Survival Analysis for Continuous Variable

Survival Analysis for Continuous Variable

Usage

survivalcont(
  data,
  elapsedtime = NULL,
  tint = FALSE,
  dxdate = NULL,
  fudate = NULL,
  contexpl = NULL,
  outcome = NULL,
  outcomeLevel,
  dod,
  dooc,
  awd,
  awod,
  analysistype = "overall",
  cutp = "12, 36, 60",
  timetypedata = "ymd",
  timetypeoutput = "months",
  uselandmark = FALSE,
  landmark = 3,
  sc = FALSE,
  kmunicate = FALSE,
  ce = FALSE,
  ch = FALSE,
  endplot = 60,
  ybegin_plot = 0,
  yend_plot = 1,
  byplot = 12,
  findcut = FALSE,
  multiple_cutoffs = FALSE,
  num_cutoffs = "two",
  cutoff_method = "quantile",
  min_group_size = 10,
  multievent = FALSE,
  ci95 = FALSE,
  risktable = FALSE,
  censored = FALSE,
  medianline = "none",
  person_time = FALSE,
  time_intervals = "12, 36, 60",
  rate_multiplier = 100,
  rmst_analysis = FALSE,
  rmst_tau = 0,
  residual_diagnostics = FALSE,
  stratified_cox = FALSE,
  strata_variable = NULL,
  loglog = FALSE,
  showExplanations = FALSE,
  showSummaries = FALSE,
  seed = 12345
)

Arguments

data

The data as a data frame.

elapsedtime

The time-to-event or follow-up duration for each patient. Should be numeric and continuous, measured in consistent units (e.g., months or years). Can be calculated automatically from dates if using the date options below.

tint

Enable this option if you want to calculate survival time from dates in your data. This is useful when you have separate columns for diagnosis date and follow-up date and want to calculate the time elapsed between them.

dxdate

The date of diagnosis or study entry. Accepts: (1) Date/datetime text (e.g., "2024-01-15"), (2) Numeric Unix epoch seconds (from DateTime Converter's corrected_datetime_numeric output), (3) Numeric datetime values from R. Time intervals calculated as difference from follow-up date.

fudate

The date of last follow-up or event. Accepts: (1) Date/datetime text (e.g., "2024-01-15"), (2) Numeric Unix epoch seconds (from DateTime Converter's corrected_datetime_numeric output), (3) Numeric datetime values from R. Must be in same format as diagnosis date.

contexpl

The continuous explanatory variable to be used in the analysis.

outcome

The outcome variable to be used in the analysis.

outcomeLevel

The level of the outcome variable that represents the event of interest.

dod

.

dooc

.

awd

.

awod

Select the levels of the outcome variable that correspond to different event types in your data. For example, you might have separate levels for "Dead of Disease" and "Alive w Disease" in a survival analysis of cancer patients.

analysistype

Select the survival estimand. Overall counts both death categories as events; Cause Specific treats other-cause death as censored; Disease-Free counts disease, recurrence, and death categories as events according to the selected mapping. Competing-risk coding is retained for compatibility but is not analysed by this continuous cut-off function; select it to receive guidance to a competing-risk analysis.

cutp

Specify time points for survival-probability estimates. The historical factory value "12, 36, 60" and the word "default" both select unit-aware 1-, 3-, and 5-year points (for example 365, 1095, and 1825 for days, or 1, 3, and 5 for years). Enter any other comma- or space-separated positive values to request custom points in the selected output time scale.

timetypedata

select the time type in data

timetypeoutput

select the time type in output

uselandmark

Enable this option to perform landmark survival analysis at a specified time point.

landmark

Specify the landmark time at which to evaluate survival probabilities in landmark analysis. This option is only available if you enable the "Use Landmark Time" option.

sc

Create Kaplan-Meier curves after the continuous explanatory variable has been divided by a selected single or multiple cut-off method.

kmunicate

Enable this option to create a KMunicate-style survival plot for the continuous explanatory variable.

ce

Enable this option to create a plot of cumulative events over time for the continuous explanatory variable.

ch

Enable this option to create a plot of cumulative hazard over time for the continuous explanatory variable.

endplot

Specify the end time for the survival plots; this is the maximum time point shown on the x-axis. The factory value 60 is interpreted as five years in the selected output time scale (1825 for days, 260 for weeks, 60 for months, 5 for years). Enter any other value to set the axis limit explicitly in that scale.

ybegin_plot

Specify the starting value for the y-axis in the survival plots. This option allows you to customize the range of the y-axis.

yend_plot

Specify the ending value for the y-axis in the survival plots. This option allows you to customize the range of the y-axis.

byplot

Specify the spacing of x-axis tick marks on the survival plots. The factory value 12 is interpreted as one year in the selected output time scale (365 for days, 52 for weeks, 12 for months, 1 for years). Enter any other value to set the spacing explicitly in that scale.

findcut

Derive a single data-dependent cut-off for the continuous explanatory variable using a maximally selected rank statistic. Treat the continuous Cox model as the primary analysis and validate the selected cut-off in independent data.

multiple_cutoffs

Enable this option to derive multiple candidate cut-off points for the continuous explanatory variable. Depending on the selected method, these may be quantile-based, tree-derived, recursively optimized, or minimum-p-value cut-offs. All data-derived cut-offs are exploratory and require external validation.

num_cutoffs

Select the number of cut-off points to identify. This creates ordered marker-value groups (for example, 2 cut-offs create 3 groups); the observed survival ordering must be read from the results and is not assumed in advance.

cutoff_method

Method for finding multiple cut-offs. Quantile-based uses tertiles/quartiles, Recursive finds sequential optimal points, Tree-based uses survival trees, Minimum P-value finds points that minimize log-rank p-values.

min_group_size

Minimum percentage of patients required in each group created by cut-offs. Prevents creating groups with insufficient sample sizes for reliable analysis.

multievent

Enable this option if your data includes multiple event levels (e.g., different types of events or outcomes). This option is required for cause-specific and competing risk survival analyses.

ci95

Enable this option to display 95 percent confidence intervals around the survival estimates in the plots.

risktable

Enable this option to display a table of risk estimates for each group in the survival analysis.

censored

Enable this option to display censored observations in the survival plots.

medianline

If true, displays a line indicating the median survival time on the survival plot.

person_time

Enable this option to calculate and display person-time metrics, including total follow-up time and incidence rates. These metrics help quantify the rate of events per unit of time in your study population.

time_intervals

Specify time intervals for stratified person-time analysis. Enter a comma-separated list of time points to create intervals. For example, "12, 36, 60" will create intervals 0-12, 12-36, 36-60, and 60+.

rate_multiplier

Specify the multiplier for incidence rates (for example, 100 reports events per 100 units of the selected output time scale).

rmst_analysis

Report Restricted Mean Survival Time (RMST) up to a specified horizon. Without a cut-off this is an overall cohort summary; with a single cut-off, the displayed groups are compared descriptively at a common observed horizon.

rmst_tau

Specify the time horizon for RMST calculation. If left as 0, will use 75th percentile of observed survival times. This represents the maximum follow-up time for RMST calculation.

residual_diagnostics

Enable Cox model residual diagnostics including case-level Martingale, Deviance, and Score residuals plus event-time Schoenfeld residuals. These are diagnostic aids; Schoenfeld residuals alone are not a formal proportional-hazards test.

stratified_cox

Enable stratified Cox regression analysis. This allows for different baseline hazards across strata while maintaining proportional hazards within strata.

strata_variable

Categorical variable used for stratification in Cox regression, allowing a different baseline hazard in each stratum. Must be a factor with a small number of reasonably sized groups (for example treating centre or stage). A near-continuous variable would give one stratum per patient and no estimable hazard ratio.

loglog

Enable log-log plot for assessing proportional hazards assumption. Parallel lines in the log-log plot suggest that proportional hazards assumption holds.

showExplanations

Display detailed explanations for each analysis component to help interpret the statistical methods and results.

showSummaries

Display natural language summaries alongside tables and plots. These summaries provide plain-language interpretations of the statistical results. Turn off to reduce visual clutter when summaries are not needed.

seed

Random seed for Monte Carlo components of minimum-p-value and selection-adjusted cut-point calculations.

Value

A results object containing:

results$eventRecodeInfo					a html
results$todo					a html
results$clinicalWarnings					a html
results$errors					a html
results$strongWarnings					a html
results$warnings					a html
results$infoMessages					a html
results$coxRegressionHeading					a preformatted
results$coxSummary					a preformatted
results$coxTable					a table
results$stratifiedCoxTable					a table
results$tCoxtext2					a html
results$coxRegressionHeading3					a preformatted
results$coxRegressionExplanation					a html
results$personTimeHeading					a preformatted
results$personTimeTable					a table
results$personTimeSummary					a html
results$personTimeExplanation					a html
results$rmstHeading					a preformatted
results$rmstTable					a table
results$rmstSummary					a preformatted
results$rmstExplanation					a html
results$residualsTable					a table
results$schoenfeldResidualsTable					a table
results$residualDiagnosticsExplanation					a html
results$cutoffAnalysisHeading					a preformatted
results$rescutTable					a table
results$cutoffAnalysisHeading3					a preformatted
results$cutoffAnalysisExplanation					a html
results$plot4					an image
results$plot5					an image
results$medianSummary					a preformatted
results$medianTable					a table
results$survTableSummary					a preformatted
results$survTable					a table
results$plot2					an image
results$plot3					an image
results$plot6					an image
results$survivalPlotsHeading3					a preformatted
results$survivalPlotsExplanation					a html
results$plot7					an image
results$loglogPlotExplanation					a html
results$residualsPlot					an image
results$calculatedtime					an output
results$outcomeredefined					an output
results$calculatedcutoff					an output
results$multipleCutTable					a table
results$multipleMedianTable					a table
results$multipleCutoffsExplanation					a html
results$multipleSurvTable					a table
results$plotMultipleCutoffs					an image
results$plotMultipleSurvival					an image
results$calculatedmulticut					an output

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$coxTable$asDF

as.data.frame(results$coxTable)