cmsens is used to conduct sensitivity analysis for unmeasured confounding via the E-value approach by Vanderweele et al. (2017) and Smith et al. (2019), and sensitivity analysis for measurement error via regression calibration by Carroll et al. (1995) and SIMEX by Cook et al. (1994) and Küchenhoff et al. (2006).

cmsens(
  object = NULL,
  sens = "uc",
  MEmethod = "simex",
  MEvariable = NULL,
  MEvartype = NULL,
  MEerror = NULL,
  lambda = c(0.5, 1, 1.5, 2),
  B = 200,
  nboot.rc = 400
)

# S3 method for cmsens.uc
print(x, ...)

# S3 method for cmsens.me
print(x, ...)

# S3 method for cmsens.me
summary(object, ...)

# S3 method for summary.cmsens.me
print(x, digits = 4, ...)

Arguments

object

an object of class cmest.

sens

sensitivity analysis for unmeasured confounding or measurement error. uc represents unmeasured confounding and me represents measurement error. See Details.

MEmethod

method for measurement error correction. rc represents regression calibration and simex represents SIMEX. See Details.

MEvariable

variable measured with error.

MEvartype

type of the variable measured with error. Can be continuous or categorical (first 3 letters are enough).

MEerror

a vector of standard deviations of the measurement error (when MEvartype is continuous) or a list of misclassification matrices (when MEvartype is categorical).

lambda

a vector of lambdas for SIMEX. Default is c(0.5, 1, 1.5, 2).

B

number of simulations for SIMEX. Default is 200.

nboot.rc

number of boots for correcting the var-cov matrix of coefficients with regression calibration. Default is 400.

x

an object of class cmsens

...

other arguments.

digits

minimal number of significant digits. See print.default.

Value

If sens is uc, an object of class cmsens.uc is returned:

call

the function call,

evalues

a data frame in which the first three columns are point estimates, lower limits of 95% confidence intervals and upper limits of 95% confidence intervals of causal effects on the risk or rate ratio scale and the last three columns are E-values on the risk or rate ratio scale,

If sens is me, an object of class cmsens.me is returned:
call

the function call,

ME

a list which might contain MEmethod, MEvariable, MEvartype, MEerror, lambda, B, nboot.rc and reliability ratio (which is calculated by 1 - MEerror[i]/sd(data[, MEvariable]) for i=1,...,length(MEerror) when MEvartype is continuous),

naive

naive causal mediation analysis results,

sens

a list of causal mediation analysis results after correcting errors in MEerror,

...

Details

Sensitivity Analysis for Unmeasured Confounding

Currently, sensitivity analysis for unmeasured confounding are available when the outcome regression model is fitted by lm, glm, glm.nb, gam, multinom, polr.

All E-values are reported on the risk or rate ratio scale. If the causal effects are estimated on the difference scale (i.e., the outcome is continuous), they are transformed into risk ratios using the transformation described by Vanderweele et al. (2017).

Sensitivity Analysis for Measurement Error

Currently, sensitivity analysis for measurement error are available:

1) when the regression model involving the variable measured with error is fitted by lm, glm (with family gaussian, binomial or poisson), multinom, polr, coxph or survreg and model is rb or gformula; 2) when estimation is paramfunc.

Sensitivity analysis for measurement error only supports a single variable measured with error. Regression calibration requires that the variable measured with error be an independent continuous variable in the regression it's involved in. SIMEX supports a continuous or categorical variable measured with error. Quadratic extrapolation method is implemented for SIMEX.

Functions

References

VanderWeele TJ, Ding P (2017). Sensitivity analysis in observational research: introducing the E-Value. Annals of Internal Medicine. 167(4): 268 - 274.

Smith LH, VanderWeele TJ (2019). Mediational E-values: Approximate sensitivity analysis for unmeasured mediator-outcome confounding. Epidemiology. 30(6): 835 - 837.

Carrol RJ, Ruppert D, Stefanski LA, Crainiceanu C (2006). Measurement Error in Nonlinear Models: A Modern Perspective, Second Edition. London: Chapman & Hall.

Cook JR, Stefanski LA (1994). Simulation-extrapolation estimation in parametric measurement error models. Journal of the American Statistical Association, 89(428): 1314 - 1328.

Küchenhoff H, Mwalili SM, Lesaffre E (2006). A general method for dealing with misclassification in regression: the misclassification SIMEX. Biometrics. 62(1): 85 - 96.

Stefanski LA, Cook JR (1995). Simulation-extrapolation: the measurement error jackknife. Journal of the American Statistical Association. 90(432): 1247 - 56.

Valeri L, Lin X, VanderWeele TJ (2014). Mediation analysis when a continuous mediator is measured with error and the outcome follows a generalized linear model. Statistics in medicine, 33(28): 4875 – 4890.

See also

Examples

if (FALSE) { library(CMAverse) # 10 boots are used for illustration naive <- cmest(data = cma2020, model = "rb", outcome = "contY", exposure = "A", mediator = c("M1", "M2"), basec = c("C1", "C2"), EMint = TRUE, mreg = list("logistic", "multinomial"), yreg = "linear", astar = 0, a = 1, mval = list(0, "M2_0"), estimation = "imputation", inference = "bootstrap", nboot = 10) exp1 <- cmsens(object = naive, sens = "uc") exp2 <- cmsens(object = naive, sens = "me", MEmethod = "rc", MEvariable = "C1", MEvartype = "con", MEerror = c(0.1, 0.2)) summary(exp2) # B = 10 is used for illustration exp3 <- cmsens(object = naive, sens = "me", MEmethod = "simex", MEvariable = "M1", MEvartype = "cat", MEerror = list(matrix(c(0.95,0.05,0.05,0.95), nrow = 2), matrix(c(0.9,0.1,0.1,0.9), nrow = 2)), B = 10) summary(exp3) }