Type:	Package
Title:	Calculate Estimates in Models with Interaction
Version:	0.1.1
Date:	2023-11-21
Description:	A tool to calculate and plot estimates from models in which an interaction between the main predictor and a continuous covariate has been specified. Methods used in the package refer to Harrell Jr FE (2015, ISBN:9783319330396); Durrleman S, Simon R. (1989) <doi:10.1002/sim.4780080504>; Greenland S. (1995) <doi:10.1097/00001648-199507000-00005>.
License:	GPL-3
Depends:	R (≥ 3.6)
Imports:	graphics, stats, utils, survival, msm, rms, boot, pspline, pryr, stringr
Suggests:	knitr, rmarkdown, dplyr, mlbench, grDevices, parallel, methods
VignetteBuilder:	knitr
RoxygenNote:	7.2.3
Encoding:	UTF-8
LazyData:	true
NeedsCompilation:	no
Packaged:	2023-11-21 21:57:55 UTC; gem37
Author:	Giorgio Melloni [aut, cre], Andrea Bellavia [aut], Hong Xiong [aut]
Maintainer:	Giorgio Melloni <melloni.giorgio@gmail.com>
Repository:	CRAN
Date/Publication:	2023-11-21 22:40:02 UTC

Returns the estimates of for an unspecified interaction model

Description

This function is a dispatcher that generate OR, HR or linear estimates values for a simple or restricted cubic spline interaction model from a logistic, Cox or linear regression

Usage

intEST(
  var2values,
  model,
  data,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a dataframe with initial values and OR/HR/linear estimates , if ci = TRUE a dataframe with 5 columns, initial values, OR/HR/linear estimates, lower CI, upper CI and SE

Examples

library(rms)
library(mlbench)
data(PimaIndiansDiabetes)
# Set age on a 5-year scale
PimaIndiansDiabetes$age <- PimaIndiansDiabetes$age/5
# Recode diabetes as 0/1
PimaIndiansDiabetes$diabetes <- ifelse(PimaIndiansDiabetes$diabetes=="pos" , 1 , 0)
# Logistic model predicting diabetes over BMI, age and glucose
myformula <- diabetes ~ mass + age * rcs( glucose , 3 )
model <- lrm(myformula , data = PimaIndiansDiabetes )
intEST( var2values = 20:80
       , model = model , data = PimaIndiansDiabetes , var1 ="age", var2="glucose"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")
# Linear model predicting BMI over diabetes, age and glucose
myformula2 <- mass ~ diabetes + age * rcs( glucose , 3 )
model2 <- glm(myformula2 , data = PimaIndiansDiabetes , family = "gaussian")
intEST( var2values = 20:80
       , model = model2 , data = PimaIndiansDiabetes , var1 ="age", var2="glucose"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")

Linear regression interaction estimates

Description

Generate linear estimates for a 1 unit increase in a variable at specified points of another interacting variable in a linear interaction model

Usage

linLIN(
  var2values,
  model,
  data,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a vector of estimate of length(var2values), if ci = TRUE a dataframe with 5 columns, initial values, linear estimates, lower CI, upper CI and SE

Examples

library(rms)
library(mlbench)
data(PimaIndiansDiabetes)
# Recode diabetes as 0/1
PimaIndiansDiabetes$diabetes <- ifelse(PimaIndiansDiabetes$diabetes=="pos" , 1 , 0)
myformula <- glucose ~ mass + diabetes * age
model <- glm(myformula , data = PimaIndiansDiabetes ,family=gaussian)
# Show the effect on glucose of being diabetic at age 20 to 80
linLIN( var2values = 20:80
       , model = model , data = PimaIndiansDiabetes , var1 ="diabetes", var2="age"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")

Linear interaction HR

Description

Generate HR values for a 1 unit increase in a variable at specified points of another interacting variable in a simple Cox interaction model

Usage

loglinHR(
  var2values,
  model,
  data,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a vector of estimate of length(var2values), if ci = TRUE a dataframe with 5 columns, initial values, HR, lower CI, upper CI and SE

Examples

library(survival)
data(cancer)
myformula <- Surv(time, status) ~ ph.karno + ph.ecog + age*sex
model <- coxph(myformula , data = lung )
loglinHR( var2values = 40:80
                     , model = model , data = lung , var1 ="sex", var2="age"
                     , ci=TRUE , conf = 0.95 , ci.method = "delta")

Linear interaction OR

Description

Generate OR values for a 1 unit increase in a variable at specified points of another interacting variable in a simple logistic interaction model

Usage

loglinOR(
  var2values,
  model,
  data,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a vector of estimate of length(var2values), if ci = TRUE a dataframe with 5 columns, initial values, OR, lower CI, upper CI and SE

Examples

library(rms)
library(mlbench)
data(PimaIndiansDiabetes)
# Set age on a 5-year scale
PimaIndiansDiabetes$age <- PimaIndiansDiabetes$age/5
# Recode diabetes as 0/1
PimaIndiansDiabetes$diabetes <- ifelse(PimaIndiansDiabetes$diabetes=="pos" , 1 , 0)
myformula <- diabetes ~ mass + age * glucose
model <- glm(myformula , data = PimaIndiansDiabetes , family = binomial())
loglinOR( var2values = 20:80
       , model = model , data = PimaIndiansDiabetes , var1 ="age", var2="glucose"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")

Plot the result of HR, OR or linear estimates

Description

Create a spline var2 by 1 unit increase of var1

Usage

plotINT(
  x,
  xlab = "",
  main = "",
  log = FALSE,
  ylab = NULL,
  line1 = TRUE,
  linecolor = "dodgerblue",
  cicolor = "darkgray",
  ...
)

Arguments

Value

simple pspline smoothed splined plot of estimates of 1 unit increase in var1 at var2 values

Examples

library(rms)
library(survival)
data(cancer)
myformula <- Surv(time, status) ~ ph.karno + ph.ecog + rcs(age, 3)*sex
model <- cph(myformula , data = lung )
myHR <- rcsHR( var2values = 40:80
               , model = model , data = lung , var1 ="sex", var2="age"
               , ci=TRUE , conf = 0.95 , ci.method = "delta")
plotINT(myHR , ylab = "HR of male VS female" , xlab = "Age")

Restricted cubic spline interaction HR for more than 3 knots

Description

Generate HR values in a Cox model for a 1 unit increase in a variable at specified points of another interacting variable splined with rcs(df >= 3)

Usage

rcsHR(
  var2values,
  model,
  data = NULL,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a dataframe with initial values and HR , if ci = TRUE a dataframe with 5 columns, initial values, HR, lower CI, upper CI and SE

Examples

library(survival)
library(rms)
data(cancer)
myformula <- Surv(time, status) ~ ph.karno + ph.ecog + rcs(age,4)*sex
model <- cph(myformula , data = lung )
rcsHR( var2values = 40:80
       , model = model , data = lung , var1 ="sex", var2="age"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")

Restricted cubic spline interaction linear estimates for more than 3 knots

Description

Generate estimates in a linear model for a 1 unit increase in a variable at specified points of another interacting variable splined with rcs(df >= 3)

Usage

rcsLIN(
  var2values,
  model,
  data = NULL,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a dataframe with initial values and linear estimates , if ci = TRUE a dataframe with 5 columns, initial values, linear estimates, lower CI, upper CI and SE

Examples

library(rms)
library(mlbench)
data(PimaIndiansDiabetes)
# Recode diabetes as 0/1
PimaIndiansDiabetes$diabetes <- ifelse(PimaIndiansDiabetes$diabetes=="pos" , 1 , 0)
myformula <- glucose ~ mass + diabetes * rcs(age, 4)
model <- glm(myformula , data = PimaIndiansDiabetes , family="gaussian")
# Show the effect on glucose of being diabetic at age 20 to 80
rcsLIN( var2values = 20:80
       , model = model , data = PimaIndiansDiabetes , var1 ="diabetes", var2="age"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")

Restricted cubic spline interaction OR for more than 3 knots

Description

Generate OR values in a logistic model for a 1 unit increase in a variable at specified points of another interacting variable splined with rcs(df >= 3)

Usage

rcsOR(
  var2values,
  model,
  data = NULL,
  var1,
  var2,
  ci = TRUE,
  conf = 0.95,
  ci.method = "delta",
  ci.boot.method = "perc",
  R = 100,
  parallel = "multicore",
  ...
)

Arguments

Value

if ci = FALSE, a dataframe with initial values and OR

Examples

library(rms)
library(mlbench)
data(PimaIndiansDiabetes)
# Set age on a 5-year scale
PimaIndiansDiabetes$age <- PimaIndiansDiabetes$age/5
# Recode diabetes as 0/1
PimaIndiansDiabetes$diabetes <- ifelse(PimaIndiansDiabetes$diabetes=="pos" , 1 , 0)
myformula <- diabetes ~ mass + age * rcs( glucose , 4 )
model <- glm(myformula , data = PimaIndiansDiabetes , family = "binomial")
rcsOR( var2values = 20:80
       , model = model , data = PimaIndiansDiabetes , var1 ="age", var2="glucose"
       , ci=TRUE , conf = 0.95 , ci.method = "delta")

UMARU IMPACT Study data

Description

A subset of data from the University of Massachusets Aids Research Unit (UMARU) IMPACT study.

Usage

umaru

Format

A data frame with 575 rows and 10 variables

X

observation count

id

identification code

age

Age in years

beckscore

Beck Depression Score at admission

ndrugtx

Number of prior drug treatments

treat

treat

site

site

los

los

time

time to event

censor

censor event

heroin

heroin use at admission

cocaine

cocaine use at admission

drugtx

Prior Drug treatment, yes/no

ivdrug

Prior IV drug treatment

nonwhite

1 if non white, 0 otherwise

Source

ftp://ftp.wiley.com/public/sci_tech_med/logistic