The Tuberculosis Hackathon begins with a question: can we make useful predictive models for sub-national tuberculosis (TB) burden in Pakistan?
In partnership with Pakistan’s National Tuberculosis Control Program, KIT’s Centre for Spatial Epidemiology (CASE) is hosting a virtual hackathon to develop novel approaches to predict TB burden.
TB prevalence surveys are usually powered to provide national prevalence estimates. Yet, these surveys also contain a wealth of sub-national information that could benefit decision makers and populations. This hackathon brings together research groups interested in TB modelling to collaborate on a joint modelling exercise to estimate sub-national TB burden in Pakistan in 2018.
The hackathon began on 20 May and concluded on 3 September 2019.
The hackathon: how it works
- Pakistan’s NTP provided their 2010-2011 TB prevalence survey data and other TB data for this hackathon.
- Each group (or individual) was given three months and the same set of data to develop their own models to estimate sub-national TB burden.
- Participants were invited to use these data, and any other publicly available data for their modelling approach to succeed
Models developed as part of this hackathon were appraised and compared by an evaluation panel comprised of global TB and statistical experts including representatives from Pakistan’s NTP and the World Health Organisation. They were also evaluated on the basis of standard evaluation criteria that reflect the precision, consistency (internal and external) and credibility of the model.
The hackathon was open to individual researchers as well as teams. The following teams participated:
- Team led by M.S. Rathinavel from ICMR-NIE including Dr. Prakash
- Team led by Pete Dodd from University of Sheffield including Debebe Adewo and Peter MacPherson
- Team United4TB led by Vincent Meurrens from Epcon including Zhi Zhen, Yumna Moosa, Farihah Malik, Thys Potgieter
- Team led by Stewart Chang Institute for Disease Modeling including Bradley Wagner, Karyn Sutton, Anna Bershteyn, William Trouleau, Amjad Khan, and Jens Levy
- Team led by Fulvia Mecatti from the University of Milano-Bicocca including Maeregu Arisido and Gaia Bertarelli
- Roel Bakker from LSHTM TB Modelling Group
- Nimalan Arinaminpathy from Imperial College London
- Team led by Jennifer Ross from the Institute for Health Metrics and Evaluation including Nathaniel Henry, Emma Spurlock, Kate LeGrand, Bobby Reiner, Mingyou Yang, Brigette Blacker, and Audrey Batzel.
- The Mystery Team
KIT awarded the winner of the hackathon at the 50th Union World Conference on Lung Health in Hyderabad – Congratulations to the Team led by Stewart Chang from the Institute for Disease Modeling including Bradley Wagner, Karyn Sutton, Anna Bershteyn, William Trouleau, Amjad Khan, and Jens Levy.
We will invite participants to collaborate on satellite session at the Union conference co-ordinated by KIT CASE, as well as a summary publication describing the various methodological approaches and performance. The publication will also reflect on the models’ strengths and limitations.