Early detection of an epidemic erythromelalgia outbreak using Baidu search data

Sci Rep. 2015 Jul 28:5:12649. doi: 10.1038/srep12649.

Abstract

Dozens of epidemic erythromelalgia (EM) outbreaks have been reported in China since the mid-twentieth century, and the most recent happened in Foshan City, Guangdong Province early 2014. This study compared the daily case counts of this recent epidemic EM outbreak from February 11 to March 3 with Baidu search data for the same period. After keyword selection, filtering and composition, the most correlated lag of the EM Search Index was used for comparison and linear regression model development. This study also explored the spatial distribution of epidemic EM in China during this period based on EM Search Index. The EM Search Index at lag 2 was most significantly associated with daily case counts in Foshan (ρ = 0.863, P < 0.001). It captured an upward trend in the outbreak about one week ahead of official report and the linear regression analysis indicated that every 1.071 increase in the EM Search Index reflected a rise of 1 EM cases 2 days earlier. The spatial analysis found that the number of EM Search Indexes increased in the middle of Guangdong Province and South China during the outbreak period. The EM Search Index may be a good early indicator of an epidemic EM outbreak.

MeSH terms

  • Asian People
  • China / epidemiology
  • Data Mining / methods*
  • Disease Outbreaks*
  • Early Diagnosis
  • Epidemics*
  • Erythromelalgia / diagnosis
  • Erythromelalgia / epidemiology*
  • Erythromelalgia / ethnology
  • Geography
  • Humans
  • Linear Models
  • Population Surveillance / methods*
  • Reproducibility of Results