2009 using-reality-mining-to-improve-public-health-and-medicine
The whitepaper, authored by Alex Pentland, David Lazer, Devon Brewer, and Tracy Heibeck, surveys how reality mining – the large‑scale collection and analysis of digital traces from mobile phones, sensors, and other ubiquitous devices – can transform public health and medicine. It outlines the technical capabilities of reality mining (individual health assessment, automatic social‑network mapping, behavior‑pattern segmentation, and population‑level health monitoring), illustrates concrete applications (depression and Parkinson’s diagnosis, disease‑spread modeling, environmental exposure tracking, mental‑health screening, treatment monitoring, and social‑support interventions), and discusses ethical, legal, and privacy challenges. The authors propose a “new deal on data” that places ownership and control of personal data in the hands of individuals, and they issue recommendations for building a computational public‑health (CPH) science: develop infrastructure, user‑friendly tools, robust privacy safeguards, and interdisciplinary training. The paper argues that reality mining can provide second‑by‑second, behavior‑rich data that enable early detection, personalized treatment, and more effective public‑health interventions, but only if privacy, data‑ownership, and institutional frameworks keep pace with the technology.
Topics
Executive Vision: Reality Mining as a Transformative Public‑Health Lens
Sets the stage by describing digital breadcrumbs, the emergence of reality mining, and its potential to reshape health research and delivery.
Core Capability 1 – Individual Health Assessment via Mobile Sensors
Explains how accelerometers, microphones, GPS, and call logs can infer depression, Parkinson’s, autonomic arousal, attention, mimicry, and cognitive load, framing these as low‑resolution brain‑state scans.
Core Capability 2 – Automatic Social‑Network Mapping
Describes MIT phone deployments, proximity sensing, and statistical models that achieve ~96 % accuracy in reconstructing friendship, coworker, and family networks.
Core Capability 3 – Behavior‑Pattern Segmentation Beyond Demographics
Shows how location traces reveal a small repertoire of movement patterns, enabling clustering of subpopulations and measurement of mixing between behavior groups.
Core Capability 4 – Population‑Level Health Indicators from Communication Diversity
Presents large‑scale analyses (250 M phone numbers, 12 B calls) linking communication diversity to socioeconomic deprivation and identifying ‘information ghettos’.
Future Application 1 – Contagious Health Behaviors and Network‑Based Interventions
Discusses evidence that smoking and obesity spread through social ties, and proposes using reality mining to pinpoint causal mechanisms and leverage key network nodes for behavior change.
Future Application 2 – Real‑Time Infectious‑Disease Tracing and Outbreak Detection
Explores GPS‑based proximity tracking for pinpointing exposure sites, early outbreak signals, and augmenting traditional surveillance (e.g., SARS, flu).
Future Application 3 – Fine‑Grained Environmental Exposure Measurement
Argues that phone‑based location data combined with ambient air‑quality sensors can produce dynamic, individual‑level pollution exposure profiles.