predict.netdose {netdose} | R Documentation |
Predicted values for dose-response network meta-analysis
Description
This function provides the predicted values based on the results of dose-response network meta-analysis.
Usage
## S3 method for class 'netdose'
predict(object, agent1 = NULL, dose1 = NULL, agent2 = NULL, dose2 = NULL, ...)
Arguments
object |
An object of class netdose (mandatory). |
agent1 |
An optional character string specifying the first agent to be used for the prediction. By default, all agents are used. |
dose1 |
An optional numeric vector specifying custom doses for the prediction. By default, the doses are set to the common observed doses as defined in the data. |
agent2 |
An optional character string specifying the second agent to be used for the prediction. By default, the reference agent is used. |
dose2 |
An optional numeric vector specifying the dose for the second agent. By default the common dose of the second agent is used. |
... |
Additional arguments (ignored). |
Details
The predict.netdose function calculates predicted effects for specified doses of one or more agents, based on a dose-response network meta-analysis. It supports both linear and non-linear dose-response relationships, accommodating various modeling methods including linear, exponential, fractional polynomials, restricted cubic splines (RCS), and quadratic relationships. This function is particularly useful for exploring comparative effectiveness at specific dose levels of the agents, facilitating the interpretation of complex dose-response relationships in a network meta-analysis setting. By allowing predictions for multiple combinations of agents and doses, it offers flexibility in evaluating hypothetical scenarios or estimating effects for doses outside the directly observed range (where extrapolation is appropriate).
Value
A data frame with additional class predict.netdose
containing the
following variables:
agent1 , dose1 , agent2 , dose2 |
As defined above |
pred |
A numeric vector with the predicted effects |
se.pred |
A numeric vector with standard errors of the predicted effects |
lower |
A numeric vector specifying the lower bounds of the predicted values |
upper |
A numeric vector specifying the upper bounds of the predicted values |
Author(s)
Maria Petropoulou <maria.petropoulou@uniklinik-freiburg.de>, Guido Schwarzer <guido.schwarzer@uniklinik-freiburg.de>
References
Petropoulou et al. (2025): Network meta-analysis with dose-response relationships.
Examples
# Load anesthesia data
data(anesthesia)
# Use a subset of 5 studies
anesthesia_subset <- subset(anesthesia, study %in% unique(anesthesia$study)[1:5])
# Prepare data for DR-NMA
dat <- meta::pairwise(
agent = list(agent1, agent2, agent3),
event = list(event1, event2, event3),
n = list(n1, n2, n3),
dose = list(dose1, dose2, dose3),
data = anesthesia_subset,
studlab = study,
append = FALSE,
sm = "RR"
)
# Perform DR-NMA with a linear dose-response function
dr1 <- netdose(
TE, seTE, agent1, dose1, agent2,
dose2, studlab,
data = dat
)
# Predicted values
pred1 <- predict(dr1)