drsmooth (version 1.9.0)

nlbcd: Non-linearity Below Cut-off Dose

Description

This function tests non-linearity below a specified dose.

Usage

nlbcd (dosecolumn = "", targetcolumn = "", cutoffdose = 0, data = NA)

Arguments

dosecolumn

Name of dose column in input dataframe.

targetcolumn

Name of response column in input dataframe.

cutoffdose

Numeric tested cut-off dose.

data

Input dataframe.

Value

The analysis of variance table comparing the non-linear spline model with the linear model to assess non-linearity across doses below the user-specified cutoff.

Details

The user may provide a limit below which the non-linearity of the dose-response relationship is tested. A significant result indicates that the dose-response relationship exhibits non-linearity below the user-specified cutoff dose. NOTE: The dose-response relationship estimated by this function is not necessarily the same as that estimated by the nlaad function, as the nlbcd only uses doses below the cutoff and nlaad uses all doses. The user should keep this in mind in interpreting the outputs of these functions.

The nlaad, nlbcd, and lbcd functions are currently only intended for use on continuous outcome data.

Examples

# Prints the F test of the difference between the spline model (output "Model 2")
# and the linear model (output "Model 1") as a test of nonlinearity
# for doses below 1.5 (i.e., all dose levels up to and including 1.49):
nlbcd("dose", "MF_Log", cutoffdose=1.5, data=DRdata)

# This produces an error, as no cutoffdose was specified:
## Not run: 
# nlbcd("dose", "MF_Log", data=DRdata)
# ## End(Not run)