RareHacks: first Hackathon for the Share4Rare project

10/07/2019

On 5–7 July, the first Hackathon was held for the Share4Rare project: RareHacks. In three days, participants created a new way of solving some of the problems associated with rare diseases.

The event was organised by the UPC, one of the members of the consortium Share4Rare. In total, 45 people participated in the design of an innovative chatbot designed to help redress the limited access to reliable medical information for patients and relatives who live with a rare disease. Specifically, the exercise was focused on paediatric melanoma, an ultra-rare disease with an extremely low prevalence that differs significantly from the disease that occurs in adults.

In total, six teams completed the process that was carried out at MOB Barcelona (Makers of Barcelona). The teams coded and demonstrated the characteristics and functioning of their chatbox before five members of the jury: Samir Kanaan from the Universitat Politècnica de Catalunya (UPC), Begonya Nafria from the Sant Joan de Déu Foundation, Xavier Escoté and Salvador Quadres from Deister Software, and Alexandre Perera from the CREB UPC.

The winning team, comprised of Sergi del Río, Celia Sánchez, Lucía Chacón, David Pujol and Oriol Aranda, was awarded €1,500, a prize sponsored by Deister Software. They created a chatbot, called RareBot, which provides information according to the various types of melanoma and helps users with details such as the nearest certified centre (hospital) to the user according to their geolocation. In addition, the team added a multilanguage capacity to the tool so that users can interact in any language, although the system is originally fed by texts in English. Other added characteristics are word correction during the interaction, the inclusion of various questions to validate the correct type of melanoma before displaying the information in response to queries, and the possibility of many users coinciding.

Second prize was given to Pau Cutrina, Marc Sos, David Sanchez, Josep Cordón, Edgar Alarcón and Josep Munuera, who used a dynamic called web scraping. With this design, there is no need for a doctor to manage the databases used to feed the bot because existing indexed databases are used. The team planned a scalable solution that updates on its own and would be flexible enough to be adapted to other diseases.

All the ideas generated in this hackathon will be analysed and incorporated to improve the user experience in the field of tools for accessing validated, accurate information on rare diseases.