Hackathon crowd control


Crowdsourced challenges are now an established part of the medical research ecosystem, especially for data analysis problems such as finding the best genomic analysis techniques or new ways to interpret mammography data. Writing a story about these competitions for Nature, I’ve been struck by the rapid spread of their most intense form: the medical hackathon.

Described today in a Science Translational Medicine review, these hackathons (or “datathons”) take place over a weekend or a few evenings, bringing together some mix of medical scientists and engineers, data scientists, clinicians, patients, medical entrepreneurs, public health advocates and other interested parties. Participants “are encouraged to collaborate, fail fast and iterate.”

Hundreds of medical hackathons have been held. Some encourage multiple groups to study the same clinical problem with the same data and compare conclusions. Other hackathons are all about idea generation. The events may target specific threats such as the Zika epidemic, or more general topics such as improved intensive care, or a free-for-all of unsolved medical problems.

Like software hackathons, the medical hackathon “integrates collaboration, idea generation, and group learning by joining various stakeholders in a mutually supportive setting for a limited period of time,” the STM authors say.

Key is the face-to-face mashup of expertise and views, which doesn’t come easily on the outside. “It is difficult to establish a platform for the realtime, respectful, and effective exchange of ideas among specialists who are usually separated by time, space, methods, attitudes, and terminology,” they point out.

This difficulty holds even for global crises like Zika. But hackathons around the world are addressing the rapidly spreading virus, among them one held at Massachusetts General Hospital earlier this month. Among the resulting proposals:

  • An app for crowdsourced mosquito surveillance data collection, with games
  • Larvicide automatic dispensers
  • A public health online rumor-squashing campaign
  • Hairnet-like nets to cover open water containers
  • Applying new diagnostic technology to detect the virus in pregnant women
  • A mosquito larva finder, with a microscope add-on to a smart phone that samples standing water, analyzes for type of larva and adds GPS location data.

As often with more established crowdsourcing competitions like the Dream Challenges, we don’t know which if any of these early results will be driven all the way into the clinic. But the promise is real.