Building a mock from data"

In omock: Creation of Mock Observational Medical Outcomes Partnership Common Data Model

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Introduction

Most of the functionalities of omock are used to build specific mock tables (e.g. mockPerson(), mockObservationPeriod(), ...), this allows the user to create mock cdm objects combining all those functions with some room for customisation. There are times where the user will want to create a mock CDM reference from its own bespoke tables. The mockCdmFromTables() function is designed to facilitates the creation of mock CDM reference from bespoke tables. This functionality will be useful to create a mock CDM from a cohort_table or a drug_exposure table, or with incomplete data (e.g. missing columns).

library(omock)
library(dplyr, warn.conflicts = FALSE)
library(PatientProfiles)

Create a mock cdm from a cohort table

For example if you want to create a CDM reference based on below bespoke cohorts. You can do it simple using the mockCdmFromTable() functions in a few lines of code.

# Define a list of user-defined cohort tables
cohortTables <- list(
  cohort1 = tibble(
    subject_id = 1:10L,
    cohort_definition_id = rep(1L, 10),
    cohort_start_date = as.Date("2020-01-01") + 1:10,
    cohort_end_date = as.Date("2020-01-01") + 11:20
  ),
  cohort2 = tibble(
    subject_id = 11:20L,
    cohort_definition_id = rep(2L, 10),
    cohort_start_date = as.Date("2020-02-01") + 1:10,
    cohort_end_date = as.Date("2020-02-01") + 11:20
  )
)

# Create a mock CDM object from the user-defined tables
cdm <- mockCdmFromTables(tables = cohortTables)

cdm

The generated CDM object will build the person, observation_period and vocabulary tables so that all the cohorts are in observation:

cdm$cohort1 |>
  addInObservation()
cdm$observation_period

Create a mock CDM from drug_exposure

Now we will create a CDM around a drug_exposure table, this functionality is quite useful to obtain mock datasets for testing purposes only specifying part of the information. In this case we will partially define person table to impose all individuals are women:

person <- tibble(person_id = 1:5L, gender_concept_id = 8532L, year_of_birth = 1992)

and we will also create the records of the drug_exposure table:

drugExposure <- tibble(
  person_id = rep(1:5L, 2),
  drug_concept_id = 19073188L,
  drug_exposure_start_date = rep(as.Date(c("2000-01-01", "2000-06-1")), each = 5),
  drug_exposure_end_date = drug_exposure_start_date + c(10L, 20L, 100L, 140L, 30L, 50L, 30L, 20L, 45L, 35L)
)

Then mockCdmFromTables() will populate the missing columns with interpolated data and add all the tables necessary to create a minimum viable CDM (it will contain at least person, observation_period and the vocabulary tables):

cdm <- mockCdmFromTables(tables = list(person = person, drug_exposure = drugExposure))

cdm

As before all the records of drug_exposure will be in observation:

cdm$drug_exposure |>
  addInObservation() |>
  group_by(in_observation) |>
  tally()

omock documentation built on Nov. 5, 2025, 6:31 p.m.