epiR (version 0.9-22)

epi.conf: Confidence intervals for means, proportions, incidence, and standardised mortality ratios

Description

Computes confidence intervals for means, proportions, incidence, and standardised mortality ratios.

Usage

epi.conf(dat, method = "mean.single", conf.level = 0.95)

Arguments

dat

the data, either a vector or a matrix depending on the method chosen.

method

a character string indicating the method to use. Options are mean.single, mean.unpaired, mean.paired, prop.single, prop.unpaired, prop.paired, prevalence, inc

conf.level

magnitude of the returned confidence interval. Must be a single number between 0 and 1.

Value

Details

Returns confidence interval bounds at the alpha level specified. The methodology implemented here follows Altman, Machin, Bryant, and Gardner (2000). Exact binomial confidence limits are calculated for inc.risk, prevalence and inc.rate (see Collett 1999 for details). Where method is inc.risk, prevalence or inc.rate if the numerator equals zero the lower bound of the confidence interval estimate is set to zero. Where method is smr the method of Dobson et al. (1991) is used.

References

Altman DG, Machin D, Bryant TN, and Gardner MJ (2000). Statistics with Confidence, second edition. British Medical Journal, London, pp. 28 - 29 and pp. 45 - 56. Collett D (1999). Modelling Binary Data. Chapman & Hall/CRC, Boca Raton Florida, p. 24. Dobson AJ, Kuulasmaa K, Eberle E, and Scherer J (1991). Confidence intervals for weighted sums of Poisson parameters. Statistics in Medicine 10: 457 - 462. Fleiss JL (1981). Statistical methods for rates and proportions. 2nd edition. John Wiley & Sons, New York. Rothman KJ (2002). Epidemiology An Introduction. Oxford University Press, London, pp. 130 - 143.

Examples

## EXAMPLE 1
dat <- rnorm(n = 100, mean = 0, sd = 1)
epi.conf(dat, method = "mean.single")

## EXAMPLE 2
group <- c(rep(1, times = 5), rep(2, times = 5))
val = round(c(rnorm(n = 5, mean = 10, sd = 5), 
   rnorm(n = 5, mean = 7, sd = 5)), digits = 0)
dat <- as.data.frame(cbind(group, val))
epi.conf(dat, method = "mean.unpaired")

## EXAMPLE 3
grp1 <- as.vector(round(rnorm(n = 100, mean = 10, sd = 5), digits = 0))
grp2 <- as.vector(round(rnorm(n = 100, mean = 7, sd = 5), digits = 0))
dat <- as.data.frame(cbind(grp1, grp2))
epi.conf(dat, method = "mean.paired")

## EXAMPLE 4
## Single sample (Altman et al. 2000, page 47):
## Out of 263 giving their views on the use of personal computers in 
## general practice, 81 thought that the privacy of their medical file
## had been reduced.

pos <- 81
neg <- (263 - 81)
dat <- as.matrix(cbind(pos, neg))
round(epi.conf(dat, method = "prop.single"), digits = 3)

## The 95\% confidence interval for the population value of the proportion
## of patients thinking their privacy was reduced was from 0.255 to 0.366. 

## EXAMPLE 5
## Two samples, unpaired (Altman et al. 2000, page 49):
## Goodfield et al. report adverse effects in 85 patients receiving either
## terbinafine or placebo treatment for dermatophyte onchomychois.
## Out of 56 patients receiving terbinafine, 5 patients experienced
## adverse effects. Out of 29 patients receiving a placebo, none experienced
## adverse effects.

grp1 <- matrix(cbind(5, 51), ncol = 2)
grp2 <- matrix(cbind(0, 29), ncol = 2)
dat <- as.matrix(cbind(grp1, grp2))
round(epi.conf(dat, method = "prop.unpaired"), digits = 3)

## The 95\% confidence interval for the difference between the two groups is
## from -0.038 to +0.193.

## EXAMPLE 6
## Two samples, paired (Altman et al. 2000, page 53):
## In a reliability exercise, 41 patients were randomly selected from those
## who had undergone a thalium-201 stress test. The 41 sets of images were
## classified as normal or not by the core thalium laboratory and, 
## independently, by clinical investigators from different centres.
## Of the 19 samples identified as ischaemic by clinical investigators 
## 5 were identified as ischaemic by the laboratory. Of the 22 samples 
## identified as normal by clinical investigators 0 were identified as 
## ischaemic by the laboratory. 

dat <- as.matrix(cbind(14, 5, 0, 22))
round(epi.conf(dat, method = "prop.paired", conf.level = 0.95), digits = 3)

## The 95\% confidence interval for the population difference in 
## proportions is 0.011 to 0.226 or approximately +1\% to +23\%.

## EXAMPLE 7
## A herd of cattle were tested for brucellosis. Two samples out of 200
## test returned a positive result. Assuming 100\% test sensitivity and 
## specificity, what is the estimated prevalence of brucellosis in this 
## group of animals?

pos <- 4
pop <- 200
dat <- as.matrix(cbind(pos, pop))
epi.conf(dat, method = "prevalence") * 100

## The estimated prevalence of brucellosis in this herd is 2 cases
## per 100 cattle (95\% CI 0.54 -- 5.0 cases per 100 cattle).

## EXAMPLE 8
## The observed disease counts and population size in four areas are provided
## below. What are the the standardised morbidity ratios of disease for each
## area and their 95\% confidence intervals?

obs <- c(5, 10, 12, 18)
pop <- c(234, 189, 432, 812)
dat <- as.matrix(cbind(obs, pop))
round(epi.conf(dat, method = "smr"), digits = 2)