Development Status: As of 08/02/2020, PyHealth is under active development and in its alpha stage. Please follow, star, and fork to get the latest functions!

PyHealth is a comprehensive and flexible Python library for healthcare AI, designed for both ML researchers and medical practitioners. The library is proudly developed and maintained by researchers at Carnegie Mellon University, IQVIA, and University of Illinois at Urbana-Champaign. PyHealth makes many important healthcare tasks become accessible, such as phenotyping prediction, mortality prediction, ICU length stay forecasting, etc. Running these prediction tasks with deep learning models can be as short as 10 lines of code.

PyHealth comes with three major modules: (i) data preprocessing module; (ii) learning module and (iii) evaluation module. Typically, one can run the data prep module to prepare the data, then feed to the learning module for prediction, and finally assess the result with the evaluation module. Users can use the full system as mentioned or just selected modules based on the own need:

• Deep learning researchers may directly use the processed data along with the proposed new models.
• Medical personnel, may leverage our data preprocessing module to convert the medical data to the format that learning models could digest, and then perform the inference tasks to get insights from the data.

PyHealth is featured for:

• Unified APIs, detailed documentation, and interactive examples across various datasets and algorithms.
• Advanced models, including latest deep learning models and classical machine learning models.
• Optimized performance with JIT and parallelization when possible, using numba and joblib.
• Customizable modules and flexible design: each module may be turned on/off or totally replaced by custom functions. The trained models can be easily exported and reloaded for fast exexution and deployment.

API Demo for LSTM on Phenotyping Prediction:

# load pre-processed CMS dataset
from pyhealth.data.expdata_generator import cms as cms_expdata_generator

cur_dataset = cms_expdata_generator(exp_id=exp_id, sel_task='phenotyping')
cur_dataset.get_exp_data()
cur_dataset.load_exp_data()

# initialize the model for training
from pyhealth.models.lstm import LSTM
clf = LSTM(exp_id, task='phenotyping')
clf.fit(cur_dataset.train, cur_dataset.valid)

# load the best model for inference
clf.load_model()
clf.inference(cur_dataset.test)
pred_results = clf.get_results()

# evaluate the model
from pyhealth import evaluation
evaluator = evaluation.__dict__['phenotyping']
r = evaluator(pred_results['hat_y'], pred_results['y'])

Citing PyHealth:

PyHealth paper is under review at JMLR (machine learning open-source software track). If you use PyHealth in a scientific publication, we would appreciate citations to the following paper:

@article{zhao2020pyhealth,
  author  = {Zhao, Yue and Qiao, Zhi and Xiao, Cao and Glass, Lucas and Hu, Xiyang and Sun, Jimeng},
  title   = {PyHealth: A Python Library for Healthcare AI},
  year    = {2020},
}

or:

Zhao, Y., Qiao, Z., Xiao, C., Glass, L., Hu, X and Sun, J., 2020. PyHealth: A Python Library for Healthcare AI.

Key Links and Resources:

• View the latest codes on Github
• Execute Interactive Jupyter Notebooks
• Check out the PyHealth paper

Table of Contents:

• Installation
• API Cheatsheet & Reference
• Preprocessed Datasets & Implemented Algorithms
• Quick Start for Data Processing
• Quick Start for Running Predictive Models
• Algorithm Benchmark
• Blueprint & Development Plan
• How to Contribute
• Inclusion Criteria

It is recommended to use pip for installation. Please make sure the latest version is installed, as PyHealth is updated frequently:

pip install pyhealth            # normal install
pip install --upgrade pyhealth  # or update if needed
pip install --pre pyhealth      # or include pre-release version for new features

Alternatively, you could clone and run setup.py file:

git clone https://github.com/yzhao062/pyhealth.git
cd pyhealth
pip install .

Required Dependencies:

• Python 3.5, 3.6, or 3.7
• combo>=0.0.8
• joblib
• numpy>=1.13
• numba>=0.35
• pandas>=0.24
• scipy>=0.19.1
• scikit_learn>=0.20
• torch
• xlrd >= 1.0.0

Warning 1: PyHealth has multiple neural network based models, e.g., LSTM, which are implemented in PyTorch. However, PyHealth does NOT install these DL libraries for you. This reduces the risk of interfering with your local copies. If you want to use neural-net based models, please make sure PyTorch is installed. Similarly, models depending on xgboost, would NOT enforce xgboost installation by default.

Full API Reference: (https://pyhealth.readthedocs.io/en/latest/pyhealth.html). API cheatsheet for most learning models:

• fit(X_train, X_valida): Fit a learning model.
• inference(X): Predict on X using the fitted estimator.
• evaluator(y, y^hat): Model evaluation.

Model load and reload:

• load_model(): Load the best model so far.

(i) Preprocessed Datasets (customized data preprocessing function is provided in the example folders):

You may download the above datasets at the links. The structure of the generated datasets can be found in datasets folder:

• \datasets\cms\x_datat\...csv
• \datasets\cms\y_data\phenotyping.csv
• \datasets\cms\y_data\mortality.csv

The processed datasets (X,y) should be put in x_data, y_data correspondingly, to be appropriately digested by deep learning models.

(ii) Machine Learning and Deep Learning Models :

Examples of running ML and DL models can be found below, or directly at \examples\learning_examples\

(iii) Evaluation Metrics :

(iv) Supported Tasks:

We propose the idea of standard template, a formalized schema for healthcare datasets. Ideally, as long as the data is scanned as the template we defined, the downstream task processing and the use of ML models will be easy and standard. In short, it has the following structure: add a figure here. The dataloader for different datasets can be found in examples/data_generation. Using "examples/data_generation/dataloader_mimic_demo.py" as an exmaple:

1. First read in patient, admission, and event tables.

   from pyhealth.utils.utility import read_csv_to_df
   patient_df = read_csv_to_df(os.path.join('data', 'mimic-iii-clinical-database-demo-1.4', 'PATIENTS.csv'))
   admission_df = read_csv_to_df(os.path.join('data', 'mimic-iii-clinical-database-demo-1.4', 'ADMISSIONS.csv'))
   ...

2. Then invoke the parallel program to parse the tables in n_jobs cores.

   from pyhealth.data.base_mimic import parallel_parse_tables
   all_results = Parallel(n_jobs=n_jobs, max_nbytes=None, verbose=True)(
   delayed(parallel_parse_tables)(
        patient_df=patient_df,
        admission_df=admission_df,
        icu_df=icu_df,
        event_df=event_df,
        event_mapping_df=event_mapping_df,
        duration=duration,
        save_dir=save_dir)
    for i in range(n_jobs))

3. The processed sequential data will be saved in the prespecified directory.

   with open(patient_data_loc, 'w') as outfile:
       json.dump(patient_data_list, outfile)

The provided examples in PyHealth mainly focus on scanning the data tables in the schema we have, and generate episode datasets. For instance, "examples/data_generation/dataloader_mimic_demo.py" demonstrates the basic procedure of processing MIMIC III demo datasets.

1. The next step is to generate episode/sequence data for mortality prediction. See "examples/data_generation/generate_mortality_prediction_mimic_demo.py"

   with open(patient_data_loc, 'w') as outfile:
       json.dump(patient_data_list, outfile)

By this step, the dataset has been processed for generating X, y for phenotyping prediction. It is noted that the API across most datasets are similar. One may easily replicate this procedure by calling the data generation scripts in \examples\data_generation. You may also modify the parameters in the scripts to generate the customized datasets.

Preprocessed datasets are also available at \datasets\cms and \datasets\mimic.

"examples/learning_models/lstm_cms_example.py" demonstrates the basic API of using LSTM for phenotyping prediction. It is noted that the API across all other algorithms are consistent/similar.

If you do not have the preprocessed datasets yet, download the \datasets folder (cms.zip and mimic.zip) from PyHealth repository, and run \examples\learning_models\extract_data_run_before_learning.py to prepare/unzip the datasets.

1. Setup the datasets. X and y should be in x_data and y_data, respectively.

   # load pre-processed CMS dataset
   from pyhealth.data.expdata_generator import cms as cms_expdata_generator
   
   cur_dataset = cms_expdata_generator(exp_id=exp_id, sel_task='phenotyping')
   cur_dataset.get_exp_data()
   cur_dataset.load_exp_data()

2. Initialize a LSTM model, you may set up the parameters of the LSTM, e.g., n_epoch, learning_rate, etc,.

   # initialize the model for training
   from pyhealth.models.lstm import LSTM
   clf = LSTM(exp_id, task='phenotyping')
   clf.fit(cur_dataset.train, cur_dataset.valid)

3. Load the best shot of the training, predict on the test datasets

   # load the best model for inference
   clf.load_model()
   clf.inference(cur_dataset.test)
   pred_results = clf.get_results()

4. Evaluation on the model. Multiple metrics are supported.

   # evaluate the model
   from pyhealth import evaluation
   evaluator = evaluation.__dict__['phenotyping']
   r = evaluator(pred_results['hat_y'], pred_results['y'])

The comparison among of implemented models will be made available later with a benchmark paper. TBA soon :)

The long term goal of PyHealth is to become a comprehensive healthcare AI toolkit that supports beyond EHR data, but also the images and clinical notes.

• The support of image datasets and clinical notes
• The compatibility and the support of OMOP format datasets
• Model persistence (save, load, and portability)
• The release of a benchmark paper with PyHealth