uplift: "Uplift" from IER and CER

In joelkuiper/personograph: Pictographic Representation of Treatment Effects

Description

Calculates the percentage (from 0 to 1) of people who have an intervention benefit, intervention harm, bad outcome regardless, and good outcome regardless from the Intervention Event Rates (IER) and Control Event Rates (CER). Note that the result depends on the direction of the outcome measure, e.g. higher_is_better = T (default) for intervention efficacy, higher_is_better = F for adverse events.

Usage

uplift(ier, cer, higher_is_better = NULL)

Arguments

ier	Intervention Event Rates
cer	Control Event Rates
higher_is_better	logical indicating the direction of the outcome measure, default TRUE

Details

The adopted terminology is similar to that of Uplift modelling https://en.wikipedia.org/wiki/Uplift_modelling

Value

A list of S3 class personograph.uplift with the following elements:

• good outcome people who have a good outcome regardless of intervention

• bad outcome people who have a bad outcome regradless of intervention

• intervention benefit people who benefit from intervention

• intervention harm people who are harmed by intervention

Can be plotted as a personograph with the S3 generic plot.

Examples

ier <- 0.06368133
cer <- 0.1115242
u <- uplift(ier, cer, higher_is_better=TRUE)
plot(u)

joelkuiper/personograph documentation built on May 19, 2019, 3 p.m.