This library was developed in order to facilitate rapid prototyping in Python of predictive machine-learning models using longitudinal medical data from an OMOP CDM-standard database. omop-learn supports the easy definition of predictive clinical tasks, featurizations of OMOP data, and cohorts of relevance. We further provide methods using sparse tensor implementations to rapidly manipulate the collected features in the rawest form possible, allowing for dynamic transformations of the data.
Two machine-learning models are included with the library. First, a windowed linear model, which uses various backwards-facing windows to aggregate features over different timescales, then feeds these features into a regularized logistic regression model. This model was inspired by the work of Razavian et. al, and despite its simplicity is often competitive with state-of-the-art algorithms. We also include SARD (Self-Attention with Reverse Distillation), a novel deep-learning algorithm that uses self-attention to allow medical events to contextualize themselves using other events in a patient's timeline. SARD also makes use of reverse distillation, a training technique we introduce that effectively initializes a deep model using a high-performing linear proxy, in this case the windowed linear model described above -- for the details of this method and the SARD architecture, we refer to our paper (to be attached here soon!).
For a more detailed summary of omop-learn's data collection pipeline, and for documentation of functions, please refer to the GitHub Pages link associated with this repo. This link also describes the process of creating one's own cohorts, predictive tasks, and features.
The following libraries are necessary to run omop-learn:
- numpy
- sqlalchemy
- pandas
- torch
- sklearn
- matplotlib
- ipywidgets
- IPython.display
- gensim.models
- scipy.sparse
- sparse
Note that sparse is the PyData Sparse library, documented here
We provide two example notebooks, End of Life Linear Model Example.ipynb and End of Life Deep Model Example which both use an example task of predicting mortality over a six-month window for patients over the age of 70. The two notebooks run the windowed linear and deep SARD models respectively -- note that your machine must be able to access a GPU in order to run the deep models.
To run the models, first set up the file config.py with connection information for your Postgres server containing an OMOP CDM database. Then, simply run through the cells of the notebook in order. Further documentation of the exact steps taken to define a task, collect data, and run a predictive model are embedded within the notebooks.
Rohan Kodialam, Rebecca Boiarsky, Jake Marcus, Ike Lage, Shannon Hwang