OpenTrials Hackathon
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.idea
docs
pytrials
results
.gitignore
.travis.yml
LICENSE
README.md
requirements.txt

README.md

License (LGPL version 3) Build Status

OpenTrials

This repository hosts the work of the OpenTrials Hackathon 2016 (#OTHackDay).
http://opentrials.net/
https://www.eventbrite.com/e/opentrials-hack-day-tickets-27046834811

Content in this repository is licensed

Introduction

OpenTrials is a collaboration between Open Knowledge and Dr Ben Goldacre from the University of Oxford DataLab. It aims to locate, match, and share all publicly accessible data and documents, on all trials conducted, on all medicines and other treatments, globally.

OpenTrials is building a collaborative and open linked database for all available structured data and documents on all clinical trials, threaded together by individual trial. With a versatile and expandable data schema, it is initially designed to host and match the following documents and data for each trial:

  • registry entries
  • links, abstracts, or texts of academic journal papers
  • portions of regulatory documents describing individual trials
  • structured data on methods and results extracted by systematic reviewers or other
  • researchers
  • Clinical Study Reports
  • additional documents such as blank consent forms, blank case report forms, and protocols

The data in OpenTrials is mainly collected via

  • scraping registries
  • donations of structured data
  • crowdsourced document contributions

HackDay Project

Within the HackDay we wanted get an introduction to the OpenTrials database structure & content, and use the API via python for queries.

In addition we were interested in the following questions:

  • What conditions are shared between trials for a given subset of trials?
  • What interventions are shared between trials for a given subset of trials?
  • What is the graph/network structure of the trial <-> intervention <-> condition graph for a certain query?
  • What can we learn from this graph structure (hubs, connections, connected components)?

Our strategy to answer these questions was

  • use the python swagger API to query OpenTrials
  • create the trial <-> intervention <-> condition graph from the query results (GML format)
  • visualize the results for example queries (NAFLD, diabetes type 2, depression) in a graph visualization software like Cytoscape or gephi

In the created graphs, trials, interventions and conditions nodes are marked according to the following legend:

graph legend

The resulting graph in Cytoscape for condition.name:NAFLD shows an interesting structure consisting of a large component and a handful of unconnected componets. Some condition hubs with many connections are observed in the large component.

NAFLD complete

Here we zoomed in into the NAFLD graph with labels NAFLD part

Alternatively we can import the GML format in other visualization software like gephi NAFLD gephi

In the following the resulting Cytoscape graph for condition.name:depression is shown. Much more trials exist for depression than NAFLD in OpenTrials.

NAFLD complete

Main things we learned:

  • the general graph structure of the trial <-> intervention <-> condition network is similar for different conditions consisting of one large connected component with a dense center and a collection of small components. It seems that some trials use distinct condition and intervention terms than all other studies. The reason could be a missing standardization of the terms or some studies which are not so mainstream, looking at uncommon conditions and interventions.
  • few hub nodes exist in the conditions and interventions which are used by many trials. Some of the hubs are identical in their semantic meaning, like condition:NAFLD and condition:Nonalcoholic fatty liver disease These identical terms are not normalized. The normalization of terms results in node duplication with identical meaning.

Installation & Usage

The python code is based on the OpenTrials example notebook
https://github.com/pwalsh/notebooks/blob/master/opentrials/opentrials.ipynb which can be accessed via

git clone https://github.com/pwalsh/notebooks.git opentrials-notebooks
cd opentrials-notebook
jupyter notebook

The requirements to run the python code and the examples are listed in requirements.txt.

To run the code clone the repository via

git clone https://github.com/matthiaskoenig/opentrials

and check the functionality by running the unittests via

nosetests

The OpenTrials queries and query graphs for the examples can than be generated executing pytrials/examples.py.
You should see an output similar to

...
*** Query: breast AND cancer AND tamoxifen ***
page: 1
Cummulative results: 0.207039 [100/483]
page: 2
Cummulative results: 0.414079 [200/483]
page: 3
Cummulative results: 0.621118 [300/483]
page: 4
Cummulative results: 0.828157 [400/483]
page: 5
Cummulative results: 1.000000 [483/483]
483 483
*** Graph: conditions.name:depression ***
*** Graph: conditions.name:NAFLD ***
*** Graph: conditions.name:diabetes AND type AND 2 ***
*** Graph: breast AND cancer AND tamoxifen ***

Process finished with exit code 0

Resources

OpenTrial API documentation
http://api.opentrials.net/v1/docs/

OpenTrial redash
https://app.redash.io/opentrials/
email: opentrials@opentrials.net
pw: othackday

OpenTrial API repository
https://github.com/opentrials/api

OpenTrials Documentation repository
https://github.com/opentrials/docs

This database is based on WHO Trial Registration Data Set:
http://www.who.int/ictrp/network/trds/en/

http://www.documentcloud.org
https://www.documentcloud.org/public/search/Group:%20okfn%20Project:%20%22OpenTrialsFDA%22
http://opentrials.net/2016/08/10/opentrialsfda-unlocking-the-trove-of-clinical-trial-data-in-drugsfda/