Submission ID 131080
| Issue/Objective | Disease surveillance in low-resource settings, including Ethiopia, is often fragmented. Haqila is an artificial intelligence (AI)-powered, multilingual platform integrating five priority diseases, measles, mpox, malaria, polio, and brucellosis, to strengthen community-based surveillance. This study assesses its real-world usability and integration. |
|---|---|
| Methodology/Approach | A mixed-methods design was employed to evaluate Haqila across three zones in the Oromia region, Ethiopia, guided by RE-AIM and CFIR frameworks. The platform was deployed across four end-user domains, including community informants, healthcare workers, laboratory personnel, and decision-making bodies. Quantitative data captured system utilization, reporting timeliness and completeness, and workflow integration using platform analytics and routine health information systems. Qualitative data from in-depth interviews explored usability, acceptability, feasibility, and contextual determinants of adoption. Quantitative data were analyzed using descriptive and inferential statistics across RE-AIM dimensions, while qualitative data were thematically analyzed using a hybrid deductive-inductive approach guided by CFIR, with triangulation to enhance validity. |
| Results | Over a nine-month period, 32 implementation personnel across participating sites actively used the platform, generating 527 suspected case reports spanning all five priority diseases. Haqila achieved high engagement, with consistent cross-domain utilization and integration into routine workflows. Median reporting time from case suspicion to notification was reduced (from 5.2 to 2.2 days), and reporting completeness improved compared to baseline facility records (from 33% to 72%), indicating enhanced surveillance performance. System usage logs showed sustained interaction across user groups, reflecting effective adoption and operational feasibility. Users reported high usability and acceptability, with multilingual functionality and context-sensitive design identified as critical enablers. The platform strengthened coordination across surveillance levels, improving communication and standardization of reporting practices. However, variability in digital literacy, intermittent internet connectivity, and operational constraints in remote settings affected the consistency and speed of uptake. |
| Discussion/Conclusion | Haqila demonstrates the feasibility of implementing a unified, AI-enabled, multi-disease community-based surveillance platform in resource-constrained settings. Its strong usability and measurable improvements in reporting performance support integration into routine workflows. Addressing infrastructure and capacity gaps will be essential for sustained adoption and scale-up, with implications for similar low-resource contexts. |
| Presenters and Affiliations | Gelan Ayana Zewdie University of Toronto Gelane Biru Jimma University Jude Kong University of Toronto |
