Introduction and Methods: Pertussis has recently re-emerged in the United States. Timely surveillance is vital to estimate the burden of this disease accurately and to guide public health response. However, the surveillance of pertussis is limited by delays in reporting, consolidation and dissemination of data to relevant stakeholders. We fit and assessed a real-time predictive Google model for pertussis in California using weekly incidence data from 2009-2014.
Results and Discussion: The linear model was moderately accurate (r = 0.88). Our findings cautiously offer a complementary, real-time signal to enhance pertussis surveillance in California and help to further define the limitations and potential of Google-based epidemic prediction in the rapidly evolving field of digital disease detection.