Package {SampleSizeR}

Convert SampleSizeR Object to Data Frame

Description

Convert SampleSizeR Object to Data Frame

Usage

## S3 method for class 'SampleSizeR'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

Arguments

Value

Data frame.

Plot Sample Size Result

Description

Plot Sample Size Result

Usage

## S3 method for class 'SampleSizeR'
plot(x, ...)

Arguments

Value

ggplot object.

Print Sample Size Result

Description

Prints a SampleSizeR object in a user-friendly format.

Usage

## S3 method for class 'SampleSizeR'
print(x, digits = 2, ...)

Arguments

Value

Invisible x.

Sample Size for an Unmatched Case-Control Study

Description

Calculates the required sample size for an unmatched case-control study based on the expected odds ratio, exposure prevalence among controls, desired statistical power and significance level.

Usage

ss_case_control(
  odds.ratio,
  p0,
  alpha = 0.05,
  power = 0.8,
  ratio = 1,
  dropout = 0
)

Arguments

Details

The implementation follows the Kelsey/Fleiss approach and supports unequal control-to-case allocation ratios.

Exposure prevalence among cases is estimated from

p_1=\frac{OR\times p_0} {1+p_0(OR-1)}

Sample size is then calculated using the Kelsey/Fleiss unmatched case-control formula.

Value

Object of class SampleSizeR

References

Kelsey JL, Whittemore AS, Evans AS, Thompson WD. Methods in Observational Epidemiology.

Fleiss JL, Levin B, Paik MC. Statistical Methods for Rates and Proportions.

Schlesselman JJ. Case-Control Studies: Design, Conduct and Analysis.

Examples

ss_case_control(
    odds.ratio = 2,
    p0 = 0.20
)

ss_case_control(
    odds.ratio = 3,
    p0 = 0.15,
    ratio = 2
)

Sample Size for a Two-Arm Parallel Clinical Trial

Description

Calculates the required sample size for superiority clinical trials with either continuous or binary outcomes.

Usage

ss_clinical_trial(
  outcome = c("continuous", "binary"),
  alpha = 0.05,
  power = 0.8,
  ratio = 1,
  dropout = 0,
  mean1 = NULL,
  mean2 = NULL,
  sd = NULL,
  p1 = NULL,
  p2 = NULL
)

Arguments

Details

Supported outcome types:

• Continuous (difference between two means)

• Binary (difference between two proportions)

Supports unequal allocation ratios.

Value

Object of class SampleSizeR

References

Chow SC, Shao J, Wang H. Sample Size Calculations in Clinical Research.

Julious SA. Sample Sizes for Clinical Trials.

Fleiss JL et al. Statistical Methods for Rates and Proportions.

Examples

## Continuous endpoint

ss_clinical_trial(
    outcome="continuous",
    mean1=15,
    mean2=12,
    sd=5
)

## Binary endpoint

ss_clinical_trial(
    outcome="binary",
    p1=0.30,
    p2=0.50
)

Sample Size for an Unmatched Cohort Study

Description

Calculates the required sample size for an unmatched cohort study based on the expected risk ratio, incidence among the unexposed group, statistical power and significance level.

Usage

ss_cohort(risk.ratio, p0, alpha = 0.05, power = 0.8, ratio = 1, dropout = 0)

Arguments

Details

Supports unequal exposed:unexposed allocation ratios.

Incidence among the exposed group is estimated as

p_1 = RR \times p_0

The required sample size is calculated using the Fleiss/Kelsey two-proportion formula.

Value

Object of class SampleSizeR

References

Kelsey JL et al. Methods in Observational Epidemiology.

Fleiss JL, Levin B, Paik MC. Statistical Methods for Rates and Proportions.

Chow SC, Shao J, Wang H. Sample Size Calculations in Clinical Research.

Examples

ss_cohort(
    risk.ratio = 2,
    p0 = 0.10
)

ss_cohort(
    risk.ratio = 1.8,
    p0 = 0.15,
    ratio = 2
)

Sample Size for Diagnostic Agreement Studies

Description

Estimates the required sample size for testing agreement using Cohen's kappa.

The default implementation ("pearson") is based on the multinomial agreement model and Pearson goodness-of-fit effect size.

This implementation is not the original Donner & Eliasziw (1992) sample size method.

Usage

ss_diagnostic_agreement(
  kappa1,
  kappa0 = 0.4,
  prevalence = 0.5,
  alpha = 0.05,
  power = 0.8,
  response.rate = 1,
  dropout = 0,
  method = c("pearson")
)

Arguments

Details

Computes the required sample size for studies evaluating agreement between two binary diagnostic methods.

Value

SampleSizeR object.

Sample Size for ROC Area Under the Curve (AUC)

Description

Calculates the required sample size for studies evaluating the area under the receiver operating characteristic (ROC) curve.

Usage

ss_diagnostic_auc(
  auc,
  auc0 = 0.5,
  prevalence,
  precision = NULL,
  alpha = 0.05,
  power = 0.8,
  ratio = 1,
  design = c("precision", "hypothesis"),
  alternative = c("two.sided", "one.sided"),
  method = c("obuchowski", "hanley"),
  response.rate = 1,
  dropout = 0
)

Arguments

Details

Two study designs are supported:

• Precision-based estimation of AUC

• Hypothesis testing of AUC

Two variance estimators are available:

• Obuchowski (default)

• Hanley & McNeil

Value

Object of class SampleSizeR.

References

Obuchowski NA. Statistics in Medicine. 1994.

Hanley JA, McNeil BJ. Radiology. 1982.

Zhou XH, Obuchowski NA, McClish DK. Statistical Methods in Diagnostic Medicine.

Sample Size for Diagnostic Test Sensitivity

Description

Calculates the required sample size to estimate the sensitivity of a diagnostic test with a specified confidence interval precision using the method of Buderer (1996).

Usage

ss_diagnostic_sensitivity(
  sensitivity,
  prevalence,
  precision = 0.05,
  conf.level = 0.95,
  finite.population = NULL,
  response.rate = 1,
  dropout = 0
)

Arguments

Details

The Buderer (1996) method estimates the total sample size required to achieve the desired precision for the sensitivity estimate while accounting for the expected prevalence of disease.

The required number of diseased subjects is

n_D = \frac{Z^2 Se(1-Se)}{L^2}

The total sample size is

n = \frac{n_D}{Prev}

Value

An object of class "SampleSizeR".

References

Buderer NM. Statistical Methodology: Incorporating the Prevalence of Disease into the Sample Size Calculation for Sensitivity and Specificity. Academic Emergency Medicine. 1996.

Flahault A, Cadilhac M, Thomas G. Sample size calculation should be performed for design accuracy in diagnostic test studies. Journal of Clinical Epidemiology. 2005.

Examples

ss_diagnostic_sensitivity(
    sensitivity = 0.90,
    prevalence = 0.25,
    precision = 0.05
)

Sample Size for Diagnostic Test Specificity

Description

Calculates the required sample size to estimate the specificity of a diagnostic test with a specified confidence interval precision using the method of Buderer (1996).

Usage

ss_diagnostic_specificity(
  specificity,
  prevalence,
  precision = 0.05,
  conf.level = 0.95,
  finite.population = NULL,
  response.rate = 1,
  dropout = 0
)

Arguments

Details

The Buderer (1996) method estimates the total sample size required to achieve the desired precision for the specificity estimate while accounting for the expected prevalence of disease.

The required number of non-diseased subjects is

n_{ND} = \frac{Z^2 Sp(1-Sp)}{L^2}

The total sample size is

n = \frac{n_{ND}}{1-Prev}

Value

An object of class "SampleSizeR".

References

Buderer NM. Statistical Methodology: Incorporating the Prevalence of Disease into the Sample Size Calculation for Sensitivity and Specificity. Academic Emergency Medicine. 1996.

Flahault A, Cadilhac M, Thomas G. Sample size calculation should be performed for design accuracy in diagnostic test studies. Journal of Clinical Epidemiology. 2005.

Examples

ss_diagnostic_specificity(
    specificity = 0.95,
    prevalence = 0.30,
    precision = 0.05
)

Sample Size for an Equivalence Trial

Description

Calculates the required sample size for a two-arm parallel equivalence clinical trial using the Two One-Sided Tests (TOST) procedure.

Usage

ss_equivalence_trial(
  outcome = c("continuous", "binary"),
  alpha = 0.05,
  power = 0.8,
  ratio = 1,
  dropout = 0,
  delta,
  mean1 = NULL,
  mean2 = NULL,
  sd = NULL,
  p1 = NULL,
  p2 = NULL
)

Arguments

Details

Supported endpoints:

• Continuous

• Binary (Risk Difference)

Supports equal or unequal allocation ratios.

Value

Object of class "SampleSizeR".

References

Chow SC, Shao J, Wang H. Sample Size Calculations in Clinical Research.

Piaggio G et al. Reporting of Noninferiority and Equivalence Randomized Trials.

ICH E9 Statistical Principles for Clinical Trials.

Examples

ss_equivalence_trial(
    outcome = "continuous",
    mean1 = 100,
    mean2 = 102,
    sd = 12,
    delta = 5
)

ss_equivalence_trial(
    outcome = "binary",
    p1 = 0.80,
    p2 = 0.78,
    delta = 0.10
)

Sample Size for a Non-Inferiority Trial

Description

Calculates the required sample size for a two-arm parallel non-inferiority clinical trial.

Usage

ss_noninferiority_trial(
  outcome = c("continuous", "binary"),
  alpha = 0.025,
  power = 0.8,
  alternative = c("one.sided", "two.sided"),
  ratio = 1,
  dropout = 0,
  delta,
  mean1 = NULL,
  mean2 = NULL,
  sd = NULL,
  p1 = NULL,
  p2 = NULL
)

Arguments

Details

Supported endpoints:

• Continuous

• Binary

Supports equal or unequal allocation ratios.

Value

Object of class SampleSizeR.

References

Chow SC, Shao J, Wang H. Sample Size Calculations in Clinical Research.

ICH E9.

Piaggio G et al. Reporting of Noninferiority and Equivalence Randomized Trials.

Sample Size for a Prevalence Study

Description

Calculates the minimum required sample size for estimating disease prevalence with a specified confidence level and desired precision.

Usage

ss_prevalence(
  prevalence,
  precision = NULL,
  relative.precision = NULL,
  conf.level = 0.95,
  finite.population = NULL,
  design.effect = 1,
  response.rate = 1,
  dropout = 0
)

Arguments

Details

The function implements Cochran's sample size formula and optionally adjusts for:

• finite population correction (FPC)

• design effect

• anticipated non-response

Relative precision can also be specified.

Cochran's formula

n=\frac{Z^2P(1-P)}{d^2}

where

• P = expected prevalence

• d = desired precision

• Z = Normal quantile corresponding to confidence level

If a finite population size is supplied, finite population correction is applied.

The resulting sample size is subsequently adjusted for

• design effect

• anticipated non-response

• dropout

Value

An object of class SampleSizeR.

References

Cochran WG (1977). Sampling Techniques. Third Edition. John Wiley & Sons.

Lwanga SK, Lemeshow S (1991). Sample Size Determination in Health Studies. WHO.

Naing L, Winn T, Rusli BN (2006). Practical issues in calculating the sample size for prevalence studies. Archives of Orofacial Sciences.

Examples

ss_prevalence(
  prevalence = 0.20,
  precision = 0.05
)

ss_prevalence(
  prevalence = 0.10,
  relative.precision = 0.20
)

ss_prevalence(
  prevalence = 0.15,
  precision = 0.04,
  finite.population = 2500
)

Sample Size for Cluster Prevalence Studies

Description

Calculates the required sample size for estimating prevalence using a cluster sampling design.

Usage

ss_prevalence_cluster(
  prevalence,
  precision = NULL,
  relative.precision = NULL,
  cluster.size,
  icc,
  conf.level = 0.95,
  finite.population = NULL,
  response.rate = 1,
  dropout = 0
)

Arguments

Details

The function first computes the simple random sample size using Cochran's formula and then inflates the sample size using the cluster design effect:

DEFF = 1 + (m-1)\rho

where

• m = average cluster size

• \rho = intra-cluster correlation coefficient (ICC)

The function optionally adjusts for

• finite population

• non-response

• dropout

Value

Object of class SampleSizeR

References

Cochran WG (1977). Sampling Techniques.

Donner A, Klar N (2000). Design and Analysis of Cluster Randomization Trials.

Hayes RJ, Bennett S (1999). Simple sample size calculation for cluster-randomized trials.

Examples

ss_prevalence_cluster(
prevalence=0.20,
precision=0.05,
cluster.size=20,
icc=0.05
)

Sample Size for a Superiority Trial

Description

Calculates the required sample size for a two-arm superiority clinical trial with either continuous or binary endpoints.

Usage

ss_superiority_trial(
  outcome = c("continuous", "binary"),
  alpha = 0.05,
  power = 0.8,
  alternative = c("two.sided", "one.sided"),
  ratio = 1,
  dropout = 0,
  delta,
  mean1 = NULL,
  mean2 = NULL,
  sd = NULL,
  p1 = NULL,
  p2 = NULL
)

Arguments

Details

Supports

• Continuous outcomes

• Binary outcomes

• One-sided superiority tests

• Two-sided superiority tests

• Unequal allocation ratios

Value

Object of class SampleSizeR.

References

Chow SC, Shao J, Wang H. Sample Size Calculations in Clinical Research.

ICH E9 Statistical Principles for Clinical Trials.

Julious SA. Sample Sizes for Clinical Trials.

Examples

ss_superiority_trial(
    outcome="continuous",
    mean1=16,
    mean2=12,
    sd=6,
    delta=2
)

ss_superiority_trial(
    outcome="binary",
    p1=0.70,
    p2=0.55,
    delta=0.05
)

Summary of Sample Size Result

Description

Summary of Sample Size Result

Usage

## S3 method for class 'SampleSizeR'
summary(object, ...)

Arguments

Value

Object of class summary.SampleSizeR.