dichot_smooth: Dose-response Modeling with Smoothing Splines

Description

Generates a spline model given dose and target response columns.

Usage

dichot_smooth(dosecolumn = "", targetcolumn = "", k = 4, return_predict = FALSE, write_predict = TRUE, STD_bias = TRUE, data = NA)

Arguments

dosecolumn

Name of dose column of interest in dataframe.

targetcolumn

Name of response column of interest in dataframe.

Dimension of the basis used to represent the smooth term; see Details.

return_predict

If TRUE (default FALSE), returns dataframe of predicted values.

write_predict

If TRUE (the default), writes the dataframe of predicted values to a .csv file in the working directory.

STD_bias

If TRUE (the default), calculates the slope transition dose, a bootstrapped and resource-intensive computation.

data

Input dataframe.

Value

A plot of the spline-estimated dose-response function along with the actual data.

Details

This function generates a spline model with the input dose and target response columns, plots the spline-estimated dose-response function with its upper and lower 95 percent confidence bounds in green and red respectively along with the actual data, and returns key metrics related to the dose-response function. Note that the confidence bounds depicted on the plot are for the dose-response function itself, and not for the raw data.

The parameter 'k', defaulted to 4, defines the number of dimensions the spline function will use in estimating the response relation. With 2 reserved for each end of the smooth, the default allows for 2 bends in the smooth. In the case that this appears to overfit the data, the user may choose to override the default to 3, which would allow only one bend.