calibrate_risk: Calibrate Acute Event Prediction

In evanjflack/polypharmacy: Behavioral Hazard and Patient Decision Making

Description

calibrate_risk() is a function that calculates the actual event rate of acute events by predicted risk quantile and initial period treatment status.

Usage

calibrate_risk(DT, keep_vars, cont_risk_var, n_quant, event_var, treated_var)

Arguments

DT	data.table with keep_vars, cont_risk_var, event_var, and treated_var
keep_vars	character vector, names of subsetting variables (e.g. "keep_jan")
cont_risk_var	character, name of contrinuous risk prediction variable (e.g. "comp_card_untreated_statin")
n_quant	interger, number of quantiles to break cont_risk_var in to
event_var	character, name of the realized event variable (e.g. "ami")
treated_var	character, name of the binary treatment indicator variable

Value

data.table with following columns:

treated	indicator for initial period treatment
risk	quantile of predicted risk
mean	realized event rate

evanjflack/polypharmacy documentation built on Oct. 12, 2020, 10:45 a.m.