Inspiration
As the initial symptoms of the COVID-19 pandemic flu are similar to those of the seasonal flu, case detection is based on social histories (travel and contact histories) and clinical manifestations, followed by diagnostic testing for confirmation. Rapid screening and referral of suspected cases will enable prompt early detection in the outbreak phase and slow down the epidemic amplification of the disease. However, resource restrictions and capacity constraints in Nigeria and other resource-poor settings have slowed down the screening and referral of suspected cases. This solution would also act as a triage system for the call centre officers of the NCDC; enabling them to focus on reaching out to the high-risk groups most in need of testing.
What it does
The solution is a Facebook chatbot that assesses COVID19 risk, notifies testing facilities of high-risk cases, provides actionable info to all risk groups, schedules follow-up screens in14 days, and analyzes & reports. We are talking with the central testing institution to adopt the solution and provide results of test data. This will enable the analyses of the most likely combination of indicators that lead to a positive COVID19 test, the current hotspots for COVID19 infection. It would also enable the tracing of close contacts of COVID19 confirmed patients.
How we built it
The assessment algorithm was natively designed to reflect the World Health Organization's guidelines on the COVID19 disease, its risk factors, and how it is spread. Each question assesses a risk factor and weights them. The risk weight is based on whether the country is undergoing transmission through an index spread or through community spread. This question bank and weighting were built into a chat system using the Chatfuel tool and hosted on Facebook. Details from high-risk cases are immediately shared - by email - with the testing institution who reaches out to them to follow up with diagnostic tests. The respondent MUST actively consent for their details to be shared.
Challenges we ran into
- Privacy concerns from the testing institution about where the metadata is stored and who had access to it. We are incorporating the ability for responders to delete their data when they wish to allay some of these concerns.
- Payments for access to Chatfuel
Accomplishments that we're proud of
We developed the entire system; from algorithm design to Facebook implementation within one week! This speed of development is remarkable for us! Furthermore, the chatbot is starting to gain traction, and we hope that it would rapidly increase COVID19 prevention by identifying those who need the test the most in resource-poor settings.
What we learned
Simple solutions delivered through widely available platforms can reach and help a lot more people around the world much faster than attempting to build platforms from scratch.
What's next for COVID19 Screening Bot
We intend to incorporate a contact tracing module that receives input on which screened cases are positive, then helps to trace their contacts through social networks and also by identifying known hotspots and all those who were close to those hotspots within a locale.
Log in or sign up for Devpost to join the conversation.