Analysis: Google, EHRs offer alternative method for flu surveillance

A recent study suggests EHR data used in conjunction with information obtained from Google and Twitter may prove to be the most up-to-date method of flu surveillance, according to an analysis by athenaInsight senior editor Gale Pryor.

Here are seven insights Ms. Pryor shared about flu surveillance.

1. Flu surveillance typically relies on public health laboratories to report influenza incidents to the CDC. However, public health laboratories' measures require four to five days for test results and up to three weeks before these results are reported to the public.

2. To create a real-time method of flu surveillance, a team of researchers from Boston-based Harvard Medical School analyzed data from Google, Twitter, cloud-based EHRs and a crowd-sourced platform to attempt to track influenza during four flu seasons in Boston.

3. Their results, published in the journal JMIR Public Health and Surveillance in January, suggested this method of aggregating information from various data sources was able to detect early indicators of influenza activity with the same level of accuracy as the Boston Public Health Commission's methods for tracking flu diagnoses in the city.

4. Athenahealth's cloud-based EHR was a key source of data for the project, according to Ms. Pryor. "Cloud-based systems offer near-real-time visibility of patient visits coded for [influenza-like illnesses], and the data can be cut by age, gender and zip code of provider," she wrote.

5. For the Harvard study, crossing datasets from EHRs with those from social media platforms proved particularly helpful. Real-time analyses of Twitter posts with #flu and #fever are a robust form of data due to the volume of activity. However, Ms. Pryor also noted such flu-related posts may not be indicative of flu incidence.

6. Google data, though imperfect, has also shown promise. A 2008 initiative called Google Flu Trends successfully tracked influenza incidence using flu-related search terms for several years. However, in 2013, a rise in searchers for N3N2 or "swine flu" overestimated incidence by 140 percent.

7, When used together, disparate data sources like EHRs, Twitter and Google proved useful in the Harvard study, despite having shortcomings when used on their own.

"You're going to have different data streams telling you different things," John Brownstein, PhD, study author, chief innovation officer at Boston Children's Hospital and professor of pediatrics at Harvard Medical School, told Ms. Pryor. "But if you can bring together many different layers of information … that is going to give you the most robust picture of a flu epidemic."

To access Ms. Pryor's analysis, click here.

More articles on data analytics & precision medicine:
Penn State launches big data project to drive population health, personalized medicine research
Researchers use predictive analytics, EHRs to forecast hypertension onset
Cincinnati Children's Hospital Medical Center launches Maternal and Infant Data Hub

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months