High Usability but Limited Case Capture: Performance of Thailand’s Digital Influenza Surveillance System in a Private Hospital Setting

Abstract

The Digital Disease Surveillance (DDS) system was introduced in 2024 to enhance real-time monitoring of influenza. However, there has been no performance assessment among private hospitals that have adopted this system. This study assessed the performance of the DDS for influenza surveillance at a private hospital in Thailand. We conducted a mixed-methods study from January to December 2024. We analyzed data from the Hospital Information System, DDS and interviewed 23 stakeholders. We assessed system attributes, including sensitivity, positive predictive value (PPV), completeness, accuracy, timeliness, and representativeness. Qualitative findings indicated high system simplicity and usability, with data outputs utilized for hospital-level resource preparation and vaccination campaign planning. Quantitative attributes of 250 reported-cases showed a high PPV (82% for the Division of Epidemiology (DOE) case definition and 100% for physician diagnosis and laboratory-based definitions), 100% data completeness, and 89% timeliness (reporting within 7 days). A critical limitation in system automation was identified, notably incorrect data extraction via an application programming interface (API) necessitated a reliance on manual data entry. This contributed to a low sensitivity (535/5,751: 9%), particularly using the DOE definition, compared to physician diagnosis (622/1,776: 35%) and laboratory-based definitions (723/1,746: 41%). This low sensitivity was attributable to systematic exclusion of outpatients and non-local residents. While the DDS demonstrates high usability and data quality for reported-cases, its reliance on manual workflows due to API failure results in low sensitivity. These gaps limit its effectiveness for comprehensive surveillance. Enhancing API integration, revising case definitions, and standardizing reporting protocols are recommended.