The role of social media in monitoring and predicting influenza cases in Beijing: s time series analysis

<bold>Background:</bold> Influenza remains a global health challenge, necessitating robust surveillance systems for timely intervention. With the prevalence of social media, its potential in health surveillance needs exploration. <bold>Methods:</bold> Using outpatient and emergency data and social media data, we analyzed trends in influenza-like illness cases and related media activity. Descriptive analysis, lead-lag correlation, and causal inference through the PCMCI algorithm were applied. The media’s impact on ILI case trends was quantified using Ordinary Least Squares regression. <bold>Results:</bold> Weekly analysis showed ILI cases peak during winter, with a consistent increase. Media index analysis revealed exponential growth and peaks aligning with ILI increases. Lead-lag correlation suggested a one-week lag between media index and ILI cases. PCMCI confirmed a causal relationship, indicating that ILI cases were positively correlated with the media index with a one-week delay. Model comparisons highlighted the improved predictive capability when incorporating media index data. <fig><object-id>erj;64/suppl_68/PA5108/F1</object-id><object-id>F1</object-id><object-id>F1</object-id><graphic></graphic></fig> <bold>Conclusion:</bold> The study confirmed the utility of social media data as an augmenting factor for influenza surveillance, providing vital insights into outbreak dynamics. The incorporation of media indices in predictive models could enhance early warning systems. <table-wrap><object-id>erj;64/suppl_68/PA5108/T1</object-id><object-id>T1</object-id><object-id>t1</object-id><table><colgroup><col></col><col></col><col></col><col></col><col></col></colgroup><tbody><tr><td></td><td>R-Squared</td><td>Adjusted R-Squared</td><td>AIC</td><td>BIC</td></tr><tr><td>Model1</td><td>0.740</td><td>0.737</td><td>1587</td><td>1596</td></tr><tr><td>Model2</td><td>0.823</td><td>0.821</td><td>1528</td><td>1537</td></tr></tbody></table></table-wrap>