nri_cox function

Net Reclassification Index for Cox Regression Models

nri_cox Net Reclassification Index for Cox Regression Models

nri_cox(data, formula0, formula1, t_risk, cutoff, B = FALSE, nboot = 10)

Arguments

• data: Data frame with relevant predictors
• formula0: A formula object to specify the reference model as normally used by glm. See under "Details" and "Examples" how these can be specified.
• formula1: A formula object to specify the new model as normally used by glm.
• t_risk: Follow-up value to calculate cases, controls. See details.
• cutoff: A numerical vector that defines the outcome probability cutoff values.
• B: A logical scalar. If TRUE bootstrap confidence intervals are calculated, if FALSE only the NRI estimates are reported.
• nboot: A numerical scalar. Number of bootstrap samples to derive the percentile bootstrap confidence intervals. Default is 10.

Returns

An object from which the following objects can be extracted:

• data dataset.
• prob_orig outcome risk probabilities at t_risk for reference model.
• prob_new outcome risk probabilities at t_risk for new model.
• time name of time variable.
• status name of status variable.
• cutoff cutoff value for survival probability.
• t_risk follow-up time used to calculate outcome (risk) probabilities.
• reclass_totals table with total reclassification numbers.
• reclass_cases table with reclassification numbers for cases.
• reclass_controls table with reclassification numbers for controls.
• totals totals of controls, cases, censored cases.
• km_est totals of cases calculated using Kaplan-Meiers risk estimates.
• nri_est reclassification measures.

Details

A typical formula object has the form Outcome ~ terms. Categorical variables has to be defined as Outcome ~ factor(variable), restricted cubic spline variables as Outcome ~ rcs(variable, 3). Interaction terms can be defined as Outcome ~ variable1*variable2 or Outcome ~ variable1 + variable2 + variable1:variable2. All variables in the terms part have to be separated by a "+". If a formula object is used set predictors, cat.predictors, spline.predictors or int.predictors at the default value of NULL.

Follow-up for which cases nd controls are determined. For censored cases before this follow-up the expected risk of being a case is calculated by using the Kaplan-Meier value to calculate the expected number of cases.These expected numbers are used to calculate the NRI proportions but are not shown by function nricens.

Examples

library(survival)
  lbpmicox1 <- subset(psfmi::lbpmicox, Impnr==1) # extract one dataset
  risk_est <- nri_cox(data=lbpmicox1, formula0 = Surv(Time, Status) ~ Duration + Pain,
                   formula1 = Surv(Time, Status) ~ Duration + Pain + Function + Radiation,
                   t_risk = 80, cutoff=c(0.45), B=TRUE, nboot=10)

References

Cook NR, Ridker PM. Advances in measuring the effect of individual predictors of cardiovascular risk: the role of reclassification measures. Ann Intern Med. 2009;150(11):795-802.

Steyerberg EW, Pencina MJ. Reclassification calculations for persons with incomplete follow-up. Ann Intern Med. 2010;152(3):195-6; author reply 196-7.

Pencina MJ, D'Agostino RB Sr, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30(1):11-21

Inoue E (2018). nricens: NRI for Risk Prediction Models with Time to Event and Binary Response Data. R package version 1.6, https://CRAN.R-project.org/package=nricens.

Author(s)

Martijn Heymans, 2023