Pakistan: Virtual Hackathon aims to develop novel approaches for TB burden prediction

Pakistan: Virtual Hackathon aims to develop novel approaches for TB burden prediction

[vc_row css_animation=”” row_type=”row” use_row_as_full_screen_section=”no” type=”full_width” angled_section=”no” text_align=”left” background_image_as_pattern=”without_pattern”][vc_column][vc_separator type=”normal” thickness=”2″ up=”20″ down=”20″][/vc_column][/vc_row][vc_row css_animation=”” row_type=”row” use_row_as_full_screen_section=”no” type=”full_width” angled_section=”no” text_align=”left” background_image_as_pattern=”without_pattern”][vc_column][vc_gallery type=”nivo” interval=”3″ images=”2871″ img_size=”full”][vc_column_text]From May to August 2019, the Royal Tropical Institute’s (KIT) Centre for Spatial Epidemiology (CASE), together with Pakistan’s National Tuberculosis Control Program and Stop TB Partnership, hosted a virtual hackathon to develop novel approaches to predict TB burden.

This hackathon brought together research groups interested in TB modelling to collaborate on a joint modelling exercise to estimate sub-national TB burden in Pakistan in 2019. The models developed in this hackathon will be evaluated by an evaluation panel of experts and the winner will be announced at the 50th Union World Conference on Lung Health in Hyderabad, India.

Sandra Alba, epidemiologist at KIT, and Abdullah Latiff of the National Tuberculosis Control Program, Pakistan, provide some insight into HackTB. To read more about this initiative, click here.

Why Pakistan?

Abdullah: Pakistan is one of the high burden countries for TB. In the last 2 decades, the National TB control program has made tremendous efforts to control TB in the country using only WHO estimates for strategic planning. A prevalence survey was conducted in 2010-11 to estimate the real burden of the disease but it only served to provide national estimates.

Pakistan is a diverse country that spreads from the Himalayas to the Indian Ocean and from India to Afghanistan and Iran. Some areas like Baluchistan are scarcely populated while other areas like Punjab have a dense population living in close proximity. This geographical diversity facilitates and harbors a diversity of clusters of infection, which does not justify a uniform strategy for infection control; hence, national estimates cannot be applied to sub-national levels.

Sandra: KIT epidemiologists and the Pakistan NTP have a long history of collaborating together. When we first had the idea of organising a hackathon, we approached different NTPs who had recently conducted a TB prevalence survey. We immediately had a very good contact with the NTP in Pakistan who perfectly understood what we had in mind and how it could potentially benefit their program. The rest is history!

Why a hackathon model?

Sandra: Hackathons are increasingly being proposed as a model to solve complex problems by leveraging the mixed skills of a group of people. They originate from the field of computer programming and software development, yet a number of hackathons have already been proposed in global health. The core idea of a hackathon is that by collaborating we can co-create something that is greater than the sum of the parts. Learning from each other is central to this.

We thought the playful and collaborative nature of this approach could be ideal to bring together different modelling approaches for sub-national TB. Sub-national TB estimates are very valuable for NTPs to plan their programs and optimize resource allocation. The Pakistan NTP is among many programs worldwide who value this data, but do not have access to reliable estimates. Many statistical and mathematical approaches have been developed for this purpose, but it’s very difficult to assess their accuracy, since we do not have ‘gold standards’ to compare the estimates with.

What do you except from the TB hackathon?

Sandra: From an NTP perspective, we hope that this hackathon can provide program planners with useful data to tailor their TB control efforts to different sub-national contexts. From a modelling perspective, we hope that we can all learn from comparing the different methods and can further refine our approaches.

Abdullah: The Pakistan NTP needs close estimates of different TB variables. So far, we do not have access to reliable estimates. The Hackathon’s proposed sub-national TB estimates will be very valuable for us to plan our programs and may help to identify districts with missed TB cases and optimize resource allocation.

Would you recommend other countries to apply this model?

Abdullah: I definitely recommend countries lacking sub-national estimates to apply this model. The hackathon is a great learning experience and we expect the results of this exercise to be very promising. We are hoping that it will at least provide us with rough estimates which are very much needed for better and effective program management.

Are there plans to expand to other countries?

Sandra: Ideally, yes! But we would leave it to others to organize it. In order to ensure fairness and transparency in the evaluation process we did not think it was appropriate for us to participate this year as a modelling team. However, we are very curious to know how our modelling approach would compare to those currently applied by the HackTB participants.

Also, the spirit of the hackathon is very much one of learning, so we expect this year’s participants to learn from the experience and to apply those lessons learned in future modelling enterprises. We encourage participants of this year’s hackathon to organize another edition in the future and we look forward to join in as participating modelers![/vc_column_text][/vc_column][/vc_row]