An R package CPRDDrugPrep

Belay Birlie Yimer, David Selby, ...

An algorithm for the transparent and efficient preparation of CPRD drug data into information on individuals’ drug use over time. The goal of CPRDDrugPrep package is to allow users to create multiverse analyses in a concise and easily interpretable manner. The CPRDDrugPrep package allows reserchers to specify sets of defensible data processing options at each decison node (e.g., different ways of imputing missing quantity and ndd, different ways of handling multiple prescriptions), implement them all, and then report the outcomes of all analyses resulting from all possible choice combinations. The package depends on the R-package doseminer for extracting drug dosage information from CPRD prescription data.

Installation

You can install the latest development version from GitHub with these R commands:

install.packages("devtools")
devtools::install_github("belayb/CPRDDrugPrep")

Examples

Data

DrugPrep has been developed to process prescriptions data from the Clinical Practice Research Datalink (CPRD). You will need a dataset containing the following variables for the drug types (prodcodes) you are interested in:

Variable description	Name in script	Name on CPRD	Where located in CPRD
Patient identifier	patid	patid	Therapy file
Product identifier	prodcode	prodcode	Therapy file
Start date of prescription	start	start	Therapy file
Quantity	qty	qty	Therapy file
Numeric daily dose	ndd	ndd	Therapy file or from result of doseminer call
Number of days of treatment prescribed	numdays	numdays	Therapy file
Dose duration	dose_duration	dose_duration	common_dosages file
Maximum and minimum length of prescriptions		NA	Not in CPRD: self-defined
Maximum and minimum numeric daily dose	max_ndd, min_ndd	NA	Not in CPRD: self-defined
Maximum and minimum quantity	max_qty, min_qty	NA	Not in CPRD: self-defined

Call-to-doseminer and ndd commputation

Immplusible values

Running universe data creation

Running universe data analysis

Running multiverse data creation

Running multiverse analysis

Contributors

Maintained by Belay Birlie Yimer (belaybirlie.yimer@manchester.ac.uk), David Selby (david.selby@manchester.ac.uk), ...