Alizeh Sonia Fatimi, Tazeen Fatima, Asghar Nasir, Muhammad Shariq Shaikh, Imran Siddiqui, Sibtain Ahmed
Google Trends search data effectively reflects and can anticipate dengue case count patterns in Pakistan. This suggests that it could serve as a cost-effective, real-time early warning tool to complement the national surveillance system. Combining digital search monitoring with IDSR reporting could lead to earlier detection and a more proactive public health response.
BACKGROUND: Dengue fever is a significant public health issue in Pakistan. Seasonal outbreaks lead to considerable illness and economic strain. Traditional surveillance systems, like the Integrated Disease Surveillance and Response (IDSR) system, depend mainly on passive reporting. They often identify outbreaks only after transmission has already worsened. Digital data sources, such as Google Trends, might provide early signs of increasing disease activity by reflecting changes in public information-seeking behavior.
OBJECTIVE: To determine if Google search activity for dengue-related terms can predict dengue case counts in Pakistan during 2024.
METHODS: Weekly dengue case count data were collected from the IDSR Weekly Bulletin. Weekly Relative Search Volume (RSV) data for "dengue," "dengue fever," "dengue symptoms," and "dengue treatment" were gathered from Google Trends. Case counts were adjusted to a 0-100 scale. Pearson correlation was used to examine the relationship between search activity and case count data. Time-series causality analysis was carried out using the Toda-Yamamoto method, following tests for stationarity and the specification of the Vector Autoregression (VAR) model.
RESULTS: Strong positive correlations were found between dengue-related search interest and case count data (r = 0.756-0.832; p < 0.001). The strongest correlation was for the term "dengue." Toda-Yamamoto causality testing showed that search activity significantly predicted subsequent increases in dengue case counts for "dengue symptoms" (p = 0.004), "dengue treatment" (p = 0.001), and "dengue" (p = 0.004), with an approximate two-week lag.
CONCLUSION: Google Trends search data effectively reflects and can anticipate dengue case count patterns in Pakistan. This suggests that it could serve as a cost-effective, real-time early warning tool to complement the national surveillance system. Combining digital search monitoring with IDSR reporting could lead to earlier detection and a more proactive public health response.