Title: | Freedom from Disease |
---|---|
Description: | Functions, S4 classes/methods and a graphical user interface (GUI) to design surveys to substantiate freedom from disease using a modified hypergeometric function (see Cameron and Baldock, 1997, <doi:10.1016/s0167-5877(97)00081-0>). Herd sensitivities are computed according to sampling strategies "individual sampling" or "limited sampling" (see M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, 2002, <doi:10.1016/S0167-5877(01)00245-8>). Methods to compute the a-posteriori alpha-error are implemented. Risk-based targeted sampling is supported. |
Authors: | Ian Kopacka |
Maintainer: | Ian Kopacka <[email protected]> |
License: | GPL (>= 2) |
Version: | 1.0-9 |
Built: | 2024-11-21 02:48:41 UTC |
Source: | https://github.com/r-forge/ffd |
For a vector of herd sizes the herd-based alpha-errors (= 1-herd sensitivity) are computed for either limited or individual sampling; see Ziller et al.
computeAlpha(nAnimalVec, method, sampleSizeLtd, herdSensitivity, intraHerdPrevalence, diagSensitivity, diagSpecificity = 1)
computeAlpha(nAnimalVec, method, sampleSizeLtd, herdSensitivity, intraHerdPrevalence, diagSensitivity, diagSpecificity = 1)
nAnimalVec |
Integer vector. Stock sizes of the herds. |
method |
Character string. |
sampleSizeLtd |
Integer. Required only if |
herdSensitivity |
Numeric between 0 and 1. Required only if |
intraHerdPrevalence |
Numeric between 0 and 1. Intra-herd prevalence. The number of diseased
animals per herd is computed as
|
diagSensitivity |
Numeric between 0 and 1. Sensitivity (= probability of a testpositive result, given the tested individual is diseased) of the diagnostic test. |
diagSpecificity |
Numeric between 0 and 1. Specificity (= probability of a testnegative result, given the tested individual is not diseased) of the diagnostic test. The default value is 1, i.e., perfect specificity, and is recommended. |
Returns a vector containing the herd-based alpha-errors, where each
entry in the vector corresponds to an entry in the input argument
nAnimalVec
.
Ian Kopacka <[email protected]>
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
Is used in the method sample
for classes IndSampling
and LtdSampling
.
data(sheepData) ## Compute the herd sensitivities usinh limited sampling: alphaVec <- computeAlpha(nAnimalVec = sheepData$nSheep, method = "limited", sampleSizeLtd = 7, intraHerdPrevalence = 0.2, diagSensitivity = 0.9)
data(sheepData) ## Compute the herd sensitivities usinh limited sampling: alphaVec <- computeAlpha(nAnimalVec = sheepData$nSheep, method = "limited", sampleSizeLtd = 7, intraHerdPrevalence = 0.2, diagSensitivity = 0.9)
For sampling strategy "limited sampling" (see Ziller et al., 2002) the function computes the herd-level alpha-errors (= 1-herd sensitivity) for each stock size, as well as the average herd-level alpha-error.
computeAlphaLimitedSampling(stockSizeVector, sampleSizeLtd, intraHerdPrevalence, diagSensitivity, diagSpecificity = 1, groupVec = NULL)
computeAlphaLimitedSampling(stockSizeVector, sampleSizeLtd, intraHerdPrevalence, diagSensitivity, diagSpecificity = 1, groupVec = NULL)
stockSizeVector |
Integer vector. Stock sizes of the herds. |
sampleSizeLtd |
Integer. Sample size for limited sampling, i.e., for each herd
|
intraHerdPrevalence |
Numeric between 0 and 1. Intra-herd prevalence. The number of diseased
animals per herd is computed as
|
diagSensitivity |
Numeric between 0 and 1. Sensitivity (= probability of a testpositive result, given the tested individual is diseased) of the diagnostic test. |
diagSpecificity |
Numeric between 0 and 1. Specificity (= probability of a testnegative result, given the tested individual is not diseased) of the diagnostic test. |
groupVec |
Character vector. Optional parameter. If specified it must have the same length
as |
List of 3 elements:
alphaDataFrame |
Data frame. Variables |
meanAlpha |
Numeric between 0 and 1. Mean alpha-error attained by strategy "limited sampling" for given sample size and herd size distribution. |
meanAlphaRiskGroups |
If |
Ian Kopacka <[email protected]>
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
data(sheepData) alphaList <- computeAlphaLimitedSampling(stockSizeVector = sheepData$nSheep, sampleSizeLtd = 7, intraHerdPrevalence = 0.2, diagSensitivity = 0.9, diagSpecificity = 1)
data(sheepData) alphaList <- computeAlphaLimitedSampling(stockSizeVector = sheepData$nSheep, sampleSizeLtd = 7, intraHerdPrevalence = 0.2, diagSensitivity = 0.9, diagSpecificity = 1)
For a sampling scheme designed to substantiate freedom from disease the function computes the a-posteriori alpha-error, i.e., the actual alpha-error based on the drawn sample.
computeAposterioriError(alphaErrorVector, nPopulation, nDiseased, method = "default")
computeAposterioriError(alphaErrorVector, nPopulation, nDiseased, method = "default")
alphaErrorVector |
Numeric vector. Alpha-error (between 0 and 1) of each herd in the sample. |
nPopulation |
Integer. Population size, i.e., total numer of herds in the population. |
nDiseased |
Integer. Number of diseased herds in the population according to the design prevalence. |
method |
Character string. "exact" for exact error, "approx" for approximation (recommended for nDiseased > 7). |
The exact evaluation of the alpha-error is computationally complex, due
to combinatirical issues. In order to increase effectivity parts of the code
were implemented in C. Still, for nDiseased
> 7 the computation may take
very long and it is generally not recommended to use the exact calculation. Rather
the approximation should be used for nDiseased
> 7.
The return value is the a-posteriori alpha-error based on the sample at hand (numeric scalar).
Ian Kopacka <[email protected]>
## Freedom from disease using limited sampling with sampleSizeLtd = 7. ## Data: sheep holdings in state "Steiermark". ###################################################################### data(sheepData) popVec <- sheepData$nSheep[sheepData$state == 6] N1 <- length(popVec) sampleSizeLtd <- 7 intraHerdPrev <- 0.13 designPrev <- 0.002 nDiseased <- round(designPrev*N1) ## Draw the sample: n1 <- 1560 samplePop <- sample(x = popVec, size = n1, replace = FALSE, prob = NULL) ## Compute alpha-errors for the sample: alphaList <- computeAlphaLimitedSampling(stockSizeVector = samplePop, sampleSizeLtd = sampleSizeLtd, intraHerdPrevalence = intraHerdPrev, diagSensitivity = 0.9, diagSpecificity = 1) alphaDataFrame <- merge(x = data.frame(size = samplePop), y = alphaList$alphaDataFrame, by = "size", all.x = TRUE, all.y = FALSE) ## Compute the a-posteriori alpha-error: alphaAPostApprox <- computeAposterioriError(alphaErrorVector = alphaDataFrame$alpha, nPopulation = N1, nDiseased = nDiseased, method = "approx") alphaAPostExact <- computeAposterioriError(alphaErrorVector = alphaDataFrame$alpha, nPopulation = N1, nDiseased = nDiseased, method = "exact")
## Freedom from disease using limited sampling with sampleSizeLtd = 7. ## Data: sheep holdings in state "Steiermark". ###################################################################### data(sheepData) popVec <- sheepData$nSheep[sheepData$state == 6] N1 <- length(popVec) sampleSizeLtd <- 7 intraHerdPrev <- 0.13 designPrev <- 0.002 nDiseased <- round(designPrev*N1) ## Draw the sample: n1 <- 1560 samplePop <- sample(x = popVec, size = n1, replace = FALSE, prob = NULL) ## Compute alpha-errors for the sample: alphaList <- computeAlphaLimitedSampling(stockSizeVector = samplePop, sampleSizeLtd = sampleSizeLtd, intraHerdPrevalence = intraHerdPrev, diagSensitivity = 0.9, diagSpecificity = 1) alphaDataFrame <- merge(x = data.frame(size = samplePop), y = alphaList$alphaDataFrame, by = "size", all.x = TRUE, all.y = FALSE) ## Compute the a-posteriori alpha-error: alphaAPostApprox <- computeAposterioriError(alphaErrorVector = alphaDataFrame$alpha, nPopulation = N1, nDiseased = nDiseased, method = "approx") alphaAPostExact <- computeAposterioriError(alphaErrorVector = alphaDataFrame$alpha, nPopulation = N1, nDiseased = nDiseased, method = "exact")
For a sampling scheme designed to substantiate freedom from disease the function computes the a-posteriori alpha-error, i.e., the actual alpha-error based on the drawn sample, when the population is stratified into risk groups for which the relative infection risk is known.
computeAposterioriErrorRiskGroups(alphaErrorVector, groupVec, groupLevels, nPopulationVec, nRelRiskVec, nDiseased, method = "default")
computeAposterioriErrorRiskGroups(alphaErrorVector, groupVec, groupLevels, nPopulationVec, nRelRiskVec, nDiseased, method = "default")
alphaErrorVector |
Numeric vector. Alpha-error (between 0 and 1) of each herd in the sample. |
groupVec |
Character vector. Risk group to which each of the
herds in the sample belongs. Must have the same length
(and order) as |
groupLevels |
Character vector. (Unique) levels/names of the risk group. Defines the order
of the values in |
nPopulationVec |
Integer vector. Population sizes of the risk groups. Must have the same length
(and order) as |
nRelRiskVec |
Numeric vector. (Relative) infection risks of the risk groups. Must have
the same length (and order) as |
nDiseased |
Integer. Number of diseased herds in the population according to the design prevalence. |
method |
Character string. "exact" for exact error, "approx" for approximation (recommended for nDiseased > 7). |
The exact evaluation of the alpha-error is computationally complex, due
to combinatirical issues. In order to increase effectivity parts of the code
were implemented in C. Still, for nDiseased
> 3 the computation may take
very long and it is generally not recommended to use the exact calculation. Rather
the approximation should be used for nDiseased
> 3.
The return value is the a-posteriori alpha-error based on the sample at hand (numeric scalar).
Ian Kopacka <[email protected]>
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Limited sampling with risk groups: myLtdSamplingRG <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 20, nSampleFixVec = NULL, probVec = c(1,2)) ## Draw sample manually: set.seed(20110708) x <- myLtdSamplingRG indexSampleRG <- sapply(seq(along = x@surveyData@riskValueData$riskGroup), function(ii){ riskGroup <- as.character(x@surveyData@riskValueData$riskGroup[ii]) nSampleRG <- x@nHerdsPerRiskGroup[riskGroup] indexRG <- which(x@surveyData@riskGroupVec == riskGroup) indexOut <- sample(x = indexRG, size = nSampleRG, replace = FALSE) }) indexSample <- sort(Reduce(function(x,y) c(x,y), indexSampleRG)) ## Compute the a-posteriori alpha error: alphaErrorVector <- computeAlpha(nAnimalVec = x@surveyData@nAnimalVec[indexSample], method = "limited", sampleSizeLtd = x@sampleSizeLtd, intraHerdPrevalence = x@surveyData@intraHerdPrevalence, diagSensitivity = x@surveyData@diagSensitivity, diagSpecificity = 1) ## Determine the number of herds in each risk group: riskValueDf <- mySurvey@riskValueData[,1:2] names(riskValueDf) <- c("riskGroup", "riskValues") riskValueDf$riskGroup <- as.character(riskValueDf$riskGroup) riskValueDf$id <- seq(along = riskValueDf[,1]) riskGroupTab <- table(mySurvey@riskGroupVec) riskGroupDf <- data.frame(riskGroup = as.character(names(riskGroupTab)), nPopulation = as.vector(riskGroupTab)) riskValueDf <- merge(x = riskValueDf, y = riskGroupDf, by = "riskGroup", sort = FALSE) riskValueDf <- riskValueDf[order(riskValueDf$id, decreasing = FALSE),] aPostAlphaRG <- computeAposterioriErrorRiskGroups(alphaErrorVector = alphaErrorVector, groupVec = x@surveyData@riskGroupVec[indexSample], groupLevels = riskValueDf$riskGroup, nPopulationVec = riskValueDf$nPopulation, nRelRiskVec = riskValueDf$riskValues, nDiseased = max(round(length(x@surveyData@nAnimalVec)*x@surveyData@designPrevalence),1), method = "approx")
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Limited sampling with risk groups: myLtdSamplingRG <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 20, nSampleFixVec = NULL, probVec = c(1,2)) ## Draw sample manually: set.seed(20110708) x <- myLtdSamplingRG indexSampleRG <- sapply(seq(along = x@surveyData@riskValueData$riskGroup), function(ii){ riskGroup <- as.character(x@surveyData@riskValueData$riskGroup[ii]) nSampleRG <- x@nHerdsPerRiskGroup[riskGroup] indexRG <- which(x@surveyData@riskGroupVec == riskGroup) indexOut <- sample(x = indexRG, size = nSampleRG, replace = FALSE) }) indexSample <- sort(Reduce(function(x,y) c(x,y), indexSampleRG)) ## Compute the a-posteriori alpha error: alphaErrorVector <- computeAlpha(nAnimalVec = x@surveyData@nAnimalVec[indexSample], method = "limited", sampleSizeLtd = x@sampleSizeLtd, intraHerdPrevalence = x@surveyData@intraHerdPrevalence, diagSensitivity = x@surveyData@diagSensitivity, diagSpecificity = 1) ## Determine the number of herds in each risk group: riskValueDf <- mySurvey@riskValueData[,1:2] names(riskValueDf) <- c("riskGroup", "riskValues") riskValueDf$riskGroup <- as.character(riskValueDf$riskGroup) riskValueDf$id <- seq(along = riskValueDf[,1]) riskGroupTab <- table(mySurvey@riskGroupVec) riskGroupDf <- data.frame(riskGroup = as.character(names(riskGroupTab)), nPopulation = as.vector(riskGroupTab)) riskValueDf <- merge(x = riskValueDf, y = riskGroupDf, by = "riskGroup", sort = FALSE) riskValueDf <- riskValueDf[order(riskValueDf$id, decreasing = FALSE),] aPostAlphaRG <- computeAposterioriErrorRiskGroups(alphaErrorVector = alphaErrorVector, groupVec = x@surveyData@riskGroupVec[indexSample], groupLevels = riskValueDf$riskGroup, nPopulationVec = riskValueDf$nPopulation, nRelRiskVec = riskValueDf$riskValues, nDiseased = max(round(length(x@surveyData@nAnimalVec)*x@surveyData@designPrevalence),1), method = "approx")
Computes the optimal sample size for a survey to substantiate
freedom from disease. The optimal sample size is the smallest
sample size that produces an alpha-error less than or equal
to a prediscribed value alpha
. The population is considered
as diseased if at least one individual has a positive test result.
The sample size is computed using a bisection method.
The function either computes the sample size for a fixed population
(lookupTable == FALSE
) or a lookup table with sampling sizes
depending on the population size for individual sampling
(lookupTable == TRUE
); see Ziller et al., 2002.
computeOptimalSampleSize(nPopulation, prevalence, alpha = 0.05, sensitivity = 1, specificity = 1, lookupTable = FALSE)
computeOptimalSampleSize(nPopulation, prevalence, alpha = 0.05, sensitivity = 1, specificity = 1, lookupTable = FALSE)
nPopulation |
Integer. Population size if |
prevalence |
Numeric between 0 and 1. Design prvalence. The number of diseased
is then computed as |
alpha |
Numeric between 0 and 1. Alpha-Error (=error of the first kind, significance level) of the underlying significance test. Default value = 0.05. |
sensitivity |
Numeric between 0 and 1. Sensitivity of the diagnostic (for one-stage sampling) or herd test (for two stage sampling). Default value = 1. |
specificity |
Numeric between 0 and 1. Specificity of the diagnostic (for one-stage sampling) or herd test (for two stage sampling). Default value = 1. |
lookupTable |
Logical. TRUE if a lookup table of sample sizes for individual sampling (see, Ziller et al., 2002) should be produced. FALSE if the sample size is desired for a fixed population size (default). |
The return value is either an integer, the minimal sample size that produces
the desired alpha-error if lookupTable == FALSE
or a matrix with
columns N_lower
, N_upper
, sampleSize
containing
the sample sizes for the different herd sizes N
.
E.g., N_lower = 17
, N_upper = 21
, sampleSize = 11
means
that for holdings with 17-21 animals 11 animals need to be tested in order
to achieve the desired accuracy (=herd sensitivity).
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
## Compute the number of herds to be tested for limited sampling ## with sampleSizeLtd = 7: ################################################################# data(sheepData) ## Compute the average herd sensitivity: alphaList <- computeAlphaLimitedSampling(stockSizeVector = sheepData$nSheep, sampleSizeLtd = 7, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, diagSpecificity = 1) sensHerd <- 1 - alphaList$meanAlpha ## Number of herds to be tested: nHerds <- computeOptimalSampleSize(nPopulation = dim(sheepData)[1], prevalence = 0.002, alpha = 0.05, sensitivity = sensHerd, specificity = 1) ## Compute the number of animals to be tested for individual ## sampling: ################################################################# sampleSizeIndividual <- computeOptimalSampleSize(nPopulation = 300, prevalence = 0.2, alpha = 0.05, sensitivity = 0.97, specificity = 1, lookupTable = TRUE)
## Compute the number of herds to be tested for limited sampling ## with sampleSizeLtd = 7: ################################################################# data(sheepData) ## Compute the average herd sensitivity: alphaList <- computeAlphaLimitedSampling(stockSizeVector = sheepData$nSheep, sampleSizeLtd = 7, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, diagSpecificity = 1) sensHerd <- 1 - alphaList$meanAlpha ## Number of herds to be tested: nHerds <- computeOptimalSampleSize(nPopulation = dim(sheepData)[1], prevalence = 0.002, alpha = 0.05, sensitivity = sensHerd, specificity = 1) ## Compute the number of animals to be tested for individual ## sampling: ################################################################# sampleSizeIndividual <- computeOptimalSampleSize(nPopulation = 300, prevalence = 0.2, alpha = 0.05, sensitivity = 0.97, specificity = 1, lookupTable = TRUE)
Computes the optimal sample size (for each risk group) for a survey to substantiate freedom from disease for a population stratified into risk groups. The optimal sample size is the smallest sample size that produces an alpha-error less than or equal to a prediscribed value for alpha. The population is considered as diseased if at least one individual has a positive test result. The sample size is computed using a bisection method. The sample size can be fixed for a subset of the risk groups via the input parameter 'nSampleFixVec' (vector containing sample sizes for the risk groups with fixed values and NA for the risk groups for which the sample size is to be computed). For those risk groups for which the sample size is to be computed a vector specifying the proportional distribution among the risk groups ('nSamplePropVec') needs to be specified.
Example: We have 3 risk groups. For the 2nd risk group we want 20 farms to be sampled. For the other risk groups we specify that the sample size for risk group 1 should be double the sample size of risk group 3. We then set: nSampleFixVec <- c(NA, 20, NA) nSamplePropVec <- c(2,1)
computeOptimalSampleSizeRiskGroups(nPopulationVec, nRelRiskVec, nSampleFixVec = NULL, nSamplePropVec = NULL, prevalence, alpha = 0.05, sensitivity = 1, specificity = 1)
computeOptimalSampleSizeRiskGroups(nPopulationVec, nRelRiskVec, nSampleFixVec = NULL, nSamplePropVec = NULL, prevalence, alpha = 0.05, sensitivity = 1, specificity = 1)
nPopulationVec |
Integer vector. Population sizes of the risk groups. |
nRelRiskVec |
Numeric vector. (Relative) infection risks of the risk groups. |
nSampleFixVec |
Numeric vector containing NAs (optional argument).
For risk groups for which the sample size is fixed
specify the sample size. For the risk groups for which
the sample size should be computed set NA (order of the
risk groups must be the same order as in |
nSamplePropVec |
Numeric vector. For those risk groups for which the
sample size should be computed a proportional
distribution of the overall sample size must be specified.
The vector must have the same length as the number of
NA entries in |
prevalence |
Numeric between 0 and 1. Design prvalence. The number of diseased
is then computed as |
alpha |
Numeric between 0 and 1. Alpha-Error (=error of the first kind, significance level) of the underlying significance test. Default value = 0.05. |
sensitivity |
Numeric between 0 and 1. Sensitivity of the diagnostic (for one-stage sampling) or herd test (for two stage sampling). Default value = 1. |
specificity |
Numeric between 0 and 1. Specificity of the diagnostic (for one-stage sampling) or herd test (for two stage sampling). Default value = 1. |
The return value is an integer vector containing the optimal sample size
for every risk group specified in the input variables nPopulationVec
and nRelRiskVec
.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
P.A.J.Martin, A.R. Cameron, M. Greiner, "Demonstrating freedom from disease using multiple complex data sources. : A new methodology based on scenario trees", Prev. Vet. Med. 79 (2007), pp. 71 - 97.
nPopulationVec <- c(500,700) nRelRiskVec <- c(1,3) prevalence <- 0.01 alpha <- 0.05 herdSensitivity <- 0.7 specificity <- 1 ## Optimal sample size with risk groups: nRisk <- computeOptimalSampleSizeRiskGroups(nPopulationVec = nPopulationVec, nRelRiskVec = nRelRiskVec, nSamplePropVec = c(1,4), prevalence = prevalence, alpha = alpha, sensitivity = herdSensitivity, specificity = specificity) ## Optimal sample size without risk groups: nNoRisk <- computeOptimalSampleSize(sum(nPopulationVec), prevalence, alpha, herdSensitivity, specificity, FALSE)
nPopulationVec <- c(500,700) nRelRiskVec <- c(1,3) prevalence <- 0.01 alpha <- 0.05 herdSensitivity <- 0.7 specificity <- 1 ## Optimal sample size with risk groups: nRisk <- computeOptimalSampleSizeRiskGroups(nPopulationVec = nPopulationVec, nRelRiskVec = nRelRiskVec, nSamplePropVec = c(1,4), prevalence = prevalence, alpha = alpha, sensitivity = herdSensitivity, specificity = specificity) ## Optimal sample size without risk groups: nNoRisk <- computeOptimalSampleSize(sum(nPopulationVec), prevalence, alpha, herdSensitivity, specificity, FALSE)
For a population of size nPopulation
with a given design prevalence
the function computes the probability of finding no testpositives
in a sample of size nSample
if an imperfect test is used (given sensitivity
and specificity). This probability corresponds to the alpha-error
(=error of the first kind) of the overall test with null hypothesis:
prevalence = design prevalence. A modified hypergeometric formula
is used; see Cameron, Baldock, 1998.
computePValue(nPopulation, nSample, nDiseased, sensitivity, specificity = 1)
computePValue(nPopulation, nSample, nDiseased, sensitivity, specificity = 1)
nPopulation |
Integer. Population size. |
nSample |
Integer. Size of sample. |
nDiseased |
Integer. Number of diseased elements in the population according to the design prevalence. |
sensitivity |
Numeric between 0 and 1. Sensitivity (= probability of a testpositive result, given the tested individual is diseased) of the test (e.g., diagnostic test or herd test). |
specificity |
Numeric between 0 and 1. Specificity (= probability of a testnegative result, given the tested individual is not diseased) of the test (e.g., diagnostic test or herd test). The default value is 1. |
The return value is a numeric between 0 and 1. It is the probability of finding no testpositives (not diseased!) in the sample.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
computeOptimalSampleSize
, computeAlphaLimitedSampling
alphaError <- computePValue(nPopulation = 3000, nSample = 1387, nDiseased = 6, sensitivity = 0.85, specificity = 1)
alphaError <- computePValue(nPopulation = 3000, nSample = 1387, nDiseased = 6, sensitivity = 0.85, specificity = 1)
For a population that is stratified into risk groups the function computes the
probability of finding no testpositives in a sample of given size using an
imperfect diagnostic test. For each of the risk groups the population size
nPopulationVec
, the sample size nSampleVec
and the relative
infection risk nRelRiskVec
must be specified. The discussed probability
corresponds to the alpha-error (=error of the first kind) of the overall test
with null hypothesis: prevalence = design prevalence.
computePValueRiskGroups(nPopulationVec, nSampleVec, nRelRiskVec, nDiseased, sensitivity, specificity = 1)
computePValueRiskGroups(nPopulationVec, nSampleVec, nRelRiskVec, nDiseased, sensitivity, specificity = 1)
nPopulationVec |
Integer vector. Population sizes of the risk groups. |
nSampleVec |
Integer vector. Sample sizes of the risk groups. |
nRelRiskVec |
Numeric vector. (Relative) infection risks of the risk groups. |
nDiseased |
Integer. Number of diseased elements in the population according to the design prevalence. |
sensitivity |
Numeric between 0 and 1. Sensitivity (= probability of a testpositive result, given the tested individual is diseased) of the test (e.g., diagnostic test or herd test). |
specificity |
Numeric between 0 and 1. Specificity (= probability of a testnegative result, given the tested individual is not diseased) of the test (e.g., diagnostic test or herd test). The default value is 1. |
The return value is a numeric between 0 and 1. It is the probability of finding no testpositives (not diseased!) in the sample.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
P.A.J.Martin, A.R. Cameron, M. Greiner, "Demonstrating freedom from disease using multiple complex data sources. : A new methodology based on scenario trees", Prev. Vet. Med. 79 (2007), pp. 71 - 97.
Calls computePValue
nPopulationVec <- c(500,700) nSampleVec <- c(300,200) nRelRiskVec <- c(1,1) nDiseased <- round(sum(nPopulationVec)*0.01) sensitivity <- 0.9 specificity <- 1 alphaError <- computePValue(sum(nPopulationVec), sum(nSampleVec), nDiseased, sensitivity, specificity)
nPopulationVec <- c(500,700) nSampleVec <- c(300,200) nRelRiskVec <- c(1,1) nDiseased <- round(sum(nPopulationVec)*0.01) sensitivity <- 0.9 specificity <- 1 alphaError <- computePValue(sum(nPopulationVec), sum(nSampleVec), nDiseased, sensitivity, specificity)
This function opens the Graphical User Interface for the FFD sampling plan calculator.
The GUI offers the functionality to
compute first and second-stage sample sizes for two-stage sampling schemes to substantiate freedom from disease,
analyze two-stage sampling schemes with respect to overall cost and sensitivity,
determine cost-optimal strategies and
draw samples from a population.
The GUI is structured into three tabs, Data Input, Parameters, Calculations:
Data Input: The farm data is specified. A list of farms is required with one row per farm and one column
containing the herd size, i.e. the number of animals on the farm. The data must be provided in the CSV file format,
however currently only the central European format with comma = ",", seperator = ";" is supported. The location of
the csv-file is specified in the field Data file, the name of the column containing the herd size is set in
the field Herd sizes column via a dropdown menu.
Parameters: Next the survey parameters: design prevalence, Type I error level, intraherd prevalence,
and test sensitivity must be set and a sampling strategy (limited sampling or individual sampling, see, e.g. Ziller et al.)
must be chosen. Furthermore the cost of each tested herd (excluding the cost per tested animal) and the cost
of each tested animal can be specified for cost analyses.
Calculations: Once a sampling strategy is chosen the tab is activated, where cost optimal sampling
plans can be determined (Sample size Diagnostics), cost and sensitivity of a fixed sampling schemes can
be computed (Compute sample size -> Calculate) and an actual sample can be drawn from the farm file
(Compute sample size -> Sample).
FFD_GUI()
FFD_GUI()
The FFD GUI provides the following menu options:
File
Load loads population data and survey parameter settings from a ffd-RData file.
Save saves population data and survey parameter settings to a ffd-RData file.
Load examples loads built-in examples.
Reset clears all entered data.
Help
R Documentation opens the help file for the FFD GUI (cf ?FFD_GUI
).
FFD Manual (PDF) opens the pdf manual
About FFD_GUI displays version and developers info.
Ian Kopacka
Cameron A.R. and Baldock F.C. (1998). A new probability formula for surveys to substantiate freedom from disease, Prev. Vet. Med. 34 (1), pp. 1-17
Cameron A.R. and Baldock F.C. (1998). Two-stage sampling in surveys to substantiate freedom from disease, Prev. Vet. Med. 34 (1), pp. 19-30
Ziller M., Selhorst T. Teuffert J., Kramer M. and Schlueter H. (2002). Analysis of sampling strategies to substantiate freedom from disease in large areas, Prev. Vet. Med. 52 (3-4), pp. 333-343
Creates an object of the class 'IndSampling'. For given survey parameters
(passed to the function as an object of the class SurveyData
)
indSampling()
computes the the number of herds to test,
the expected total number of animals to test, the expected total cost of a survey
using limited sampling with a given herd sensitivity herdSensitivity
, as
well as a lookup table for the number of animals to test per herd, depending on the
herd size.
indSampling(survey.Data, herdSensitivity, nSampleFixVec = NULL, probVec = NULL)
indSampling(survey.Data, herdSensitivity, nSampleFixVec = NULL, probVec = NULL)
survey.Data |
Object of class |
herdSensitivity |
Numeric between 0 and 1. Desired herd sensitivity. |
nSampleFixVec |
Numeric vector containing some NAs (optional argument).
For risk groups for which the sample size is fixed
specify the sample size. For the risk groups for which
the sample size should be computed set NA (order of the
risk groups must be the same order as in |
probVec |
Numeric vector. For those risk groups for which the
sample size should be computed sample probabilities must
be specified.
The vector must have the same length as the number of
NA entries in |
The function returns an object of the class IndSampling
.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
A.R. Cameron and F.C. Baldock, "Two-stage sampling surveys to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 19-30.
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
See IndSampling
and SurveyData
for additional details.
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Individual sampling without risk groups: myIndSampling <- indSampling(survey.Data = mySurvey, herdSensitivity = 0.7) ## Individual sampling with risk groups: myIndSamplingRG <- indSampling(survey.Data = mySurvey, herdSensitivity = 0.7, nSampleFixVec = NULL, probVec = c(1,4))
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Individual sampling without risk groups: myIndSampling <- indSampling(survey.Data = mySurvey, herdSensitivity = 0.7) ## Individual sampling with risk groups: myIndSamplingRG <- indSampling(survey.Data = mySurvey, herdSensitivity = 0.7, nSampleFixVec = NULL, probVec = c(1,4))
Contains the parameters and the data necessary for a survey to substantiate freedom from disease using "individual sampling". Additionally to the survey parameters: design prevalence (=prevalence of the disease under the null hypothesis), overall significance level (=1-confidence), intra-herd prevalence, sensitivity of the diagnostic test, cost per tested animal and cost per tested herd, the object contains the herd sensitivity the number of herds to be tested, the mean overall number of animals to be tested, the expected costs, as well as a lookup table for the number of animals to test depending on the herd size.
Objects can be created by calls of the form new("IndSampling", ...)
.
surveyData
:Object of class "SurveyData"
. Contains all the necessary data
and specifications for the survey.
herdSensitivity
:Object of class "numeric"
with values between
0 and 1. Desired herd sensitivity.
nHerds
:Object of class "numeric"
. Number of herds
to be tested according to the herd sensitivity herdSensitivity
.
nHerdsPerRiskGroup
:Object of class "numeric"
. Number of herds
to be tested per risk group (if population is stratified by risk groups).
nSampleFixVec
:Object of class "numeric"
. Numeric vector
containing some NAs (optional argument). For risk groups for which the
sample size is fixed it specifies the sample size. For the risk groups
for which the sample size was computed it was set to NA (order of the
risk groups is the same as in survey.Data@riskValueData
).
probVec
:Object of class "numeric"
. Contains the
sample probabilities for those risk groups for which the sample size
was computed (=NA entries in nSampleFixVec
).
nAnimalsMean
:Object of class "numeric"
. Expected
total number of animals to be tested in the survey.
expectedCost
:Object of class "numeric"
. Expected
costs of the survey.
lookupTable
:Object of class "matrix"
with columns N_lower
,
N_upper
and sampleSize
containing the number of animals to
test for each herd size.
signature(x = "IndSampling")
: Creates an html file containing the
summary data and the diagnostic plots. Title, file name, output directory, css-file,
etc. can additionally be specified using the parameters, filename
, outdir
,
CSSFile
, Title
, as well as all the other parameters of
the R2HTML-function HTMLInitFile
.
signature(x = "IndSampling", size = c("fixed", "dynamic"))
: Sample herds
using individual sampling. Additionally
to the argument x
of type IndSampling
the method takes an argument size
,
which is a character string. For size == "fixed"
the fixed number of herds
given in x@nHerds
is sampled using simple random sampling. For size == "dynamic"
dynamic sampling is used, i.e., based on real-time computation of the a-posteriori alpha-
error the sample is updated until the a-posteriori alpha-error falls below the predefined
significance level x@alpha. The return value is a list with two items: indexSample
is
a vector of indices of the sample corresponding to x@surveyData@nAnimalVec
and
aPostAlpha
containing the a-posteriori alpha-error of the sample.
signature(object = "IndSampling")
: Display structure of the class and content
of the slots.
signature(object = "IndSampling")
: Display structure of the class and a
summary of the content of the slots.
No notes yet.
Ian Kopacka <[email protected]>
The slot surveyData
contains an object of the class
SurveyData
which is created using surveyData
.
Objects of the class IndSampling
are create using the constructor
indSampling
.
## Show the structure of the class: showClass("IndSampling") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) myIndSampling <- indSampling(survey.Data = mySurvey, herdSensitivity = 0.7) ## Display results: summary(myIndSampling) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "IndSampling") HTML(myIndSampling) HTMLEndFile() ## End(Not run)
## Show the structure of the class: showClass("IndSampling") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) myIndSampling <- indSampling(survey.Data = mySurvey, herdSensitivity = 0.7) ## Display results: summary(myIndSampling) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "IndSampling") HTML(myIndSampling) HTMLEndFile() ## End(Not run)
Creates an object of the class 'IndSamplingSummary'. For given survey parameters
(passed to the function as an object of the class SurveyData
)
indSamplingSummary()
computes the number of herds to test,
the expected total number of animals to test and the expected total cost of a survey
using individual sampling with a given sequence of herd sensitivities. This sequence
ranges from 0.1 to the sensitivity of the diagnostic test specified in survey.Data
.
The step size for the herd sensitivities can be specified by the user via the argument
stepSize
. If no step size is specified a step size of 0.02 is used.
indSamplingSummary(survey.Data, stepSize = 0.02, nSampleFixVec = NULL, probVec = NULL)
indSamplingSummary(survey.Data, stepSize = 0.02, nSampleFixVec = NULL, probVec = NULL)
survey.Data |
Object of class |
stepSize |
Numeric. A series of parameters is computed for a
sequence of herd sensitivities. The argument |
nSampleFixVec |
Numeric vector containing some NAs (optional argument).
For risk groups for which the sample size is fixed
specify the sample size. For the risk groups for which
the sample size should be computed set NA (order of the
risk groups must be the same order as in |
probVec |
Numeric vector. For those risk groups for which the
sample size should be computed sample probabilities must
be specified.
The vector must have the same length as the number of
NA entries in |
The function returns an object of the class IndSamplingSummary
.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
A.R. Cameron and F.C. Baldock, "Two-stage sampling surveys to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 19-30.
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
See IndSamplingSummary
and SurveyData
for additional details.
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Individual sampling without risk groups: myIndSamplingSummary <- indSamplingSummary(survey.Data = mySurvey, stepSize = 0.06) ## Individual sampling with risk groups: myIndSamplingSummaryRG <- indSamplingSummary(survey.Data = mySurvey, stepSize = 0.06, nSampleFixVec = NULL, probVec = c(1,4))
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Individual sampling without risk groups: myIndSamplingSummary <- indSamplingSummary(survey.Data = mySurvey, stepSize = 0.06) ## Individual sampling with risk groups: myIndSamplingSummaryRG <- indSamplingSummary(survey.Data = mySurvey, stepSize = 0.06, nSampleFixVec = NULL, probVec = c(1,4))
Contains the parameters and the data necessery for a survey to substantiate freedom from disease using "individual sampling". Additionally to the survey parameters: design prevalence (=prevalence of the disease under the null hypothesis), overall significance level (=1-confidence), intra-herd prevalence, sensitivity of the diagnostic test, cost per tested animal and cost per tested herd, the object contains the the number of herds to be tested, the mean overall number of animals to be tested and the expected costs for a range of possible herd sensitivities.
Objects can be created by calls of the form new("IndSamplingSummary", ...)
.
surveyData
:Object of class "SurveyData"
. Containing the
necessary survey parameters.
herdSensVec
:Object of class "numeric"
with
values between 0 and 1. Mean herd sensitivity in the population (vector).
nHerdsVec
:Object of class "numeric"
. Number of herds
to be tested according to the herd sensitivity herdSensVec
(vector).
nHerdsPerRiskGroupMx
:Object of class "matrix"
. Number of herds
to be tested per risk group [columns] and sample limit [rows]
(if population is stratified by risk groups).
nSampleFixVec
:Object of class "numeric"
. Numeric vector
containing some NAs (optional argument). For risk groups for which the
sample size is fixed it specifies the sample size. For the risk groups
for which the sample size was computed it was set to NA (order of the
risk groups is the same as in survey.Data@riskValueData
).
probVec
:Object of class "numeric"
. Contains the
sample probabilities for those risk groups for which the sample size
was computed (=NA entries in nSampleFixVec
).
nAnimalsMeanVec
:Object of class "numeric"
. Expected
total number of animals to be tested in the survey (vector).
expectedCostVec
:Object of class "numeric"
. Expected
costs of the survey (vector).
signature(x = "IndSamplingSummary")
: Creates an html file containing the
summary data and the diagnostic plots. Title, file name, output directory, css-file,
etc. can additionally be specified using the parameters, filename
, outdir
,
CSSFile
, Title
, as well as all the other parameters of
the R2HTML-function HTMLInitFile
.
signature(x = "IndSamplingSummary")
: Create plots of 1) the mean number of animals
to be tested per herd, 2) the number of herds to be tested, 3) the expected total number of
animals to be tested, 4) the expected total cost of the survey plotted against the vector
of herd sensitivities.
signature(object = "IndSamplingSummary")
: Display structure of the class and content
of the slots.
signature(object = "IndSamplingSummary")
: Display structure of the class and a
summary of the content of the slots.
No notes yet.
Ian Kopacka <[email protected]>
The slot surveyData
contains an object of the class
SurveyData
.
## Show the structure of the class: showClass("IndSamplingSummary") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 15, costAnimal = 20) myIndSamplingSummary <- indSamplingSummary(survey.Data = mySurvey, stepSize = 0.06) ## Display results: summary(myIndSamplingSummary) plot(myIndSamplingSummary) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "ItdSampling") HTML(myIndSamplingSummary) HTMLEndFile() ## End(Not run)
## Show the structure of the class: showClass("IndSamplingSummary") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 15, costAnimal = 20) myIndSamplingSummary <- indSamplingSummary(survey.Data = mySurvey, stepSize = 0.06) ## Display results: summary(myIndSamplingSummary) plot(myIndSamplingSummary) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "ItdSampling") HTML(myIndSamplingSummary) HTMLEndFile() ## End(Not run)
Function works like ls() but it returns the names of the objects in the workspace, together with class, dimension and size in the form of a data frame.
lls(name, pos = -1, envir = as.environment(pos), all.names = FALSE, pattern, classFilter, sort = "size")
lls(name, pos = -1, envir = as.environment(pos), all.names = FALSE, pattern, classFilter, sort = "size")
name |
which environment to use in listing the available objects. Defaults to the current environment. Although called name for back compatibility, in fact this argument can specify the environment in any form; see the details section. |
pos |
an alternative argument to name for specifying the environment as a position in the search list. Mostly there for back compatibility. |
envir |
an alternative argument to name for specifying the environment. Mostly there for back compatibility. |
all.names |
a logical value. If TRUE, all object names are returned. If FALSE, names which begin with a . are omitted. |
pattern |
an optional regular expression. Only names matching pattern are returned. glob2rx can be used to convert wildcard patterns to regular expressions. |
classFilter |
optional character string. Only objects of the specified class are returned. |
sort |
optional character string either "size" (default) or "name". Objects are either sorted by size or alphabetically. |
The name argument can specify the environment from which object names are taken in one of several forms: as an integer (the position in the search list); as the character string name of an element in the search list; or as an explicit environment (including using sys.frame to access the currently active function calls). By default, the environment of the call to lls or objects is used. The pos and envir arguments are an alternative way to specify an environment, but are primarily there for back compatibility.
Returns a data frame.
Original author unknown, modified by Ian Kopacka
lls()
lls()
Creates an object of the class 'LtdSampling'. For given survey parameters
(passed to the function as an object of the class SurveyData
)
ltdSampling()
computes the mean herd sensitivity, the number of herds to test,
the expected total number of animals to test and the expected total cost of a survey
using limited sampling with a given animal-level sample size sampleSizeLtd
.
If values for nSampleFixVec
and/or probVec
are specified, sampling
if performed with stratification of the population by risk groups.
ltdSampling(survey.Data, sampleSizeLtd, nSampleFixVec = NULL, probVec = NULL)
ltdSampling(survey.Data, sampleSizeLtd, nSampleFixVec = NULL, probVec = NULL)
survey.Data |
Object of class |
sampleSizeLtd |
Positive integer. Pre-fixed number of animals to be tested per holding, irrespective of the herd size (if the herd contains fewer animals then the entire herd needs to be tested). |
nSampleFixVec |
Numeric vector containing some NAs (optional argument).
For risk groups for which the sample size is fixed
specify the sample size. For the risk groups for which
the sample size should be computed set NA (order of the
risk groups must be the same order as in |
probVec |
Numeric vector. For those risk groups for which the
sample size should be computed sample probabilities must
be specified.
The vector must have the same length as the number of
NA entries in |
The function returns an object of the class LtdSampling
.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
A.R. Cameron and F.C. Baldock, "Two-stage sampling surveys to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 19-30.
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
See LtdSampling
and SurveyData
for additional details.
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Limited sampling without risk groups: myLtdSampling <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 7) ## Limited sampling with risk groups: myLtdSamplingRG <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 7, nSampleFixVec = NULL, probVec = c(1,4))
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Limited sampling without risk groups: myLtdSampling <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 7) ## Limited sampling with risk groups: myLtdSamplingRG <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 7, nSampleFixVec = NULL, probVec = c(1,4))
Contains the parameters and the data necessary for a survey to substantiate freedom from disease using "limited sampling". Additionally to the survey parameters: design prevalence (=prevalence of the disease under the null hypothesis), overall significance level (=1-confidence), intra-herd prevalence, sensitivity of the diagnostic test, cost per tested animal and cost per tested herd, the object contains the mean herd sensitivity, the number of herds to be tested, the mean overall number of animals to be tested and the expected costs.
Objects can be created by calls of the form new("LtdSampling", ...)
.
surveyData
:Object of class "SurveyData"
. Contains all the necessary data
and specifications for the survey.
sampleSizeLtd
:Object of class "numeric"
. Pre-fixed number of
animals to be tested per holding, irrespective of the herd size.
If a herd contains fewer animals the entire herd is tested.
meanHerdSensitivity
:Object of class "numeric"
with
values between 0 and 1. Mean herd sensitivity in the population.
meanHerdSensPerRG
:Object of class "numeric"
with values between 0 and 1. Mean herd sensitivity of each risk group
(if population is stratified by risk groups).
nHerds
:Object of class "numeric"
. Number of herds
to be tested according to the herd sensitivity meanHerdSensitivity
.
nHerdsPerRiskGroup
:Object of class "numeric"
. Number of herds
to be tested per risk group (if population is stratified by risk groups).
nSampleFixVec
:Object of class "numeric"
. Numeric vector
containing some NAs (optional argument). For risk groups for which the
sample size is fixed it specifies the sample size. For the risk groups
for which the sample size was computed it was set to NA (order of the
risk groups is the same as in survey.Data@riskValueData
).
probVec
:Object of class "numeric"
. Contains the
sample probabilities for those risk groups for which the sample size
was computed (=NA entries in nSampleFixVec
).
nAnimalsMean
:Object of class "numeric"
. Expected
total number of animals to be tested in the survey.
expectedCost
:Object of class "numeric"
. Expected
costs of the survey.
signature(x = "LtdSampling")
: Creates an html file containing the
summary data and the diagnostic plots. Title, file name, output directory, css-file,
etc. can additionally be specified using the parameters, filename
, outdir
,
CSSFile
, Title
, as well as all the other parameters of
the R2HTML-function HTMLInitFile
.
signature(x = "LtdSampling", size = c("fixed", "dynamic"))
:
Sample herds using limited sampling. Additionally
to the argument x
of type LtdSampling
the method takes an argument size
,
which is a character string. For size == "fixed"
the fixed number of herds
given in x@nHerds
is sampled using simple random sampling. For size == "dynamic"
dynamic sampling is used, i.e., based on real-time computation of the a-posteriori alpha-
error the sample is updated until the a-posteriori alpha-error falls below the predefined
significance level x@alpha. The return value is a list with two items: indexSample
is
a vector of indices of the sample corresponding to x@surveyData@nAnimalVec
and
aPostAlpha
containing the a-posteriori alpha-error of the sample.
signature(object = "LtdSampling")
: Display structure of the class and content
of the slots.
signature(object = "LtdSampling")
: Display structure of the class and a
summary of the content of the slots.
No notes yet.
Ian Kopacka <[email protected]>
The slot surveyData
contains an object of the class
SurveyData
which is created using surveyData
.
Objects of the class LtdSampling
are create using the constructor
ltdSampling
.
## Show the structure of the class: showClass("LtdSampling") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) myLtdSampling <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 7) ## Display results: summary(myLtdSampling) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "LtdSampling") HTML(myLtdSampling) HTMLEndFile() ## End(Not run)
## Show the structure of the class: showClass("LtdSampling") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) myLtdSampling <- ltdSampling(survey.Data = mySurvey, sampleSizeLtd = 7) ## Display results: summary(myLtdSampling) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "LtdSampling") HTML(myLtdSampling) HTMLEndFile() ## End(Not run)
Creates an object of the class 'LtdSamplingSummary'. For given survey parameters
(passed to the function as an object of the class SurveyData
)
ltdSamplingSummary()
computes the mean herd sensitivity, the number of herds to test,
the expected total number of animals to test and the expected total cost of a survey
using limited sampling with a given sequence of animal-level sample sizes. This sequence
ranges from 1 to a given upper bound sampleSizeLtdMax
. If no upper bound is
specified the maximal herd size is used.
ltdSamplingSummary(survey.Data, sampleSizeLtdMax, nSampleFixVec = NULL, probVec = NULL)
ltdSamplingSummary(survey.Data, sampleSizeLtdMax, nSampleFixVec = NULL, probVec = NULL)
survey.Data |
Object of class |
sampleSizeLtdMax |
Positive integer. A series of parameters is computed
for a sequence of sample limits. These sample limits
range from 1 to a given upper bound, defined by
|
nSampleFixVec |
Numeric vector containing some NAs (optional argument).
For risk groups for which the sample size is fixed
specify the sample size. For the risk groups for which
the sample size should be computed set NA (order of the
risk groups must be the same order as in |
probVec |
Numeric vector. For those risk groups for which the
sample size should be computed sample probabilities must
be specified.
The vector must have the same length as the number of
NA entries in |
The function returns an object of the class LtdSamplingSummary
.
Ian Kopacka <[email protected]>
A.R. Cameron and F.C. Baldock, "A new probablility formula to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 1-17.
A.R. Cameron and F.C. Baldock, "Two-stage sampling surveys to substantiate freedom from disease", Prev. Vet. Med. 34 (1998), pp. 19-30.
M. Ziller, T. Selhorst, J. Teuffert, M. Kramer and H. Schlueter, "Analysis of sampling strategies to substantiate freedom from disease in large areas", Prev. Vet. Med. 52 (2002), pp. 333-343.
See LtdSamplingSummary
and SurveyData
for additional details.
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Limited sampling without risk groups: myLtdSamplingSummary <- ltdSamplingSummary(survey.Data = mySurvey, sampleSizeLtdMax = 10) ## Limited sampling with risk groups: myLtdSamplingRG <- ltdSamplingSummary(survey.Data = mySurvey, sampleSizeLtdMax = 10, nSampleFixVec = NULL, probVec = c(1,4))
data(sheepData) sheepData$size <- ifelse(sheepData$nSheep < 30, "small", "large") riskValueData <- data.frame(riskGroup = c("small", "large"), riskValues = c(1,2)) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, riskGroupVec = sheepData$size, riskValueData = riskValueData, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Limited sampling without risk groups: myLtdSamplingSummary <- ltdSamplingSummary(survey.Data = mySurvey, sampleSizeLtdMax = 10) ## Limited sampling with risk groups: myLtdSamplingRG <- ltdSamplingSummary(survey.Data = mySurvey, sampleSizeLtdMax = 10, nSampleFixVec = NULL, probVec = c(1,4))
Contains the parameters and the data necessery for a survey to substantiate freedom from disease using "limited sampling". Additionally to the survey parameters: design prevalence (=prevalence of the disease under the null hypothesis), overall significance level (=1-confidence), intra-herd prevalence, sensitivity of the diagnostic test, cost per tested animal and cost per tested herd, the object contains the mean herd sensitivity, the number of herds to be tested, the mean overall number of animals to be tested and the expected costs for a range of possible sample limits (= fixed number of animals to test per herd).
Objects can be created by calls of the form new("LtdSamplingSummary", ...)
.
surveyData
:Object of class "SurveyData"
. Containing the
necessary survey parameters.
sampleSizeLtdVec
:Object of class "numeric"
. Pre-fixed number of
animals to be tested per holding, irrespective of the herd size.
If a herd contains fewer animals the entire herd is tested (vector).
meanHerdSensVec
:Object of class "numeric"
with
values between 0 and 1. Mean herd sensitivity in the population (vector).
meanHerdSensPerRGMx
:Object of class "matrix"
with values between 0 and 1. Mean herd sensitivity of each risk group
[columns] and sample limit [rows] (if population is stratified by
risk groups).
nHerdsVec
:Object of class "numeric"
. Number of herds
to be tested according to the herd sensitivity meanHerdSensitivity
(vector).
nHerdsPerRiskGroupMx
:Object of class "matrix"
. Number of herds
to be tested per risk group [columns] and sample limit [rows]
(if population is stratified by risk groups).
nSampleFixVec
:Object of class "numeric"
. Numeric vector
containing some NAs (optional argument). For risk groups for which the
sample size is fixed it specifies the sample size. For the risk groups
for which the sample size was computed it was set to NA (order of the
risk groups is the same as in survey.Data@riskValueData
).
probVec
:Object of class "numeric"
. Contains the
sample probabilities for those risk groups for which the sample size
was computed (=NA entries in nSampleFixVec
).
nAnimalsMeanVec
:Object of class "numeric"
. Expected
total number of animals to be tested in the survey (vector).
expectedCostVec
:Object of class "numeric"
. Expected
costs of the survey (vector).
signature(x = "LtdSamplingSummary")
: Creates an html file containing the
summary data and the diagnostic plots. Title, file name, output directory, css-file,
etc. can additionally be specified using the parameters, filename
, outdir
,
CSSFile
, Title
, as well as all the other parameters of
the R2HTML-function HTMLInitFile
.
signature(x = "LtdSamplingSummary")
: Create plots of 1) the mean herd sensitivity,
2) the number of herds to be tested, 3) the expected total number of animals to be tested,
4) the expected total cost of the survey plotted against the vector of sample limits.
signature(object = "LtdSamplingSummary")
: Display structure of the class and content
of the slots.
signature(object = "LtdSamplingSummary")
: Display structure of the class and a
summary of the content of the slots.
No notes yet.
Ian Kopacka <[email protected]>
The slot surveyData
contains an object of the class
SurveyData
.
## Show the structure of the class: showClass("LtdSamplingSummary") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) myLtdSamplingSummary <- ltdSamplingSummary(survey.Data = mySurvey, sampleSizeLtdMax = 7) ## Display results: summary(myLtdSamplingSummary) plot(myLtdSamplingSummary) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "LtdSampling") HTML(myLtdSamplingSummary) HTMLEndFile() ## End(Not run)
## Show the structure of the class: showClass("LtdSamplingSummary") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) myLtdSamplingSummary <- ltdSamplingSummary(survey.Data = mySurvey, sampleSizeLtdMax = 7) ## Display results: summary(myLtdSamplingSummary) plot(myLtdSamplingSummary) ## Write results to an html-file: ## Not run: target <- HTMLInitFile(getwd(), filename = "LtdSampling") HTML(myLtdSamplingSummary) HTMLEndFile() ## End(Not run)
This test data frame contains a ficticious list of Austrian sheep holdings with information on the state and the number of animals.
data(sheepData)
data(sheepData)
A data frame with 15287 observations on the following 3 variables.
herdId
Numeric vector. Unique identification number of the herd.
state
Numeric vector. The 9 Austrian states are coded using the numbers 1 to 9.
nSheep
Numeric vector. Number of animals for each holding.
Simulated test data.
data(sheepData)
data(sheepData)
Constructor for objects of the class 'SurveyData'.
surveyData(nAnimalVec = numeric(), riskGroupVec = character(), riskValueData = data.frame(), populationData = data.frame(), designPrevalence = numeric(), alpha = numeric(), intraHerdPrevalence = numeric(), diagSensitivity = numeric(), costHerd = numeric(), costAnimal = numeric())
surveyData(nAnimalVec = numeric(), riskGroupVec = character(), riskValueData = data.frame(), populationData = data.frame(), designPrevalence = numeric(), alpha = numeric(), intraHerdPrevalence = numeric(), diagSensitivity = numeric(), costHerd = numeric(), costAnimal = numeric())
nAnimalVec |
Positive numeric vector containing the number of animals per holding. |
riskGroupVec |
Character vector. Vector containing the the name of a risk group to which the farm belongs. Optional argument. If provided, it must have the same length as nAnimalVec. |
riskValueData |
Data frame. Data frame where the first column contains the labels in riskGroupVec and the second column contains the numeric values for the relative infection risk. |
populationData |
Data frame. Columns of the data frame
must have the same length as the vector in |
designPrevalence |
Numeric with values between 0 and 1. Prevalence of the disease under the null hypothesis. |
alpha |
Numeric with values between 0 and 1. Type one error for the statistical test (=significance level). |
intraHerdPrevalence |
Numeric with values between 0 and 1. Intra-herd prevalence, i.e., the assumed prevalence of the disease within an infected herd. |
diagSensitivity |
Numeric with values between 0 and 1. Sensitivity of the diagnostic test used. |
costHerd |
Numeric. Cost per tested herd excluding costs for sampling of animals (e.g., travel costs of the vet). |
costAnimal |
Numeric. Cost per tested animal, e.g., drawing of samples + analysis in the lab. |
The function returns an object of the class SurveyData
.
Ian Kopacka <[email protected]>
See SurveyData
for additional details on the class.
data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1)
data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1)
Contains the parameters and the data necessery for a survey to substantiate freedom from disease using "individual sampling" or "limited sampling". The parameters are: design prevalence (=prevalence of the disease under the null hypothesis), overall significance level (=1-confidence), intra-herd prevalence, sensitivity of the diagnostic test, as well as cost per tested animal and cost per tested herd.
Objects can be created by calls of the form new("SurveyData", ...)
.
nAnimalVec
:Object of class "numeric"
(nonnegative). Vector containing the
number of animals per holding.
riskGroupVec
:Object of class "character"
. Vector containing the
the name of a risk group to which the farm belongs. Optional argument.
If provided, it must have the same length as nAnimalVec
.
riskValueData
:Object of class "data.frame"
. Data frame
where the first column contains the labels in riskGroupVec
and
the second column contains the numeric values for the relative infection risk.
populationData
:Object of class "data.frame"
. Columns of the data frame
must have the same length as the vector in slot nAnimalVec
. The data frame can
contain additional data such as herd id, name and address of the owner etc.
designPrevalence
:Object of class "numeric"
with values between 0 and 1.
Prevalence of the disease under the null hypothesis.
alpha
:Object of class "numeric"
with values between 0 and 1.
Type one error for the statistical test (significance level).
intraHerdPrevalence
:Object of class "numeric"
with values between 0 and 1.
Intra-herd prevalence, i.e., the assumed prevalence of the disease within an infected herd.
diagSensitivity
:Object of class "numeric"
with values between 0 and 1.
Sensitivity of the diagnostic test used.
costHerd
:Object of class "numeric"
. Cost per tested herd excluding costs for
sampling of animals (e.g., travel costs of the vet.)
costAnimal
:Object of class "numeric"
. Cost per tested animal, e.g., drawing
of samples + analysis in the lab.
signature(object = "SurveyData")
: Display structure of the class and content
of the slots.
signature(object = "SurveyData")
: Display structure of the class and a
summary of the content of the slots.
No notes added yet.
Ian Kopacka <[email protected]>
Objects of the class 'SurveyData' can be created using the constructor
surveyData
and are used as types for slots in the class
LtdSampling
.
## Show the structure of the class: showClass("SurveyData") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Display results: summary(mySurvey)
## Show the structure of the class: showClass("SurveyData") ## Create an object: data(sheepData) mySurvey <- surveyData(nAnimalVec = sheepData$nSheep, populationData = sheepData, designPrevalence = 0.002, alpha = 0.05, intraHerdPrevalence = 0.13, diagSensitivity = 0.9, costHerd = 30, costAnimal = 7.1) ## Display results: summary(mySurvey)