Correlating regional aeroallergen effects on internet search activity.

OBJECTIVE: To investigate the correlation between the change in regional aeroallergen levels and Internet search activity related to allergies.

STUDY DESIGN: A retrospective time series analysis using a graphical analytical approach and statistical modeling was used.

SETTING: Tertiary academic hospital setting.

SUBJECTS AND METHODS: There were no specific enrolled subjects. Data from Google Trends were obtained (google.com/trends) for the following search terms: "allergy," "allergies," "pollen," "runny nose," "congestion," and "post nasal drainage." Daily pollen and mold spore count data were obtained for the same period from throughout Texas. Graphical analysis, correlation, and autoregressive integrated moving average (ARIMA) were employed to assess the relationship between aeroallergens on Google search activity.

RESULTS: A strong positive correlation was observed between observed pollen counts and search activity for the terms "allergies" (r pollen = 0.798), "allergy" (r pollen = 0.781), and "pollen" (r pollen = 0.849). Symptom term searches were weakly correlated with pollen and mold counts. Also, ARIMA modeling supported the relationships indicated by the correlations.

CONCLUSION: Search activities for surrogate terms such as "allergy," "allergies," and "pollen" correlate strongly with observed pollen counts but not mold counts. These data demonstrate the usefulness of Google Trends search data in assessing regional disease burdens and offer insight into how the public seeks information about their own illness.