Create a Measurement Error Object

Description

This function creates a measurement error object, usually used as a covariate or the outcome in the formula argument of mecor if one wants to correct for the measurement error in that variable using a reference variable or a replicate measure.

Usage

MeasError(substitute, reference, replicate, differential)

Arguments

Value

MeasError returns an object of class "MeasError".

An object of class MeasError is a list containing the substitute and reference (and replicate or differential if applicable) variables and has attributes input (the name of the substitute and reference or replicate and differential (if applicable) variables) and call (the matched call).

Author(s)

Linda Nab, l.nab@lumc.nl

Examples

## measurement error in a covariate:
# internal covariate-validation study
data(vat)
with (vat, MeasError(substitute = wc,
                     reference = vat))
# replicates study
data(bloodpressure)
with (bloodpressure, MeasError(substitute = sbp30,
                               replicate = cbind(sbp60, sbp120)))
# outcome-calibration study
data(sodium)
with(sodium, MeasError(substitute = recall,
                       replicate = cbind(urinary1, urinary2)))
## measurement error in the outcome:
# internal outcome-validation study
data(haemoglobin)
with(haemoglobin, MeasError(substitute = capillary,
                            reference = venous))
# internal outcome- validation study with differential measurement error in
# the dependent variable
data(haemoglobin)
with(haemoglobin, MeasError(substitute = capillary,
                            reference = venous,
                            differential = supplement))

Create an External Measurement Error Object

Description

This function creates an external measurement error object, usually used as a covariate or the outcome in the formula argument of mecor if one wants to correct for the measurement error in that variable using external data or externally estimated coefficients of the calibration model (covariate-measurement error) or measurement error model (outcome-measurement error)

Usage

MeasErrorExt(substitute, model)

Arguments

Value

MeasErrorExt returns an object of class "MeasErrorExt".

An object of class MeasErrorExt is a list containing the substitute variable and the fitted calibration model or measurement error model and has attributes input (the name of the substitute variable) and call (the matched call).

Author(s)

Linda Nab, l.nab@lumc.nl

Examples

## measurement error in a outcome:
# external outcome-validation study
data(haemoglobin_ext)
# calibration model
calmod_fit <- lm(capillary ~ venous, data = haemoglobin)
# the external covariate-validation study can be used to correct for the
# measurement error in X_star in the dataset 'icvs', using the fitted
# calibration model
data(haemoglobin)
with (haemoglobin, MeasErrorExt(substitute = capillary,
                                model = calmod_fit))
# identical to:
calmod_coef <- coefficients(calmod_fit)
calmod_vcov <- vcov(calmod_fit)
with (haemoglobin, MeasErrorExt(substitute = capillary,
                                model = list(coef = calmod_coef,
                                             vcov = calmod_vcov)))
# when no external data is available, guesstimations of the coefficients of
# the calibration model can be used instead:
with (haemoglobin, MeasErrorExt(substitute = capillary,
                                model = list(coef = c('(Intercept)' = -7,
                                                      'venous' = 1.1))))

Create a Random Measurement Error Object

Description

This function creates a random measurement error object, usually used as a covariate in the formula argument of mecor if one wants to correct for random measurement error in that variable

Usage

MeasErrorRandom(substitute, variance)

Arguments

Value

MeasErrorRandom returns an object of class "MeasErrorRandom".

An object of class MeasErrorRandom is a list containing the substitute variable, the assumed variance of the random measurement error in that variable and, the attributes input (the name of the substitute variable) and call (the matched call).

Author(s)

Linda Nab, l.nab@lumc.nl

Examples

## random measurement error in a covariate:
# internal covariate-validation study
data(bloodpressure)
with(bloodpressure, MeasErrorRandom(sbp30, variance = 0.25))

PDAC blood pressure data [replicates study]

Description

Blood pressure, age and creatinine levels of 450 pregnant women from the Pregnancy Day Assessment Clinic.

Usage

bloodpressure

Format

A data frame with 450 rows and 6 variables:

creatinine

Serum creatinine (umol/L)

age

Age (years)

sbp30

Systolic blood pressure at 30 minutes (mm Hg)

sbp60

Systolic blood pressure at 60 minutes (mm Hg)

sbp90

Systolic blood pressure at 90 minutes (mm Hg)

sbp120

Systolic blood pressure at 120 minutes (mm Hg)

Details

This is a simulated dataset inspired by data that was originally published at the Dryad Digital Repository: <doi:10.5061/dryad.0bq15>

References

Elizabeth Anne McCarthy, Thomas A Carins, Yolanda Hannigan, Nadia Bardien, Alexis Shub, and Susan P Walker. Data from: Effectiveness and safety of 1 vs 4h blood pressure profile with clinical and laboratory assessment for the exclusion of gestational hypertension and pre-eclampsia: a retrospective study in a university affiliated maternity hospital. Dryad (2015). <doi:10.5061/dryad.0bq15>.

Examples

data("bloodpressure", package = "mecor")

Low-dose iron supplements haemoglobin data [internal outcome-validation study]

Description

Capillary haemoglobin and venous haemoglobin levels of 400 subjects of a trial investigating the efficacy of low-dose iron supplements during pregnancy. Venous haemoglobin levels were observed of approximately 25% of the subjects included in the trial.

Usage

haemoglobin

Format

A data frame with 400 rows and 3 variables:

capillary

Haemoglobin levels measured in capillary blood (g/L)

supplement

Low-dose iron supplement (20 mg/d) (0 = no, 1 = yes)

venous

Haemoglobin levels measured in venous blood (g/L)

Details

This is a simulated data set inspired by a trial investigating low-dose iron supplements <doi:10.1093/ajcn/78.1.145>. A motivating example using the example data can be found here: <doi:10.1002/sim.8359>

References

Maria Makrides, Caroline A Crowther, Robert A Gibson, Rosalind S Gibson, and C Murray Skeaff. Efficacy and tolerability of low-dose iron supplements during pregnancy: a randomized controlled trial. The American Journal of Clinical Nutrition (2003). <doi:10.1093/ajcn/78.1.145>

Linda Nab, Rolf HH Groenwold, Paco MJ Welsing, and Maarten van Smeden. Measurement error in continuous endpoints in randomised trials: Problems and solutions. Statistics in Medicine (2019). <doi:10.1002/sim.8359>

Examples

data("haemoglobin", package = "mecor")

Haemoglobin external data [external outcome-validation study]

Description

Capillary haemoglobin and venous haemoglobin levels of 100 individuals.

Usage

haemoglobin_ext

Format

A data frame with 100 rows and 2 variables:

capillary

Haemoglobin levels measured in capillary blood (g/L)

venous

Haemoglobin levels measured in venous blood (g/L)

Details

This is a simulated data set accompanying the dataset "haemoglobin", that is inspired by a trial investigating low-dose iron supplements <doi:10.1093/ajcn/78.1.145>. A motivating example using the example data can be found here: <doi:10.1002/sim.8359>

References

Maria Makrides, Caroline A Crowther, Robert A Gibson, Rosalind S Gibson, and C Murray Skeaff. Efficacy and tolerability of low-dose iron supplements during pregnancy: a randomized controlled trial. The American Journal of Clinical Nutrition (2003). <doi:10.1093/ajcn/78.1.145>

Linda Nab, Rolf HH Groenwold, Paco MJ Welsing, and Maarten van Smeden. Measurement error in continuous endpoints in randomised trials: Problems and solutions. Statistics in Medicine (2019). <doi:10.1002/sim.8359>

Examples

data("haemoglobin_ext", package = "mecor")

Weighting for Confounding and Joint Misclassification of Exposure and Outcome

Description

ipwm implements a method for estimating the marginal causal odds ratio by constructing weights (modified inverse probability weights) that address both confounding and joint misclassification of exposure and outcome.

Usage

ipwm(
  formulas,
  data,
  outcome_true,
  outcome_mis = NULL,
  exposure_true,
  exposure_mis = NULL,
  nboot = 1000,
  conf_level = 0.95,
  fix_nNAs = FALSE,
  semiparametric = FALSE,
  optim_args = list(method = "BFGS"),
  force_optim = FALSE,
  sp = Inf,
  print = TRUE
)

Arguments

Details

This function is an implementation of the weighting method described by Penning de Vries et al. (2018). The method defaults to the estimator proposed by Gravel and Platt (2018) in the absence of exposure misclassification.

The function assumes that the exposure or the outcome has a misclassified version. An error is issued when both outcome_mis and exposure_mis are set to NULL.

Provided force_optim = FALSE, ipwm is considerably more efficient when the optim function is not invoked; i.e., when (1) exposure_mis = NULL and the formula for outcome_true does not contain stats::terms involving outcome_mis or exposure_true, (2) outcome_mis = NULL and the formula for exposure_true does not contain stats::terms involving exposure_mis or outcome_true, or (3) all(is.na(data[, exposure_true]) == is.na(data[, outcome_true])) and the formulas for exposure_true and outcome_true do not contain stats::terms involving exposure_mis or outcome_mis. In these cases, ipwm uses iteratively reweighted least squares via the glm function for maximum likelihood estimation. In all other cases, optim_args is passed on to optim for optimisation of the joint likelihood of outcome_true, outcome_mis, exposure_true and exposure_mis.

Value

ipwm returns an object of class ipwm. The returned object is a list containing the following elements:

If nboot != 0, the list also contains

Author(s)

Bas B. L. Penning de Vries, b.b.l.penning_de_vries@lumc.nl

References

Gravel, C. A., & Platt, R. W. (2018). Weighted estimation for confounded binary outcomes subject to misclassification. Statistics in medicine, 37(3), 425-436. https://doi.org/10.1002/sim.7522

Penning de Vries, B. B. L., van Smeden, M., & Groenwold, R. H. H. (2020). A weighting method for simultaneous adjustment for confounding and joint exposure-outcome misclassifications. Statistical Methods in Medical Research, 0(0), 1-15. https://doi.org/10.1177/0962280220960172

Examples

data(sim) # simulated data on 10 covariates, exposure A and outcome Y.
formulas <- list(
  Y ~ A + L1 + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 + L10 + B + Z,
  A ~ L1 + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 + L10 + B + Z,
  Z ~ L1 + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 + L10 + B,
  B ~ L1 + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 + L10
)
## Not run: 
ipwm_out <- ipwm(
  formulas = formulas,
  data = sim,
  outcome_true = "Y",
  outcome_mis = "Z",
  exposure_true = "A",
  exposure_mis = "B",
  nboot = 200,
  sp = 1e6
)
ipwm_out

## End(Not run)

mecor: a Measurement Error Correction Package

Description

mecor provides correction methods for measurement error in a continuous covariate or outcome in linear regression models with a continuous outcome

Usage

mecor(formula, data, method = "standard", B = 0)

Arguments

Value

mecor returns an object of class "mecor".

An object of class mecor is a list containing the following components:

Author(s)

Linda Nab, l.nab@lumc.nl

References

L. Nab, R.H.H. Groenwold, P.M.J. Welsing, and M. van Smeden. Measurement error in continuous endpoints in randomised trials: problems and solutions

L. Nab, M. van Smeden, R.H. Keogh, and R.H.H. Groenwold. mecor: an R package for measurement error correction in linear models with continuous outcomes

Examples

## measurement error in a covariate/outcome:
# internal covariate-validation study
data(vat)
out <-
mecor(ir_ln ~ MeasError(wc, reference = vat) + sex + age + tbf,
      data = vat,
      method = "standard",
      B = 999)
# replicates study
data(bloodpressure)
mecor(creatinine ~ MeasError(sbp30, replicate = cbind(sbp60, sbp120)) + age,
      data = bloodpressure,
      method = "mle")
# outcome-calibration study
data(sodium)
mecor(MeasError(recall, replicate = cbind(urinary1, urinary2)) ~ diet,
      data = sodium,
      method = "efficient")
# external outcome-validation study
data(haemoglobin_ext)
calmod_fit <- lm(capillary ~ venous, data = haemoglobin_ext)
data(haemoglobin) # suppose reference venous is not available
mecor(MeasErrorExt(capillary, model = calmod_fit) ~ supplement,
      data = haemoglobin)
# sensitivity analyses
data(vat) # suppose reference vat is not available
# guesstimate the coefficients of the calibration model:
mecor(ir_ln ~ MeasErrorExt(wc, model = list(coef = c(0.2, 0.5, -1.3, 0, 0.6))) + sex + age + tbf,
      data = vat)
# assume random measurement error in wc of magnitude 0.25:
mecor(ir_ln ~ MeasErrorRandom(wc, variance = 0.25) + sex + age + tbf,
      data = vat)
data(bloodpressure) # suppose replicates sbp60 and sbp60 are not available
mecor(creatinine ~ MeasErrorRandom(sbp30, variance = 25) + age,
      data = bloodpressure)

## differential measurement error in the outcome:
# internal outcome-validation study
mecor(MeasError(capillary, reference = venous, differential = supplement) ~ supplement,
      data = haemoglobin,
      method = "standard")

Simulated dataset for the ipwm function

Description

A simulated dataset containing 5000 observations of the covariates L1-L10, the true exposure A and true outcome Y, and the misclassified exposure B and misclassified outcome Z.

Usage

sim

Format

A data frame with 5000 rows and 14 variables:

L1

covariate, binary

L2

covariate, continuous

L3

covariate, binary

L4

covariate, continuous

L5

covariate, binary

L6

covariate, binary

L7

covariate, continuous

L8

covariate, binary

L9

covariate, binary

L10

covariate, continuous

A

exposure, binary

Y

outcome, binary

B

misclassified exposure, binary

Z

misclassified outcome, binary

Examples

data("sim", package = "mecor")

TONE sodium data [outcome-calibration study]

Description

Self-reported sodium intake and urinary sodium in the TONE study, a randomized controlled trial designed to investigate whether a reduction in sodium intake results in satisfactory blood pressure control. Two replicate urinary sodium measures were available in 50% of the subjects included in the trial.

Usage

sodium

Format

A data frame with 1000 rows and 4 variables:

recall

Sodium intake measured by a 24h recall (mg)

diet

Usual diet or sodium-lowering diet (0 = usual, 1 = sodium-lowering)

urinary1

Sodium intake measured in urine (1st measure, mg)

urinary2

Sodium intake measured in urine (2nd measure, mg)

Details

This is a simulated data set inspired by the TONE study <doi: 10.1016/1047-2797(94)00056-y>. A motivating example using the example data can be found here: <doi:10.1002/sim.7011>

References

Lawrence J Appel, Mark Espeland, Paul K Whelton, Therese Dolecek, Shiriki Kumanyika, William B Applegate, Walter H Ettinger, John B Kostis, Alan C Wilson, Clifton Lacy, and Stephen T Miller. Trial of Nonpharmacologic Intervention in the Elderly (TONE). Design and rationale of a blood pressure control trial. Annals of Epidemiology (1995). <doi: 10.1016/1047-2797(94)00056-y>

Ruth H Keogh, Raymond J Carroll, Janet A Tooze, Sharon I Kirkpatrick, Laurence S Freedman. Statistical issues related to dietary intake as the response variable in intervention trials. Statistics in Medicine (2016). <doi:10.1002/sim.7011>

Examples

data("sodium", package = "mecor")

Summarizing Measurement Error Correction

Description

summary method for class "mecor"

Usage

## S3 method for class 'mecor'
summary(object, alpha = 0.05, zerovar = FALSE, fieller = FALSE, ...)

Arguments

Value

The function summary.mecor returns a list of summary statistics of the fitted corrected model and fitted uncorrected model.

See Also

The model fitting function mecor, summary

Examples

## measurement error in a covariate:
# internal covariate-validation study
data(vat)
mecor_fit <- mecor(ir_ln ~ MeasError(wc, reference = vat) + sex + age + tbf,
                   data = vat,
                   method = "standard")
summary(mecor_fit)
summary(mecor_fit, zerovar = TRUE, fieller = TRUE)
summary(mecor_fit, alpha = 0.10)

NEO visceral adipose tissue data [internal covariate-validation study]

Description

Insulin resistance, waist circumference, sex, age, total body fat and visceral adipose tissue of 650 individuals from the Netherlands Epidemiology of Obesity (NEO) study. Visceral adipose tissue measurements were taken of approximately 40% of the individuals, at random.

Usage

vat

Format

A data frame with 650 rows and 6 variables:

ir_ln

Natural logarithm of insulin resistance (fasting glucose (mmol/L) x fasting insulin (mU/L) / 22.5)

wc

Waist circumference (standardised, cm)

sex

Sex (0 = male, 1 = female)

age

Age (years)

tbf

Total body fat (standardised, %)

vat

Visceral adipose tissue (standardised, cm^2)

Details

This is a simulated data set inspired by the NEO data <doi:10.1007/s10654-013-9801-3>. A motivating example using the example data can be found here: <doi:10.1093/aje/kwab114>

References

Renee de Mutsert, Martin den Heijer, Ton J Rabelink, Johannes WA Smit, Johannes A Romijn, Johan W Jukema, Albert de Roos, Christa M Cobbaert, Margreet Kloppenburg, Saskia le Cessie, Saskia Middeldorp, Frits R Rosendaal. The Netherlands epidemiology of obesity (NEO) study: Study design and data collection. European Journal of Epidemiology (2013). <doi:10.1007/s10654-013-9801-3>

Linda Nab, Maarten van Smeden, Renee de Mutsert, Frits R Rosendaal, and Rolf HH Groenwold. Sampling strategies for internal validation samples for exposure measurement error correction: A study of visceral adipose tissue measures replaced by waist circumference measures. American Journal of Epidemiology (2021). <doi:10.1093/aje/kwab114>

Examples

data("vat", package = "mecor")

Visceral adipose tissue external data [external covariate-validation study]

Description

Waist circumference, visceral adipose tissue, sex, age, and total body fat of 100 individuals

Usage

vat_ext

Format

A data frame with 100 rows and 5 variables:

wc

Waist circumference (standardised, cm)

vat

Visceral adipose tissue (standardised, cm^2)

sex

Sex (0 = male, 1 = female)

age

Age (years)

tbf

Total body fat (standardised, %)

Details

This is a simulated data set accompanying the dataset "vat", that is inspired by the NEO data <doi:10.1007/s10654-013-9801-3>. A motivating example using the example data can be found here: <doi:10.1093/aje/kwab114>

References

Renee de Mutsert, Martin den Heijer, Ton J Rabelink, Johannes WA Smit, Johannes A Romijn, Johan W Jukema, Albert de Roos, Christa M Cobbaert, Margreet Kloppenburg, Saskia le Cessie, Saskia Middeldorp, Frits R Rosendaal. The Netherlands epidemiology of obesity (NEO) study: Study design and data collection. European Journal of Epidemiology (2013). <doi:10.1007/s10654-013-9801-3>

Linda Nab, Maarten van Smeden, Renee de Mutsert, Frits R Rosendaal, and Rolf HH Groenwold. Sampling strategies for internal validation samples for exposure measurement error correction: A study of visceral adipose tissue measures replaced by waist circumference measures. American Journal of Epidemiology (2021). <doi:10.1093/aje/kwab114>

Examples

data("vat_ext", package = "mecor")