https://github.com/clevelandclinicqhs/qhscrnomo

An R package to construct nomograms from competing risks regression models

https://github.com/clevelandclinicqhs/qhscrnomo

Repository

An R package to construct nomograms from competing risks regression models

README.Rmd

---
output: github_document
---



```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
```

# QHScrnomo 


[![Lifecycle: stable](https://img.shields.io/badge/lifecycle-stable-brightgreen.svg)](https://lifecycle.r-lib.org/articles/stages.html#stable)
[![CRAN status](https://www.r-pkg.org/badges/version/QHScrnomo)](https://CRAN.R-project.org/package=QHScrnomo)
![CRAN\_Download\_Counter](http://cranlogs.r-pkg.org/badges/grand-total/QHScrnomo)


[Nomograms](https://en.wikipedia.org/wiki/Nomogram) serve as practical, useful tools and communication devices in the context of clinical decision making that enable clinicians to quickly understand and gauge individual patients' risk of outcomes from (potentially) complex statistical models. The goal of `QHScrnomo` is to provide functionality to construct [nomograms](https://rdrr.io/cran/rms/man/nomogram.html) in the context of time-to-event (survival) analysis in the presence of competing risks. It also contains functions to build, validate, and summarize these models.

## Installation

You can install the development version of QHScrnomo from [GitHub](https://github.com/) with:

``` r
devtools::install_github("ClevelandClinicQHS/QHScrnomo")
```
Or from CRAN:

``` r
install.packages("QHScrnomo")
```
### Dependencies

This package has its most prominent dependencies on the [`rms`](https://CRAN.R-project.org/package=rms) package. In fact, it actually _Depends_ on it (see `DESCRIPTION`), so that package will load with `QHScrnomo`. It also makes heavy usage of [`cmprsk`](https://CRAN.R-project.org/package=cmprsk) and [`Hmisc`](https://CRAN.R-project.org/package=Hmisc) (which comes with `rms`). All methodology implemented here comes from these packages, so they should serve as a resource to further understand what is happening behind the scenes of `QHScrnomo`.

## Example

The following is an example of how to construct a nomogram from a competing risks regression model. First, we'll load the package.

```{r example}
library(QHScrnomo)
```

### 1. Fit the regression model

Start by fitting a Cox proportional-hazards model.

```{r}
# Register the data set
dd <- datadist(prostate.dat)
options(datadist = "dd")

# Fit the Cox-PH model for prostate cancer-specific mortality
prostate.f <- cph(Surv(TIME_EVENT,EVENT_DOD == 1) ~ TX  + rcs(PSA,3) +
           BX_GLSN_CAT + CLIN_STG + rcs(AGE,3) +
           RACE_AA, data = prostate.dat,
           x = TRUE, y= TRUE, surv=TRUE, time.inc = 144)
```

Then convert (adjust) it to account for the presence of competing risks.

```{r}
# Refit to a competing risks regression to account for death from other causes
prostate.crr <- crr.fit(prostate.f, cencode = 0, failcode = 1)
anova(prostate.crr)
```

### 2. Validate model output

We can generate cross-validated risk predictions at a particular time horizon of interest.

```{r}
# Generate the cross-validated probability of the event of interest
set.seed(123)
prostate.dat$preds.tenf <- tenf.crr(prostate.crr, time = 120, trace = FALSE) # 120 = 10 years
str(prostate.dat$preds.tenf)
```

And then check the discrimination of those probabilities via the _concordance index_.

```{r}
with(prostate.dat, cindex(preds.tenf, EVENT_DOD, TIME_EVENT, type = "crr"))["cindex"]
```

### 3. Construct the nomogram

Finally, we can build the nomogram that can be used to quickly generate model predictions manually.

```{r nomogram, fig.width=7, fig.height=6}
# Set some nice display labels (also see ?Newlevels)
prostate.g <-
  Newlabels(
    fit = prostate.crr,
    labels = 
      c(
        TX = "Treatment options",
        PSA = "PSA (ng/mL)",
        BX_GLSN_CAT = "Biopsy Gleason Score Sum",
        CLIN_STG = "Clinical Stage",
        AGE = "Age (Years)",
        RACE_AA = "Race"
      )
  )

# Construct the nomogram
nomogram.crr(
  fit = prostate.g,
  failtime = 120,
  lp = FALSE,
  xfrac = 0.65,
  fun.at = seq(0.2, 0.45, 0.05),
  funlabel = "Predicted 10-year risk"
)
```

Owner

GitHub Events

Committers

Last synced: over 2 years ago

Issues and Pull Requests

Last synced: over 2 years ago

Packages