aouFI

R package of helper functions for the Roux OHDSI Frailty project

Installation

remotes::install_github("roux-ohdsi/aouFI")

Functions

• getEligible() retrieves a single column dataframe of person_ids from the AoU database (when provided with a database connection) that are > 50 years old and < 120 years old (defaults).
• getConcepts() retrieves index concept_ids and names
• omop2fi() returns a dataframe of frailty category occurrences when given a database connection to any (in theory) OMOP CDM database and a table of person_id values (e.g., from getEligible()), and a frailty index (one of efi, efragicap, vafi, or hfrs). In theory, this data can be used with getFI() though this is not yet tested…
• getFI() returns frailty scores for a several FIs given a dataframe of person_id’s, start, and end dates. It requires a specific data input format and isn’t terribly flexible yet.

Data

FI Tables

These tables hold concept_ids and concept names for each of the frailty indices. They can be queried using getConcepts(). Used by the omop2fi() function.

• efi_concepts.rda
• efragicap_concepts.rda
• hfrs_concepts.rda
• vafi_concepts.rda

Data generating scripts are either in data-raw on on the AoU workbench…

Other useful tables

• survey_names List of surveys given in All of US
• FI_labels Paper-ready names for AouFI variables with a linking column for derived data
• simulated_ehr Simulated ehr data to search - similar to what would be returned by omop2fi(). This will be used for testing and demo purposes. It can be used with getFI(). The values in this data are entirely fabricated based on code in /data-raw. Next step is to add a third dataset with occurrences that don’t fall within any of the FIs as well to be able to test the omop2fi function.
• simulated_cohort Simulated cohort with a starting search date. Can be used with simulated_ehr to use the getFI() function. It contains a sample of fake participants in simulated_ehr with randomly generated start dates in the observation period for each person_id.