| Description: |
The digital era enables real-time monitoring of preventive health behaviours using unconventional data from social media and online searches. These sources can provide early indications of public attitudes toward protective measures, thereby supporting public health strategies. This systematic review examines the use of such data to track behaviours related to influenza and dengue, introducing an innovative approach to epidemiological surveillance. This work was a part of the INF-ACT project “BEHAVE-MOD: behaviour and sentiment monitoring and modelling for outbreak control”. The systematic review involved five databases (PubMed, Scopus, Embase, medRxiv and PsycINFO), with 4,563 articles screened after the removal of duplicates. Four researchers analysed the titles and abstracts in a double-blind process. Data were extracted regarding the type of source, monitored health behaviour, advantages and limitations in the use of unconventional data. The study quality was evaluated using the JBI checklist, and data analysis is currently ongoing. A total of 44 studies were included: 33 on influenza, 9 on dengue, and 2 on both. Most influenza studies originated from North America (61%), while dengue studies were mainly from Asia (44%). The primary data sources were online searches, particularly Google Trends, and social media, notably Twitter. Other studies analysed websites, apps, pharmaceutical sales, and school registries. The monitored behaviours ranged from attitudes and informational searches to social media interactions and public perceptions. Unconventional data sources are promising for rapidly monitoring preventive health behaviours and public sentiment. However, critical challenges are still related to representativeness, bias, and privacy. These innovative methods should complement traditional surveillance and require ongoing validation. Further analysis is needed to better assess their effectiveness across different contexts. |