R/datasets.R

In survival-lumc/CauseSpecCovarMI: Multiple imputation for cause-specific Cox models

#' Summarised simulation results - regression coefficients
#'
#' @docType data
#' 
#' @format 
#' \describe{
#' \item{var}{Variable for which a regression coefficient is being estimated, for example
#' "X.1" corresponds to the cause-specific effect of X on event 1 (Relapse)}
#' \item{m}{Number of imputed datasets}
#' \item{analy}{Method indicator, please refer to manuscript for abbreviations used}
#' \item{true}{True data-generating value for regression coefficient}
#' \item{scen_num}{Number label for scenario}
#' \item{n}{Sample size in simulated datasets}
#' \item{est}{Estimated regression coefficient}
#' \item{se}{Estimate model standard error (SE) for regression coefficient}
#' \item{se_mcse}{Monte carlo standard error for model SE}
#' \item{emp_se}{Empirical SE}
#' \item{cover}{Coverage}
#' \item{bias}{Bias}
#' \item{rmse}{Coverage}
#' \item{rmse_mcse}{Coverage}
#' \item{warns}{Number of warnings across all simulations of a particular scenario, for 
#' a coefficient. Generally corresponds to number of smcfcs rejection sampling failures.}
#' \item{bias_mcse}{Monte carlo error of bias}
#' \item{cover_mcse}{Monte carlo error of coverage}
#' \item{prop_miss}{Proportion of missing values in X}
#' \item{beta1}{Data-generating ffect of X on event 1}
#' \item{miss_mech}{Missingness mechanism}
#' \item{X_level}{Measurement level of X}
#' \item{rho}{Correlation between X and Z}
#' \item{eta1}{Strength of missingness mechanism for non-MCAR scenarios}
#' \item{haz_shape}{Shapes of baseline hazard for competing events, either "similar"
#'  or "different"}
#' }
#'
#' @usage regr_results
#'
#' @keywords datasets
"regr_results"

#' Summarised simulation results - predictions
#'
#' @docType data
#' 
#' @format 
#' \describe{
#' \item{analy}{Method indicator, please refer to manuscript for abbreviations used}
#' \item{m}{Number of imputed datasets}
#' \item{combo-X_Z}{Covariate combination defining reference patient for which
#' we predict}
#' \item{times}{Prediction horizon (in years)}
#' \item{state}{State for which we estimate probability at times}
#' \item{true}{True data-generating cumulative incidence at a horizon, and for 
#' a certain reference patient}
#' \item{scen_num}{Number label for scenario}
#' \item{n}{Sample size in simulated datasets}
#' \item{prob}{Estimated probability}
#' \item{emp_se}{Empirical SE of the probabilities}
#' \item{bias}{Bias}
#' \item{rmse}{Coverage}
#' \item{rmse_mcse}{Coverage}
#' \item{bias_mcse}{Monte carlo error of bias}
#' \item{prop_miss}{Proportion of missing values in X}
#' \item{beta1}{Data-generating ffect of X on event 1}
#' \item{miss_mech}{Missingness mechanism}
#' \item{X_level}{Measurement level of X}
#' \item{rho}{Correlation between X and Z}
#' \item{eta1}{Strength of missingness mechanism for non-MCAR scenarios}
#' \item{haz_shape}{Shapes of baseline hazard for competing events, either "similar"
#'  or "different"}
#' }
#'
#' @usage preds_results
#'
#' @keywords datasets
"preds_results"


#' Scenarios table for simulation study
#' 
#' Contains all 119 scenario from simulation study as generated by
#' \code{analysis/simulations/make-scenarios.R}.
#' 
#' @docType data 
#' 
#' @format 
#' \describe{
#' \item{n}{Sample size in simulated datasets}
#' \item{prop_miss}{Proportion of missing values in X}
#' \item{beta1}{Data-generating ffect of X on event 1}
#' \item{miss_mech}{Missingness mechanism}
#' \item{X_level}{Measurement level of X}
#' \item{rho}{Correlation between X and Z}
#' \item{eta1}{Strength of missingness mechanism for non-MCAR scenarios}
#' \item{haz_shape}{Shapes of baseline hazard for competing events, either "similar"
#'  or "different"}
#' \item{pilot}{Indicator of whether scenario is part of initial "pilot" scenarios
#' which were run to trial the simulations. Corresponds to first 15 scenarios, and 
#' is not relevant for the reader}
#' \item{scen_num}{Number label for a scenario - helps in defining the seeds}
#' \item{seed}{Base seed (used in \code{set.seed}) for the first replication of 
#' a scenario. The seed used for any other replication is then this base seed + 
#' replication number}
#' }
#' 
#' @usage scenarios
#' 
#' @keywords datasets
"scenarios"


#' Estimated baseline hazard parameters from MDS data
#'
#' @docType data 
#' 
#' @format 
#' \describe{
#' \item{state}{Event for which we would like to generate event times, either REL
#' (relapse), NRM (non-relapse mortality) or EFS (event-free)}
#' \item{shape}{Shape of Weibull distribution}
#' \item{rate}{Rate of Weibull distribution}
#' }
#' 
#' @usage mds_shape_rates
#' 
#' @keywords datasets
"mds_shape_rates"


#' Synthetic version of MDS long-term data
#'
#' @docType data 
#' 
#' @encoding UTF-8
#' 
#' @format 
#' \describe{
#' \item{ci_s_allo1}{Competing event indicator: 0 = censored, 1 = relapse,
#' 2 = non-relapse mortality.}
#' \item{ci_allo1}{Time to first competing event (years).}
#' \item{match_allo1_1}{Sex match patient and donor. Unordered categorical with four levels,
#' corresponding to all combinations of recipient male/female and donor male/female.}
#' \item{mdsclass}{MDS groups based on subclassification at alloHCT. Categorical with three
#' levels: MDS without excess blasts, with excess blast and sAML.}
#' \item{donorrel}{Human leukocyte antigen match between patient and donor. Binary: 
#' Identical sibling or Other.}
#' \item{karnofsk_allo1}{Karnofsky performance status. Three-level ordered categorical:
#' >=90, 80 and <=70.}
#' \item{crnocr}{Stage at alloHCT. Three-level unordered categorical: CR, 
#'  noCR and Untreated/not aimed at remission.}
#' \item{cmv_combi_allo1_1}{Cytomegalovirus (CMV) status in patient and donor. 
#' Four-level categorical corresponding to all combinations of patient positive/negative
#' and donor positive/negative. For example, level "+/-" corresponds to CMV positive
#' in patient and negative in donor.}
#' \item{cytog_threecat}{Cytogenetics categories for International Prognostic 
#' Scoring System (IPSS-R) - see Greenberg et al. (2012).}
#' \item{hctci_risk}{Hematopoietic stemcell transplantation-comorbidity index score.
#' Three-level ordered categorical: low risk (0), intermediate risk (1-2) and 
#' high risk (>= 3). See Sorror et al. (2005).}
#' \item{agedonor_allo1_decades}{Donor age at alloHCT (decades).}
#' \item{age_allo1_decades}{Patient age at alloHCT (decades).}
#' }
#' 
#' @usage dat_mds_synth
#' 
#' @references {
#' Greenberg, Peter L., Heinz Tuechler, Julie Schanz, Guillermo Sanz, 
#' Guillermo Garcia-Manero, Francesc Solé, John M. Bennett et al. 
#' "Revised international prognostic scoring system for myelodysplastic syndromes." 
#' \emph{Blood} 120, no. 12 (2012): 2454-2465.
#' 
#' Sorror, Mohamed L., Michael B. Maris, Rainer Storb, Frederic Baron, Brenda M. Sandmaier, 
#' David G. Maloney, and Barry Storer. 
#' "Hematopoietic cell transplantation (HCT)-specific comorbidity index: 
#' a new tool for risk assessment before allogeneic HCT." 
#' \emph{Blood} 106, no. 8 (2005): 2912-2919.
#' }
#' 
#' @keywords datasets
"dat_mds_synth"