Help for package PVBcorrect

Type:

Package

Title:

Partial Verification Bias Correction for Diagnostic Accuracy

Version:

0.3.1

Maintainer:

Wan Nor Arifin <wnarifin@gmail.com>

URL:

https://github.com/wnarifin/PVBcorrect/

Description:

Performs partial verification bias (PVB) correction for binary diagnostic tests, where PVB arises from selective patient verification in diagnostic accuracy studies. Supports correction of important accuracy measures – sensitivity, specificity, positive predictive values and negative predictive value – under missing-at-random and missing-not-at-random missing data mechanisms. Available methods and references are "Begg and Greenes' methods" in Alonzo & Pepe (2005) <doi:10.1111/j.1467-9876.2005.00477.x> and deGroot et al. (2011) <doi:10.1016/j.annepidem.2010.10.004>; "Multiple imputation" in Harel & Zhou (2006) <doi:10.1002/sim.2494>, "EM-based logistic regression" in Kosinski & Barnhart (2003) <doi:10.1111/1541-0420.00019>; "Inverse probability weighting" in Alonzo & Pepe (2005) <doi:10.1111/j.1467-9876.2005.00477.x>; "Inverse probability bootstrap sampling" in Nahorniak et al. (2015) <doi:10.1371/journal.pone.0131765> and Arifin & Yusof (2022) <doi:10.3390/diagnostics12112839>; "Scaled inverse probability resampling methods" in Arifin & Yusof (2025) <doi:10.1371/journal.pone.0321440>.

License:

MIT + file LICENSE

Encoding:

UTF-8

LazyData:

true

Imports:

boot, mice

RoxygenNote:

7.3.3

Suggests:

testthat (≥ 3.0.0)

Config/testthat/edition:

Depends:

R (≥ 2.10)

NeedsCompilation:

Packaged:

2025-10-02 06:59:45 UTC; wnarifin

Author:

Wan Nor Arifin

[aut, cre, cph]

Repository:

CRAN

Date/Publication:

2025-10-08 08:40:18 UTC

PVBcorrect: A package to perform partial verification bias correction for estimates of accuracy measures in diagnostic accuracy studies

Description

The package contains a number of functions to perform partial verification bias (PVB) correction for estimates of accuracy measures in diagnostic accuracy studies. The available methods are: Begg and Greenes' method (as extended by Alonzo & Pepe, 2005), Begg and Greenes' method 1 and 2 (with PPV and NPV as extended by deGroot et al, 2011), Inverse Probability Bootstrap (IPB) sampling method (Arifin & Yusof, 2022; Nahorniak et al., 2015), Scaled Inverse Probability Resampling methods (Arifin & Yusof, 2023; Arifin & Yusof, 2025), multiple imputation method by logistic regression (Harel & Zhou, 2006), and EM-based logistic regression method (Kosinski & Barnhart, 2003).

Author(s)

Maintainer: Wan Nor Arifin wnarifin@gmail.com (ORCID) [copyright holder]

References

Alonzo, T. A., & Pepe, M. S. (2005). Assessing accuracy of a continuous screening test in the presence of verification bias. Journal of the Royal Statistical Society: Series C (Applied Statistics), 54(1), 173–190.
Arifin, W. N., & Yusof, U. K. (2025). Partial Verification Bias Correction Using Scaled Inverse Probability Resampling for Binary Diagnostic Tests. medRxiv. https://doi.org/10.1101/2025.03.09.25323631
Arifin, W. N., & Yusof, U. K. (2022). Partial Verification Bias Correction Using Inverse Probability Bootstrap Sampling for Binary Diagnostic Tests. Diagnostics, 12(11), 2839.
Arifin, W. N. (2023). Partial verification bias correction in diagnostic accuracy studies using propensity score-based methods (PhD thesis, Universiti Sains Malaysia). https://erepo.usm.my/handle/123456789/19184
Arifin, W. N., & Yusof, U. K. (2022). Partial Verification Bias Correction Using Inverse Probability Bootstrap Sampling for Binary Diagnostic Tests. Diagnostics, 12, 2839.
Begg, C. B., & Greenes, R. A. (1983). Assessment of diagnostic tests when disease verification is subject to selection bias. Biometrics, 207–215.
de Groot, J. A. H., Janssen, K. J. M., Zwinderman, A. H., Bossuyt, P. M. M., Reitsma, J. B., & Moons, K. G. M. (2011). Correcting for partial verification bias: a comparison of methods. Annals of Epidemiology, 21(2), 139–148.
Harel, O., & Zhou, X.-H. (2006). Multiple imputation for correcting verification bias. Statistics in Medicine, 25(22), 3769–3786.
He, H., & McDermott, M. P. (2012). A robust method using propensity score stratification for correcting verification bias for binary tests. Biostatistics, 13(1), 32–47.
Kosinski, A. S., & Barnhart, H. X. (2003). Accounting for nonignorable verification bias in assessment of diagnostic tests. Biometrics, 59(1), 163–171.

PVB correction by Begg and Greenes' method with asymptotic normal CI

Description

PVB correction by Begg and Greenes' method with asymptotic normal CI. This is limited to no covariate.

Usage

acc_bg(data, test, disease, ci = FALSE, ci_level = 0.95, description = TRUE)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

acc_results: The accuracy results.

References

Begg, C. B., & Greenes, R. A. (1983). Assessment of diagnostic tests when disease verification is subject to selection bias. Biometrics, 207–215.
Harel, O., & Zhou, X.-H. (2006). Multiple imputation for correcting verification bias. Statistics in Medicine, 25(22), 3769–3786.
Zhou, X.-H. (1993). Maximum likelihood estimators of sensitivity and specificity corrected for verification bias. Communications in Statistics-Theory and Methods, 22(11), 3177–3198.
Zhou, X.-H. (1994). Effect of verification bias on positive and negative predictive values. Statistics in Medicine, 13(17), 1737–1745.
Zhou, X.-H., Obuchowski, N. A., & McClish, D. K. (2011). Statistical Methods in Diagnostic Medicine (2nd ed.). John Wiley & Sons.

Examples

acc_bg(data = cad_pvb, test = "T", disease = "D")  # equivalent to result by acc_ebg()
acc_bg(data = cad_pvb, test = "T", disease = "D", ci = TRUE)
  # the CIs are slightly differerent from result by acc_ebg()

Complete Case Analysis, CCA

Description

Perform Complete Case Analysis, CCA, used for complete data and multiple imputation, MI.

Usage

acc_cca(data, test, disease, ci = FALSE, ci_level = 0.95, description = TRUE)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

acc_results: The accuracy results.

Examples

acc_cca(data = cad_pvb, test = "T", disease = "D")
acc_cca(data = cad_pvb, test = "T", disease = "D", ci = TRUE)

PVB correction by Begg and Greenes' method 1 (deGroot et al, no covariate)

Description

Perform PVB correction by Begg and Greenes' method 1 as described in deGroot et al (2011), in which it also includes PPV and NPV calculation.

Usage

acc_dg1(data, test, disease, description = TRUE)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A data frame object containing the accuracy results.

References

de Groot, J. A. H., Janssen, K. J. M., Zwinderman, A. H., Bossuyt, P. M. M., Reitsma, J. B., & Moons, K. G. M. (2011). Correcting for partial verification bias: a comparison of methods. Annals of Epidemiology, 21(2), 139–148.

Examples

acc_dg1(data = cad_pvb, test = "T", disease = "D")  # equivalent to result by acc_ebg()

PVB correction by Begg and Greenes' method 2 (deGroot et al, one covariate)

Description

Perform PVB correction by Begg and Greenes' method 2 as described in deGroot et al (2011), in which it also includes PPV and NPV calculation. This is limited to only one covariate.

Usage

acc_dg2(data, test, disease, covariate, description = TRUE)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

The name(s) of covariate(s), i.e. other variables associated with either test or disease status. Specify as name vector, e.g. c("X1", "X2") for two or more variables. The variables must be in formats acceptable to GLM.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A data frame object containing the accuracy results.

References

de Groot, J. A. H., Janssen, K. J. M., Zwinderman, A. H., Bossuyt, P. M. M., Reitsma, J. B., & Moons, K. G. M. (2011). Correcting for partial verification bias: a comparison of methods. Annals of Epidemiology, 21(2), 139–148.

Examples

acc_dg2(data = cad_pvb, test = "T", disease = "D", covariate = "X3")
  # equivalent to acc_ebg(), saturated_model

PVB correction by extended Begg and Greenes' method

Description

Perform PVB correction by Begg and Greenes' method (as extended by Alonzo & Pepe, 2005).

Usage

acc_ebg(
  data,
  test,
  disease,
  covariate = NULL,
  saturated_model = FALSE,
  ci = FALSE,
  ci_level = 0.95,
  ci_type = "basic",
  R = 999,
  seednum = NULL,
  show_fit = FALSE,
  show_boot = FALSE,
  r_print_freq = 100,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

saturated_model

Set as TRUE to obtain the original Begg and Greenes' (1983) when all possible interactions are included.

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

ci_type

Set confidence interval (CI) type. Acceptable types are "norm", "basic", "perc", and "bca", for bootstrapped CI. See boot.ci for details.

R

The number of bootstrap samples. Default R = 999.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

show_fit

Set to TRUE to view model fit summary for the logistic regression model.

show_boot

Set to TRUE to show bootstrap iterations.

r_print_freq

Print the current bootstrap sample number at each specified interval. Default r_print_freq = 100.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

boot_data: An object of class "boot" from boot. Contains Sensitivity, Specificity, PPV, and NPV
boot_ci_data: A list of objects of type "bootci" from boot.ci. Contains Sensitivity, Specificity, PPV, NPV.
acc_results: The accuracy results.

References

Alonzo, T. A., & Pepe, M. S. (2005). Assessing accuracy of a continuous screening test in the presence of verification bias. Journal of the Royal Statistical Society: Series C (Applied Statistics), 54(1), 173–190.
Begg, C. B., & Greenes, R. A. (1983). Assessment of diagnostic tests when disease verification is subject to selection bias. Biometrics, 207–215.
He, H., & McDermott, M. P. (2012). A robust method using propensity score stratification for correcting verification bias for binary tests. Biostatistics, 13(1), 32–47.

Examples

# point estimates
acc_ebg(data = cad_pvb, test = "T", disease = "D")
acc_ebg(data = cad_pvb, test = "T", disease = "D", covariate = "X3")

# with bootstrapped confidence interval
acc_ebg(data = cad_pvb, test = "T", disease = "D", ci = TRUE, seednum = 12345)

PVB correction by EM-based logistic regression method

Description

Perform PVB correction by EM-based logistic regression method.

Usage

acc_em(
  data,
  test,
  disease,
  covariate = NULL,
  mnar = TRUE,
  ci = FALSE,
  ci_level = 0.95,
  ci_type = "basic",
  R = 999,
  seednum = NULL,
  show_t = TRUE,
  t_max = 500,
  cutoff = 1e-04,
  t_print_freq = 100,
  return_t = FALSE,
  r_print_freq = 100,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

mnar

The default is assuming missing not at random (MNAR) missing data mechanism, MNAR = TRUE. Set this to FALSE to obtain results assuming missing at random (MAR) missing data mechanism. This will be equivalent to using acc_ebg.

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

ci_type

Set confidence interval (CI) type. Acceptable types are "norm", "basic", "perc", and "bca", for bootstrapped CI. See boot.ci for details.

R

The number of bootstrap samples. Default R = 999.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

show_t

Print the current EM iteration number t. The default is TRUE.

t_max

The maximum iteration number for EM. Default t_max = 500. It is recommended to increase the number when covariates are included.

cutoff

The cutoff value for the minimum change between iteration. This defines the convergence of the EM procedure. Default cutoff = 0.0001. This can be set to a larger value to test the procedure.

t_print_freq

Print the current EM iteration number t at each specified interval. Default t_print_freq = 100.

return_t

Return the final EM iteration number t. This can be used for the purpose of checking the EM convergence. The default is FALSE, but is set to TRUE when ci = TRUE.

r_print_freq

Print the current bootstrap sample number at each specified interval. Default r_print_freq = 100.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

boot_data: An object of class "boot" from boot. Contains Sensitivity, Specificity, PPV, NPV and t (i.e. EM iteration taken for convergence). Use acc_em_object$boot_data$t[,5] to check the t.
boot_ci_data: A list of objects of type "bootci" from from boot.ci. Contains Sensitivity, Specificity, PPV, and NPV.
acc_results: The accuracy results.

References

Kosinski, A. S., & Barnhart, H. X. (2003). Accounting for nonignorable verification bias in assessment of diagnostic tests. Biometrics, 59(1), 163–171.

Examples

# For sample run, test with low R boot number, low t_max, low cutoff
# The results will not be good

# without covariate
em_out = acc_em(data = cad_pvb, test = "T", disease = "D", ci = TRUE, seednum = 12345,
                R = 2, t_max = 100, cutoff = 0.01)
em_out$acc_results
em_out$boot_data$t  # bootstrapped data, 1:5 columns are Sn, Sp, PPV, NPV,
                    # t (i.e. EM iteration taken for convergence)
em_out$boot_ci_data

PVB correction by inverse probability bootstrap sampling (IPB)

Description

Perform PVB correction by inverse probability bootstrap sampling.

Usage

acc_ipb(
  data,
  test,
  disease,
  covariate = NULL,
  saturated_model = FALSE,
  option = 2,
  ci = FALSE,
  ci_level = 0.95,
  ci_type = "norm",
  b = 1000,
  seednum = NULL,
  return_data = FALSE,
  return_detail = FALSE,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

saturated_model

Set as TRUE to obtain the original Begg and Greenes' (1983) when all possible interactions are included.

option

1 = IPW weight, 2 = W_h weight, described in Arifin (2023), modified weight of Krautenbacher (2017). The default is option = 2. For small weights, option = 2 is more stable (Arifin, 2023).

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

ci_type

Set confidence interval (CI) type. Acceptable types are "norm", "basic", "perc", and "bca", for bootstrapped CI.

b

The number of bootstrap samples, b.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

return_data

Return data for the bootstrapped samples.

return_detail

Return accuracy measures for each of the bootstrapped samples.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

data_each_sample: Raw data for each bootstrap sample, available with return_data = TRUE
acc_each_sample: Accuracy results for each bootstrap sample, available with return_detail = TRUE
acc_results: The accuracy results.

References

Arifin, W. N., & Yusof, U. K. (2022). Partial Verification Bias Correction Using Inverse Probability Bootstrap Sampling for Binary Diagnostic Tests. Diagnostics, 12(11), 2839.
Arifin, W. N. (2023). Partial verification bias correction in diagnostic accuracy studies using propensity score-based methods (PhD thesis, Universiti Sains Malaysia). https://erepo.usm.my/handle/123456789/19184
Krautenbacher, N., Theis, F. J., & Fuchs, C. (2017). Correcting Classifiers for Sample Selection Bias in Two-Phase Case-Control Studies. Computational and Mathematical Methods in Medicine, 2017, 1–18.
Nahorniak, M., Larsen, D. P., Volk, C., & Jordan, C. E. (2015). Using inverse probability bootstrap sampling to eliminate sample induced bias in model based analysis of unequal probability samples. PLoS One, 10(6), e0131765.

Examples

# point estimates
acc_ipb(data = cad_pvb, test = "T", disease = "D", b = 100, seednum = 12345)
acc_ipb(data = cad_pvb, test = "T", disease = "D", covariate = "X3",
        b = 100, seednum = 12345)

# with confidence interval
acc_ipb(data = cad_pvb, test = "T", disease = "D", ci = TRUE,
        b = 100, seednum = 12345)  # use small b for testing

PVB correction by Inverse Probability Weighting Estimator method

Description

Perform PVB correction by Inverse Probability Weighting Estimator method (Alonzo & Pepe, 2005).

Usage

acc_ipw(
  data,
  test,
  disease,
  covariate = NULL,
  saturated_model = FALSE,
  ci = FALSE,
  ci_level = 0.95,
  ci_type = "basic",
  R = 999,
  seednum = NULL,
  show_fit = FALSE,
  show_boot = FALSE,
  r_print_freq = 100,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

saturated_model

Set as TRUE to obtain the original Begg and Greenes' (1983) when all possible interactions are included.

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

ci_type

Set confidence interval (CI) type. Acceptable types are "norm", "basic", "perc", and "bca", for bootstrapped CI. See boot.ci for details.

R

The number of bootstrap samples. Default R = 999.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

show_fit

Set to TRUE to view model fit summary for the logistic regression model.

show_boot

Set to TRUE to show bootstrap iterations.

r_print_freq

Print the current bootstrap sample number at each specified interval. Default r_print_freq = 100.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

acc_results: The accuracy results.

References

Alonzo, T. A., & Pepe, M. S. (2005). Assessing accuracy of a continuous screening test in the presence of verification bias. Journal of the Royal Statistical Society: Series C (Applied Statistics), 54(1), 173–190.
He, H., & McDermott, M. P. (2012). A robust method using propensity score stratification for correcting verification bias for binary tests. Biostatistics, 13(1), 32–47.

Examples

# point estimates
acc_ipw(data = cad_pvb, test = "T", disease = "D")
acc_ipw(data = cad_pvb, test = "T", disease = "D", covariate = "X3")

# with bootstrapped confidence interval
acc_ipw(data = cad_pvb, test = "T", disease = "D", ci = TRUE, R = 99, seednum = 12345)

PVB correction by multiple imputation

Description

Perform PVB correction by multiple imputation.

Usage

acc_mi(
  data,
  test,
  disease,
  covariate = NULL,
  ci = FALSE,
  ci_level = 0.95,
  m = 100,
  seednum = NA,
  method = "logreg",
  mi_print = FALSE,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

m

The number of imputation, m.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

method

Imputation method. The default is "logreg". Other allowed methods are "logreg.boot", "pmm", "midastouch", "sample", "cart", "rf". See mice for details of these methods.

mi_print

Print multiple imputation history on console. This is mice print option. The default is FALSE.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

acc_results: The accuracy results.

References

Harel, O., & Zhou, X.-H. (2006). Multiple imputation for correcting verification bias. Statistics in Medicine, 25(22), 3769–3786.

Examples

# with logreg
acc_mi(data = cad_pvb, test = "T", disease = "D", ci = TRUE, seednum = 12345, m = 5)

# with other imputation method. e.g. predictive mean matching "pmm"
acc_mi(data = cad_pvb, test = "T", disease = "D", ci = TRUE, seednum = 12345, m = 5,
       method = "pmm")

# with covariate and confidence interval
acc_mi(data = cad_pvb, test = "T", disease = "D", covariate = "X3",
       ci = TRUE, seednum = 12345, m = 5)

PVB correction by scaled inverse probability weighted resampling (SIPW)

Description

Perform PVB correction by scaled inverse probability weighted resampling.

Usage

acc_sipw(
  data,
  test,
  disease,
  covariate = NULL,
  saturated_model = FALSE,
  option = 2,
  ci = FALSE,
  ci_level = 0.95,
  ci_type = "basic",
  b = 1000,
  R = 999,
  seednum = NULL,
  return_data = FALSE,
  return_detail = FALSE,
  show_boot = FALSE,
  r_print_freq = 100,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

saturated_model

Set as TRUE to obtain the original Begg and Greenes' (1983) when all possible interactions are included.

option

1 = IPW weight, 2 = W_h weight, described in Arifin (2023), modified weight of Krautenbacher (2017). The default is option = 2, which is more stable for small weights (Arifin, 2023).

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

ci_type

Set confidence interval (CI) type. Acceptable types are "norm", "basic", "perc", and "bca", for bootstrapped CI. See boot.ci for details.

b

The number of repeated samples, b.

R

The number of bootstrap samples. Default R = 999.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

return_data

Return data for the bootstrapped samples.

return_detail

Return accuracy measures for each of the bootstrapped samples.

show_boot

Set to TRUE to show bootstrap iterations.

r_print_freq

Print the current bootstrap sample number at each specified interval. Default r_print_freq = 100.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

boot_data: An object of class "boot" from boot. Contains Sensitivity, Specificity, PPV, and NPV
boot_ci_data: A list of objects of type "bootci" from boot.ci. Contains Sensitivity, Specificity, PPV, NPV.
acc_results: The accuracy results.

References

Arifin, W. N., & Yusof, U. K. (2025). Partial verification bias correction using scaled inverse probability resampling for binary diagnostic tests. PloS One, 20(9), e0321440.
Arifin, W. N., & Yusof, U. K. (2022). Partial Verification Bias Correction Using Inverse Probability Bootstrap Sampling for Binary Diagnostic Tests. Diagnostics, 12(11), 2839.
Arifin, W. N. (2023). Partial verification bias correction in diagnostic accuracy studies using propensity score-based methods (PhD thesis, Universiti Sains Malaysia). https://erepo.usm.my/handle/123456789/19184
Krautenbacher, N., Theis, F. J., & Fuchs, C. (2017). Correcting Classifiers for Sample Selection Bias in Two-Phase Case-Control Studies. Computational and Mathematical Methods in Medicine, 2017, 1–18.
Nahorniak, M., Larsen, D. P., Volk, C., & Jordan, C. E. (2015). Using inverse probability bootstrap sampling to eliminate sample induced bias in model based analysis of unequal probability samples. PLoS One, 10(6), e0131765.

Examples

# point estimates
acc_sipw(data = cad_pvb, test = "T", disease = "D", b = 100, seednum = 12345)
acc_sipw(data = cad_pvb, test = "T", disease = "D", covariate = "X3",
         b = 100, seednum = 12345)

# with bootstrapped confidence interval
acc_sipw(data = cad_pvb, test = "T", disease = "D", ci = TRUE,
         b = 100, R = 9, seednum = 12345)  # use small b, R for testing

PVB correction by scaled inverse probability weighted balanced resampling (SIPW-B).

Description

Perform PVB correction by scaled inverse probability weighted balanced resampling. SIPW-B only gives resultsfor Sensitivity and Specificity, for PPV and NPV please use SIPW instead.

Usage

acc_sipwb(
  data,
  test,
  disease,
  covariate = NULL,
  saturated_model = FALSE,
  option = 2,
  rel_size = 1,
  ci = FALSE,
  ci_level = 0.95,
  ci_type = "basic",
  b = 1000,
  R = 999,
  seednum = NULL,
  return_data = FALSE,
  return_detail = FALSE,
  show_boot = FALSE,
  r_print_freq = 100,
  description = TRUE
)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

covariate

saturated_model

Set as TRUE to obtain the original Begg and Greenes' (1983) when all possible interactions are included.

option

1 = IPW weight, 2 = W_h weight, described in Arifin (2023), modified weight of Krautenbacher (2017). The default is option = 2, which is more stable for small weights (Arifin, 2023).

rel_size

ratio control:case, D=0:D=1. The default is 1.

ci

View confidence interval (CI). The default is FALSE.

ci_level

Set the CI width. The default is 0.95 i.e. 95% CI.

ci_type

Set confidence interval (CI) type. Acceptable types are "norm", "basic", "perc", and "bca", for bootstrapped CI. See boot.ci for details.

b

The number of repeated samples, b.

R

The number of bootstrap samples. Default R = 999.

seednum

Set the seed number for the bootstrapped CI. The default is not set, so it depends on the user to set it outside or inside the function.

return_data

Return data for the bootstrapped samples.

return_detail

Return accuracy measures for each of the bootstrapped samples.

show_boot

Set to TRUE to show bootstrap iterations.

r_print_freq

Print the current bootstrap sample number at each specified interval. Default r_print_freq = 100.

description

Print the name of this analysis. The default is TRUE. This can be turned off for repeated analysis, for example in bootstrapped results.

Value

A list object containing:

acc_results: The accuracy results.

References

Arifin, W. N., & Yusof, U. K. (2025). Partial verification bias correction using scaled inverse probability resampling for binary diagnostic tests. PloS One, 20(9), e0321440.
Arifin, W. N., & Yusof, U. K. (2022). Partial Verification Bias Correction Using Inverse Probability Bootstrap Sampling for Binary Diagnostic Tests. Diagnostics, 12(11), 2839.
Arifin, W. N. (2023). Partial verification bias correction in diagnostic accuracy studies using propensity score-based methods (PhD thesis, Universiti Sains Malaysia). https://erepo.usm.my/handle/123456789/19184
Krautenbacher, N., Theis, F. J., & Fuchs, C. (2017). Correcting Classifiers for Sample Selection Bias in Two-Phase Case-Control Studies. Computational and Mathematical Methods in Medicine, 2017, 1–18.
Nahorniak, M., Larsen, D. P., Volk, C., & Jordan, C. E. (2015). Using inverse probability bootstrap sampling to eliminate sample induced bias in model based analysis of unequal probability samples. PLoS One, 10(6), e0131765.

Examples

# point estimates
acc_sipwb(data = cad_pvb, test = "T", disease = "D", b = 100, seednum = 12345)
acc_sipwb(data = cad_pvb, test = "T", disease = "D", covariate = "X3",
         b = 100, seednum = 12345)

# with bootstrapped confidence interval
acc_sipwb(data = cad_pvb, test = "T", disease = "D", ci = TRUE,
         b = 100, R = 9, seednum = 12345)  # use small b, R for testing

SPECT Thallium test data set

Description

Single-photon-emission computed-tomography (SPECT) thallium is a non-invasive diagnostic test used to diagnose coronary artery disease (CAD). SPECT thallium test was performed on 2688 patients. CAD is diagnosed when stenosis exceeds 50% of the artery, as evaluated by coronary angiography (gold standard). Only 471 patients underwent the coronary angiography for verification of the CAD status. The rest of the patients were unverified (82.5%).

Usage

cad_pvb

Format

A data frame with 2688 rows and five variables:

T:: SPECT thallium test, T: Binary, 1 = Positive, 0 = Negative
D:: CAD, D: Binary, 1 = Yes, 0 = No
X1:: Gender (covariate), X_1: Binary, 1 = Male, 0 = Female
X2:: Stress mode (covariate), X_2: Binary, 1 = Dipyridamole (Medication for stress test when the patient is unable to exercise), 0 = Exercise
X3:: Age (covariate), X_3: Binary, 1 = 60 years and above, 0 = Below 60 years

Source

Cecil, M. P., Kosinski, A. S., Jones, M. T., Taylor, A., Alazraki, N. P., Pettigrew, R. I., & Weintraub, W. S. (1996). The importance of work-up (verification) bias correction in assessing the accuracy of SPECT thallium-201 testing for the diagnosis of coronary artery disease. Journal of Clinical Epidemiology, 49(7), 735–742.
Kosinski, A. S., & Barnhart, H. X. (2003). Accounting for nonignorable verification bias in assessment of diagnostic tests. Biometrics, 59(1), 163–171.

Diaphanography test data set

Description

Diaphanography test is a noninvasive method (diagnostic test) of breast examination by transillumination using visible or infrared light to detect the presence of breast cancer. The test was performed on 900 patients. Only 88 patients were verified by breast tissue biopsy for histological examination (gold standard test). The percentage of unverified patients is 90.2%.

Usage

diapha_pvb

Format

A data frame with 900 rows and three variables:

disease:: Breast cancer, disease: Binary, 1 = Yes, 0 = No
test:: Diaphanography, test: Binary, 1 = Positive, 0 = Negative
verified:: Verified, verified: Binary, 1 = Yes, 0 = No

Source

Marshall, V., Williams, D. C., & Smith, K. D. (1984). Diaphanography as a means of detecting breast cancer. Radiology, 150(2), 339–343.

Hepatic scintigraphy test data set

Description

The data set pertains to hepatic scintigraphy, a diagnostic imaging technique used for detecting liver cancer. The test was performed on 650 patients, where 344 patients were verified by liver pathological examination (gold standard test). The percentage of unverified patients is 47.1%.

Usage

hepatic_pvb

Format

A data frame with 650 rows and three variables:

disease:: Liver cancer, disease: Binary, 1 = Yes, 0 = No
test:: Hepatic scintigraphy, test: Binary, 1 = Positive, 0 = Negative
verified:: Verified, verified: Binary, 1 = Yes, 0 = No

Source

Drum, D. E., & Christacopoulos, J. S. (1972). Hepatic scintigraphy in clinical decision making. Journal of Nuclear Medicine, 13(12), 908–915.

Test vs Disease/Gold Standard cross-classification table

Description

View Test vs Disease/Gold Standard cross-classification table.

Usage

view_table(data, test, disease, show_unverified = FALSE, show_total = FALSE)

Arguments

data

A data frame, with at least "Test" and "Disease" variables.

test

The "Test" variable name, i.e. the test result. The variable must be in binary; positive = 1, negative = 0 format.

disease

The "Disease" variable name, i.e. the true disease status. The variable must be in binary; positive = 1, negative = 0 format.

show_unverified

Optional. Set to TRUE to view observations with unverified disease status. The default is FALSE.

show_total

Optional. Set to TRUE to view total by test result. The default is FALSE.

Value

A cross-classification table.

Examples

str(cad_pvb)  # built-in data

view_table(data = cad_pvb, test = "T", disease = "D")  # without unverified observations
view_table(data = cad_pvb, test = "T", disease = "D", show_total = TRUE)
  # also with total observations by test result

view_table(data = cad_pvb, test = "T", disease = "D", show_unverified = TRUE)
  # with unverified observations
view_table(data = cad_pvb, test = "T", disease = "D", show_unverified = TRUE,
           show_total = TRUE)  # also with total observations by test result

PVBcorrect: A package to perform partial verification bias correction for estimates of accuracy measures in diagnostic accuracy studies

Description

General function

PVB correction main functions

PVB correction additional functions

Data set

Author(s)

References

See Also

PVB correction by Begg and Greenes' method with asymptotic normal CI

Description

Usage

Arguments

Value

References

Examples

Complete Case Analysis, CCA

Description

Usage

Arguments

Value

Examples

PVB correction by Begg and Greenes' method 1 (deGroot et al, no covariate)

Description

Usage

Arguments

Value

References

Examples

PVB correction by Begg and Greenes' method 2 (deGroot et al, one covariate)

Description

Usage

Arguments

Value

References

Examples

PVB correction by extended Begg and Greenes' method

Description

Usage

Arguments

Value

References

Examples

PVB correction by EM-based logistic regression method

Description

Usage

Arguments

Value

References

Examples

PVB correction by inverse probability bootstrap sampling (IPB)

Description

Usage

Arguments

Value

References

Examples

PVB correction by Inverse Probability Weighting Estimator method

Description

Usage

Arguments

Value

References

Examples

PVB correction by multiple imputation

Description

Usage

Arguments

Value

References

Examples

PVB correction by scaled inverse probability weighted resampling (SIPW)

Description

Usage

Arguments

Value

References

Examples

PVB correction by scaled inverse probability weighted balanced resampling (SIPW-B).

Description