mccount-package: mccount: Estimate Recurrent Event Burden with Competing Risks

Description

Calculates mean cumulative count (MCC) to estimate the expected cumulative number of recurrent events per person over time in the presence of competing risks and censoring. Implements both the Dong-Yasui equation method and sum of cumulative incidence method described in Dong, et al. (2015) tools:::Rd_expr_doi("10.1093/aje/kwu289"). Supports inverse probability weighting for causal inference as outlined in Gaber, et al. (2023) tools:::Rd_expr_doi("10.1093/aje/kwad031"). Provides S3 methods for printing, summarizing, plotting, and extracting results. Handles grouped analyses and integrates with 'ggplot2' https://ggplot2.tidyverse.org/ for visualization.

Arguments

Main Function

mcc() - estimates the MCC

S3 Object System

The package uses S3 classes to provide a consistent, extensible interface:

Base Class:

mcc - All MCC results inherit from this class

Method-Specific Classes:

mcc_equation - Results from Dong-Yasui estimator
mcc_sci - Results from the Sum of Cumulative Incidence estimator

Analysis-Type Classes:

mcc_weighted - Results using weighting
mcc_grouped - Results from grouped/stratified analysis

Classes combine hierarchically (e.g., c("mcc_grouped", "mcc_weighted", "mcc_equation", "mcc")).

Available Methods

Generic S3 Methods:

print.mcc() - Formatted display of results
summary.mcc() - Statistical summaries
plot.mcc() - Visualization with ggplot2
autoplot.mcc() - ggplot2-style plotting (when ggplot2 loaded)
as.data.frame.mcc() - Convert to standard data.frame
as_mcc() - Convert other objects to MCC class

Utility Functions:

is_mcc() - Test if object is MCC result
mcc_estimates() - Extract main results table
mcc_details() - Extract calculation details
mcc_method() - Get calculation method used
is_weighted(), is_grouped() - Check analysis properties
mcc_groups(), mcc_grouping_var() - Access grouping information
filter_mcc() - Filter grouped results
mcc_final_values() - Extract final MCC values
compare_mcc() - Compare two MCC objects

Basic Usage

# Calculate MCC
result <- mcc(data, "id", "time", "cause")
# Examine results
result              # Uses print.mcc()
summary(result)     # Uses summary.mcc()
plot(result)        # Uses plot.mcc()
# Extract components
estimates <- mcc_estimates(result)
details <- mcc_details(result)
final_values <- mcc_final_values(result)
# Grouped analysis
grouped_result <- mcc(data, "id", "time", "cause", by = "treatment")
plot(grouped_result)
filter_mcc(grouped_result, "Treatment A")

Plotting

The package provides flexible plotting through S3 methods that automatically adapt to analysis type:

# Basic plotting
plot(mcc_result)                    # MCC over time
plot(mcc_result, type = "details")  # Calculation components
# Customization
plot(mcc_result, colors = c("red", "blue"), title = "Custom Title")
# ggplot2 integration
library(ggplot2)
autoplot(mcc_result) + theme_classic()
# Further customization
plot(mcc_result) +
  geom_hline(yintercept = 1, linetype = "dashed") +
  labs(caption = "Dashed line at MCC = 1")

Author

Maintainer: Kenneth A. Taylor kenneth.taylor.dpt@gmail.com (ORCID) [copyright holder]

References

Core Methods: Dong H, Robison LL, Leisenring WM, Martin LJ, Armstrong GT, Yasui Y. Estimating the burden of recurrent events in the presence of competing risks: the method of mean cumulative count. Am J Epidemiol. 2015;181(7):532-40.

Weighted Extension: Gaber CE, Edwards JK, Lund JL, Peery AF, Richardson DB, Kinlaw AC. Inverse Probability Weighting to Estimate Exposure Effects on the Burden of Recurrent Outcomes in the Presence of Competing Events. Am J Epidemiol. 2023;192(5):830-839.