graphPAF (version 2.0.0)

plot_continuous: Plot hazard ratios, odds ratios or risk ratios comparing differing values of a continuous exposure to a reference level

Description

Plot hazard ratios, odds ratios or risk ratios comparing differing values of a continuous exposure to a reference level

Usage

plot_continuous(
  model,
  riskfactor,
  data,
  S = 10,
  ref_val = NA,
  ci_level = 0.95,
  min_risk_q = 0.1,
  plot_region = TRUE,
  plot_density = TRUE,
  n_x = 10000,
  theylab = "OR",
  qlist = seq(from = 0.001, to = 0.999, by = 0.001),
  interact = FALSE
)

Value

A ggplot2 plotting object

Arguments

model

A fitted model (either glm, clogit or coxph)

riskfactor

The string name of a continuous exposure or risk factor represented in the data and model

data

Data frame used to fit the model

S

Default 10. The integer number of random samples used to calculate average differences in linear predictors. Only relevant to set when interact=TRUE

ref_val

The reference value used in plotting. If left at NA, the median value of the risk factor is used

ci_level

Numeric. A number between 0 and 1 specifying the confidence level

min_risk_q

Default .1. A number between 0 and 1 representing the desired risk quantile for the continuous risk factor

plot_region

Default TRUE. Logical specifying whether the targeted region corresponding to an intervention setting the continuous risk factor at a quantile min_risk_q or lower is to be plotted

plot_density

Default TRUE. Logical specifying whether density of distribution of risk factor is to be added to the plot

n_x

Default 10000. How many values of riskfactor will be used to plot spline (when interact=FALSE)

theylab

Default "OR". Y-axis label of the plot

qlist

Vector of quantile values for q, corresponding to the plotted values of PAF_q for each risk factor/exposure

interact

Default "FALSE". Set to TRUE spline models enter as interactions.

References

Ferguson, J., Maturo, F., Yusuf, S. and O’Donnell, M., 2020. Population attributable fractions for continuously distributed exposures. Epidemiologic Methods, 9(1)

Examples

library(survival)
library(splines)
model_continuous <- glm(formula = case ~ region * ns(age, df = 5) +
 sex * ns(age, df = 5) + education +exercise + ns(diet, df = 3) +
 alcohol + stress + ns(lipids,df = 3) + ns(waist_hip_ratio, df = 3) +
  high_blood_pressure, family = "binomial", data = stroke_reduced)
plot_continuous(model_continuous,riskfactor="diet",data=stroke_reduced)