Surveillance of tuberculosis by analysing Google Trends

• Autor(a) principal: Santos, Luana Gorgueira
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.14/42246
• topic: Tuberculosis surveillance; Google trends; Search volume; Machine learning; Accuracy metrics; Tuberculosis incidence forecasting; Portugal; Vigilância da tuberculose; Google Trends; Volume de pesquisa; Metricas de precisão; Previsão da incidência da tuberculose; Domínio/Área Científica::Ciências Sociais::Economia e Gestão

Tuberculosis remains a global health concern, having caused around 1.5 million deaths in 2020. The Portuguese medical authorities are facing challenges to meet the goals of the Word Health Organization in the area of Tuberculosis. Early detection of potential Tubercu losis outbreaks is crucial for effective intervention and control, but traditional surveillance systems often suffer from reporting lags and resource limitations, which were aggravated by the COVID-19 pandemic. This thesis explores the potential of using Google Trends to predict Tuberculosis incidence in Portugal. Past research have shown promising results in this area, suggesting that Google Trends search volume could complement existing surveil lance methods. To improve Tuberculosis surveillance system, we developed a syndromic approach using 19 Tuberculosis-related terms extracted from Google Trends. Historical data on the incidence of Tuberculosis was extracted from the European Centre for Disease Prevention and Control. After joining both datasets, we applied different machine learn ing models to forecast the monthly Tuberculosis incidence. Nextly, four accuracy metrics, including the Akaike Information Criterion, were used to select the best predictive model. Our empirical analysis shows that the forecast matches the seasonal patterns of Tubercu losis incidence in Portugal. While there are possible limitations that need to be addressed in future research, the surveillance system developed in this study might be a valuable tool for public health authorities as it provides real-time information on potential new cases. In the long run, this system might help alleviate the burden of Tuberculosis and potentially mitigate future outbreaks.