Introduction
DeepPatientLevelPrediction is an R package for building and validating deep learning patient-level predictive models using data in the OMOP Common Data Model format and OHDSI PatientLevelPrediction framework.
Reps JM, Schuemie MJ, Suchard MA, Ryan PB, Rijnbeek PR. Design and implementation of a standardized framework to generate and evaluate patient-level prediction models using observational healthcare data. J Am Med Inform Assoc. 2018;25(8):969-975. doi:10.1093/jamia/ocy032.
Features
- Adds deep learning models to use in the OHDSI PatientLevelPrediction framework.
- Allows users to add custom deep learning models.
- Includes MLP, ResNet, Transformer, and RealMLP models.
- Allows users to use all the features of PatientLevelPrediction to validate and explore their model performance.
Technology
DeepPatientLevelPrediction is an R package. It uses Python PyTorch through reticulate for deep learning model training and inference.
System Requirements
Requires R (version 4.1.0 or higher). Installation from source on Windows may require Rtools. Python 3.10 or newer is required for model training and inference. A CPU can be used for small models; an NVIDIA GPU is recommended for larger deep learning model development.
Getting Started
- To install the package, read the package installation guide.
- Python requirements are declared to
reticulatewhen the package loads and resolved on first Python use. Advanced users can pointRETICULATE_PYTHONat a prebuilt environment for offline or controlled deployments. - Please read the main vignette for the package: Building Deep Learning Models
User Documentation
Documentation can be found on the package website.
PDF versions of the documentation are also available, as mentioned above.
Support
- Developer questions/comments/feedback: OHDSI Forum
- We use the GitHub issue tracker for all bugs/issues/enhancements
Contributing
Read here how you can contribute to this package.