Methods library

We are developing open-source tools for large-scale analytics. This includes methods for population-level estimation and patient-level prediction. Our population-level estimation workgroup is focused on developing open-source software for safety surveillance and comparative effectiveness. This will be achieved through large-scale implementations of traditional observational study designs, including cohort, case-control, self-controlled case series, and self-controlled cohort. This group is also designing and implementing other orthogonal analyses to support causal inference, informed by Hill’s causal viewpoints as presented as the proof-of-concept tool HOMER at the 2013 OMOP symposium.

To learn more about the ongoing development of these tools, check out the following links to our methods tutorials and GitHub repositories:

Population-Level Estimation

An R package for performing new-user cohort studies in an observational database in the OMOP Common Data Model.

MSCCS R package

Method to estimate risk by comparing time exposed with time unexposed among the exposed cohort

This R package is an implementation of the IC Temporal Pattern Discovery method to estimate risk by combining a self-controlled and cohort design. It is designed to run against observational databases in the OMOP Common Data Model.

This package is for performing case-control studies with options to match on age, gender, visit data, provider, and length of observation, as well as adjusting for many covariates.

This package is for performing case-crossover studies with options to adjust for time-trends in exposure (case-time-control), and specifying multiple control windows.


Patient-Level Prediction

An R package for building patient level predictive models using data in Common Data Model format.


  • Takes a cohort and outcome of interest as input.
  • Extracts the necessary data from a database in OMOP Common Data Model format.
  • Uses a large set of covariates including for example all drugs, diagnoses, procedures, as well as age, comorbidity indexes, etc.
  • Large scale regularized regression to fit the predictive models.
  • Includes function for evaluating predictive models.
  • Supported outcome models are logistic, Poisson, and survival (time to event).