Cell-phone traces reveal infection-associated behavioral change.

Area of Science:

• Epidemiology
• Public Health
• Mobile Technology Data Analysis

Background:

• Predicting epidemic trajectories and understanding human behavior are crucial for effective epidemic preparedness.
• Behavioral changes during outbreaks can reduce the reliability of traditional surveillance methods like social media monitoring.
• Mobile phone call-detail records (CDRs) offer a novel source of real-time behavioral data.

Purpose of the Study:

• To measure behavioral changes during an epidemic using mobile phone call-detail records (CDRs).
• To assess the utility of anonymously linked CDRs and health data for augmenting epidemic surveillance.
• To inform infectious disease modeling by incorporating explicit behavior-change mechanisms.

Main Methods:

• Utilized an anonymously linked dataset of cell-phone users and their influenza-like illness diagnosis dates during the 2009 H1N1v pandemic.
• Analyzed call-detail records (CDRs) to quantify changes in movement and communication patterns during illness.
• Compared mobile phone usage during illness to routine behavior baselines.

Main Results:

• Individuals diagnosed with influenza-like illness showed reduced movement, with 1.1 to 1.4 fewer unique cell tower locations used around diagnosis.
• Diagnosed individuals made fewer calls (2.3 to 3.3 fewer calls) but spent longer on the phone (41- to 66-s increase) the day after diagnosis.
• Measurable differences in mobile phone use indicate significant behavioral shifts during illness.

Conclusions:

• Anonymously linked CDRs and health data provide granular insights into behavioral changes during epidemics.
• Mobile phone data can potentially augment traditional epidemic surveillance and outbreak prediction efforts.
• Infectious disease models should incorporate explicit behavior-change mechanisms for improved accuracy.