An interactive national digital surveillance system to fight against COVID-19 in Bangladesh

• Published in: Frontiers in digital health 2023-05, Vol.5, p.1059446-1059446
• Main Authors: Sarker, Farhana, Chowdhury, Moinul H, Ratul, Ishrak Jahan, Islam, Shariful, Mamun, Khondaker A
• Format: Article
• Language: English
• Subjects: artificial intelligence; COVID-19; Digital Health; screening; surveillance system
• ISSN: 2673-253X; 2673-253X
• DOI: 10.3389/fdgth.2023.1059446

COVID-19 has affected many people globally, including in Bangladesh. Due to a lack of preparedness and resources, Bangladesh has experienced a catastrophic health crisis, and the devastation caused by this deadly virus has not yet been halted. Hence, precise and rapid diagnostics and infection tracing are essential for managing the condition and limiting its spread. The conventional screening procedure, such as reverse transcription polymerase chain reaction (RT-PCR), is not available in most rural areas and is time-consuming. Therefore, a data-driven intelligent surveillance system can be advantageous for rapid COVID-19 screening and risk estimation. This study describes the design, development, implementation, and characteristics of a nationwide web-based surveillance system for educating, screening, and tracking COVID-19 at the community level in Bangladesh. The system consists of a mobile phone application and a cloud server. The data is collected by community health professionals home visits or telephone calls and analyzed using rule-based artificial intelligence (AI). Depending on the results of the screening procedure, a further decision is made regarding the patient. This digital surveillance system in Bangladesh provides a platform to support government and non-government organizations, including health workers and healthcare facilities, in identifying patients at risk of COVID-19. It refers people to the nearest government healthcare facility, collecting and testing samples, tracking and tracing positive cases, following up with patients, and documenting patient outcomes. This study began in April 2020, and the results are provided in this paper till December 2022. The system has successfully completed 1,980,323 screenings. Our rule-based AI model categorized them into five separate risk groups based on the acquired patient information. According to the data, around 51% of the overall screened populations are safe, 35% are low risk, 9% are high risk, 4% are mid risk, and the remaining 1% is very high risk. The dashboard integrates all collected data from around the nation onto a single platform. This screening can help the symptomatic patient take immediate action, such as isolation or hospitalization, depending on the severity. This surveillance system can also be utilized for risk mapping, planning, and allocating health resources to more vulnerable areas to reduce the virus's severity.