auc_wide: Calculate AUC from a wide-format dataframe

In JMLuther/metabolicR: Functions for analysis of Metabolic Studies

Description

This function calculates the trapezoidal method AUC for data in WIDE format dataframe, with each row encoding an individial subject. The function uses the sfsmisc::integrate.xy() function, which can handle unordered time input. Variables used must follow the naming convention "xxxmeasure_123"

Usage

auc_wide(dfWide, measure, sep = "_", varname = measure,
  use.spline = FALSE, ...)

Arguments

dfWide	a dataframe with repeated measures in "wide" format
measure	Defines the measure to assess, and must be encoded in variable name. Must be quoted. Also defines the outcome variable name, unless 'varname' is defined as something else
sep	Defines the separator between the measure and time. Defaults to "_". Time must be encoded as a number AT THE END of the variable name: ( e.g., "measure_000, measure_001, measure_002...")
varname	Defines the output variable name, to be appended to the prefix "auc_".
use.spline	Whether to use linear trapezoid (use.spline=FALSE), or to integrate under cubic spline (use.spline=TRUE). defaults to FALSE.
...	Passes to the 'integrate.xy' function. Options include "a, b, digits, and xtol"

Value

Modified dataframe appended with a new variable (column) for AUC. The new variable will have named attributes "method", "sourceVars", and "times" which capture the methods used for the calculation

See Also

integrate.xy

Examples

# A very clean dataframe
dfWide1 <- data.frame(
  glucose_0  =c(120, 120, 140, 140, 150),
  glucose_30 =c(250, 250, 270, 325, 400),
  glucose_60 =c(150, 210, 300, 350, 275))
auc_wide(dfWide1, "glucose")

# A messier dataframe: glucose from different study type, time unordered
dfWide5 <- data.frame(
  glucose_ogtt_030=c(250, 250, 270, 325, 400),
  XXXglucose_ogtt_000=c(120, 120, 140, 140, 150),
  glucose_ogtt_060 =c(150, 210, 300, 350, 275),
  glucose_hg_060 =c(150, 210, 300, 350, 275))

auc_wide(dfWide5, "glucose", sep="_ogtt_")  # works

JMLuther/metabolicR documentation built on May 7, 2019, 10:12 a.m.