Package 'ofhsyn'

Generate Synthetic OFH Cohort Datasets

Description

Generate linked synthetic health datasets for a configurable cohort.

Usage

generate_ofh_cohort(
  n = 5000,
  seed = 42,
  icd10 = NULL,
  icd10_file = NULL,
  opcs4 = NULL,
  opcs4_file = NULL,
  bnf_codes = NULL,
  bnf_codes_file = NULL,
  proportions = NULL,
  record_multipliers = NULL,
  code_config = NULL,
  save_csv = TRUE,
  return_objects = TRUE,
  output_dir = NULL
)

Arguments

n	Total synthetic cohort size.
seed	Random seed.
icd10	Optional named character vector of ICD-10 descriptions.
icd10_file	Optional path to a TXT/CSV file containing ICD-10 code and description pairs. TXT format should be tab-separated with code and description columns. CSV format should provide code and description columns.
opcs4	Optional named character vector of OPCS-4 descriptions.
opcs4_file	Optional path to a TXT/CSV file containing OPCS-4 code and description pairs. TXT format should be tab-separated with code and description columns. CSV format should provide code and description columns.
bnf_codes	Optional BNF input for primary care meds. Can be either a character vector of BNF codes or a data frame with columns for code, name, and formulation (optional strength).
bnf_codes_file	Optional path to a TXT/CSV file for BNF input. TXT supports one BNF code per line. CSV supports either code-only or structured medication rows containing code, name, and formulation (optional strength).
proportions	Optional named list of dataset-level coverage proportions. Names should match names(ofh_default_proportions()).
record_multipliers	Optional named list of multipliers for multi-record datasets. Names should match names(ofh_default_record_multipliers()).
code_config	Optional nested list overriding field-level code generation probabilities and pools. Structure should follow ofh_default_code_config().
save_csv	Whether to write CSV outputs to disk.
return_objects	Whether to return generated data frames as an R object.
output_dir	Output directory when save_csv = TRUE.

Value

Named list of generated data frames when return_objects = TRUE; otherwise invisible NULL.

Acknowledgement

We extend our thanks to GitHub user @icallumwebb for contributing a bug fix that improved custom code handling.

Examples

out <- generate_ofh_cohort(n = 200, seed = 123, save_csv = FALSE, return_objects = TRUE)
names(out)

---

Standalone Synthetic Generation Primitives

Description

Utility functions for generating participant populations and event-level synthetic records.

Usage

generate_ofh_population(n = 1000, seed = 123)

add_inpatient_events(
  data,
  events_per_person = 5,
  icd10_codes = c("I210", "I500", "I639", "E110", "J440"),
  opcs4_codes = c("K401", "K451", "K561", "M011", "E033"),
  seed = 123
)

synthesize_drug_exposure(
  data,
  drug_list = c("0212000B0", "0601023A0"),
  seed = 123,
  mean_items_per_person = 2
)

Arguments

data	Input data frame containing a pid column.
n	Number of participants.
seed	Random seed.
events_per_person	Mean events per participant.
icd10_codes	ICD-10 code pool.
opcs4_codes	OPCS-4 code pool.
drug_list	Medication code pool.
mean_items_per_person	Mean prescription items per participant.

Value

Return value depends on the function called:

generate_ofh_population()

Data frame with one row per participant and columns including pid, sex, and birth_year.

add_inpatient_events()

Data frame of synthetic inpatient events with columns pid, admidate, icd10, and opcs4.

synthesize_drug_exposure()

Data frame of synthetic primary-care medication records with participant IDs and prescribing/dispensing fields (for example prescribedbnfcode, paidbnfcode).

---

Configuration Helpers for OFH Generation

Description

Helper functions that return default settings and compose full generation configuration lists.

Usage

ofh_default_proportions()
ofh_default_record_multipliers()
ofh_default_code_config()
ofh_build_config(
  n = 5000,
  proportions = ofh_default_proportions(),
  record_multipliers = ofh_default_record_multipliers(),
  code_config = list()
)

Arguments

n	Total cohort size.
proportions	Dataset proportions list.
record_multipliers	Record multiplier list for event datasets.
code_config	Optional code configuration overrides.

Value

Return value depends on the function called:

ofh_default_proportions()

Named numeric list of dataset proportions in [0, 1].

ofh_default_record_multipliers()

Named numeric list of multipliers for multi-record datasets.

ofh_default_code_config()

Nested named list containing default code pools, weights, and generation controls by dataset.

ofh_build_config()

Named list with total_pid_count (integer), datasets (nested list of dataset sizing settings), and code_config (merged code configuration list).

---

Reference Class for OFH Cohort Generation

Description

Reference class API for configuring and running synthetic cohort generation.

Usage

OFHCohortSynthesizer

Details

Create an instance with OFHCohortSynthesizer$new(...) and run generation via $run_all(n = ...).

Value

A ReferenceClass generator object. Use OFHCohortSynthesizer$new(...) to create an instance. Instance methods return the instance invisibly for chaining where applicable, and $run_all() returns a named list of data frames when return_objects = TRUE (otherwise invisible NULL).

Examples

syn <- OFHCohortSynthesizer$new(project_root = ".", seed = 123)
out <- syn$run_all(n = 100, save_csv = FALSE, return_objects = TRUE)

---