rawProcessing: rawProcessing

In sidiwang/hgzips: hgzips

Description

This rawProcessing function finds the actual (N_ij) and expected counts (E_ij) of each AE - drug pair by implementing the methodology described by DuMouchel (1999); This function outputs N and E in both matrix format and vector format.

Usage

rawProcessing(
  rawdata,
  stratify = FALSE,
  frequency_threshold,
  data.squashing = FALSE,
  zeroes = FALSE,
  keep_pts
)

Arguments

rawdata	A data frame containing columns named: primaryid, PT (Adverse Events), prod_ai (drug), strat1 (optional), strat2 (optional), ...... stratx
stratify	A logical scalar specifiying whether stratification is to be used (TRUE) or not (FALSE)
frequency_threshold	the minimum frequency of each AE or drug for it to be kept in the dataset
data.squashing	whether to conduct data squashing or not (TRUE or FALSE)
zeroes	A logical scalar specifying if zero counts should be included.
keep_pts	A vector of whole numbers for the number of points to leave unsquashed for each count (N). See the 'openEBGM' details section.

Details

An primaryid column must be included. Columns must be named as as primaryid, PT, prod_ai, strat1, strat2, ... stratx. Only variables containing 'strat' (case sensitive) will be used as stratification variables.

Value

a list including the following:

• N_ij actual counts of each AE - drug pair in matrix format

• Nij actual counts of each AE - drug pair in vector format

• E_ij expected counts of each AE - drug pair in matrix format

• Eij expected counts of each AE - drug pair in vector format

• processedData dataset after deleting AE or drug with frequencies less than frequency_threshold

• drugList a vector of drug names that are kept in processedData

• AElist a vector of adverse event names that are kept in processedData

• N_ij_squashed a list of squashed N - E - weight dataset for each drug

• N_ij_list a list of original N - E - weight (= 1) dataset for each drug

• squashed squashed processedData dataset

References

DuMouchel W (1999). "Bayesian Data Mining in Large Frequency Tables, With an Application to the FDA Spontaneous Reporting System." The American Statistician, 53(3), 177-190. openEBGM R - package

See Also

openEBGM

Examples

primaryid = c(1:10)
PT = c("fatigue", "restlessness", "sleepdisorder", "fatigue", "anaemia", "blood creatinine increased", "eosinophilia", "generalised oedema", "hypertension", "nephropathy toxic")
prod_ai = c("peginterferon alfa-2a", "peginterferon alfa-2a", "peginterferon alfa-2a", "pantoprazole", "ribavirin", "cyclosporine", "cyclosporine", "cyclosporine", "ribavirin", "flunitrazepam")
strat1 = c("M", "M", "M", "F", "F", "M", "M", "M", "F", "F")
strat2 = c("Young", "Young", "Young", "Old", "Old", "Old", "Old", "Young", "Young", "Young")
dat = data.frame(primaryid = primaryid, PT = PT, prod_ai = prod_ai, strat1 = strat1, strat2 = strat2)
result = hgzips::rawProcessing(dat, stratify = TRUE, frequency_threshold = 0, data.squashing = FALSE, zeroes = TRUE, keep_pts = 190000)

sidiwang/hgzips documentation built on Jan. 19, 2021, 4:09 p.m.