se.n and sp.n).

Arguments

se

scalar number, the prior estimate of diagnostic test sensitivity (0 to 1).

sp

scalar number, the prior estimate of diagnostic test specificity (0 to 1).

Py

scalar number, an estimate of the prevalence of the outcome in the study population.

epsilon

scalar number, the maximum difference between the estimate and the unknown population value expressed in absolute or relative terms.

error

character string. Options are absolute for absolute error and relative for relative error.

nfractional

logical, return fractional sample size.

conf.level

scalar number, the level of confidence in the computed result.

References

Buderer, NM (1996). Statistical methodology: I. Incorporating the prevalence of disease into the sample size calculation for sensitivity and specificity. Academic Emergency Medicine, 3: 895 - 900. DOI: 10.1111/j.1553-2712.1996.tb03538.x

Hajian-Tilaki K (2014). Sample size estimation in diagnostic test studies of biomedical informatics. Journal of Biomedical Informatics 48: 193 - 204. DOI: 10.1016/j.jbi.2014.02.013.

van Hal S, Lahra M, Whiley D (2025) Determining targets for ceftriaxone resistance detection in Neisseria gonorrhoeae based on genotype-phenotype data: An observational study. Lancet Microbe: 101199. DOI: 10.1016/j.lanmic.2025.101199.

Examples

## EXAMPLE 1 (from Hajian-Tilaki 2014, p 195):
## A new diagnostic test has been developed and we'd like to conduct a study
## to determine its diagnostic sensitivity which we believe should be in the
## order of 0.80. How many subjects should be enrolled if the prevalence of 
## the disease outcome of interest is 0.10 and we'd like to be 95% confident 
## that our estimate of sensitivity is within 0.07 of the true population 
## value? For this example we want to estimate diagnostic sensitivity (only)
## so the estimated performance measure for specificity is set to NA.

epi.ssdxsesp(se = 0.80, sp = NA, Py = 0.10, epsilon = 0.07, 
   error = "absolute", nfractional = FALSE, conf.level = 0.95)

## A total of 1255 subjects need to be enrolled to meet the requirements of the 
## study.


## EXAMPLE 2 (from van Hal et al. 2025):
## van Hal et al (2025) determined targets for ceftriaxone resistance detection 
## in Neisseria gonorrhoeae based on genotype-phenotype. To estimate 
## diagnostic testing performance with high precision [sic], a sample size 
## of 8295 distinct MIC-geonotype pairings was required, to achieve a 
## sensitivity and specificity of 95%, based on a ceftriaxone resistance 
## prevalence of 0.22% --- a value reflecting current rates in high-income 
## settings with robust surveillance systems.

## Code to reproduce the sample size calculations reported in the van Hal 
## et al. (2025) paper as follows:

epi.ssdxsesp(se = 0.95, sp = 0.95, Py = 0.0022, epsilon = 0.10, 
   error = "absolute", nfractional = FALSE, conf.level = 0.95)