Innovative Use of AI and Wearable Devices to Detect Inflammation Before Symptoms Arise

In the realm of modern medicine, the traditional approach to treating illness has largely been reactive—addressing health issues only once symptoms become evident. However, groundbreaking research from McGill University’s Research Institute at the McGill University Health Center proposes a shift towards a more proactive strategy. Researchers have developed an artificial intelligence (AI) system capable of predicting acute systemic inflammation, a swift immune response to viral respiratory infections, by analyzing biometric data collected from everyday wearable devices such as smart rings, watches, and shirts.

This pioneering platform can identify early immune signals before clinical symptoms appear, paving the way for earlier medical interventions. Early detection can significantly reduce the risk of complications, hospitalizations, and potentially save lives, while also lowering healthcare costs.

The study involved administering a live attenuated influenza vaccine to 55 healthy adults aged 18 to 59, who were monitored continuously from a week before to five days after inoculation. Participants wore three commercial wearable devices simultaneously, tracking vital signs including heart rate, heart rate variability, body temperature, respiratory rate, blood pressure, physical activity, and sleep patterns. These physiological measures, combined with blood tests for systemic inflammatory biomarkers, PCR testing for respiratory pathogens, and self-reported symptom data collected via a smartphone app, generated over two billion data points.

Advanced machine learning algorithms trained on this extensive dataset enabled the development of nine models that could predict surges in systemic inflammation based on subtle physiological changes, as well as a symptom-based model for comparison. The most practical model, relying on the fewest features, achieved nearly 90% sensitivity, demonstrating high accuracy in predicting inflammation ahead of visible symptoms.

Dr. Amir Hadid, the study’s first author, emphasized that physiological signals measured by wearable sensors can reflect immune activity that is otherwise invisible. This technology translates those signals into real-time alerts, providing a crucial window for early intervention.

The models outperformed traditional symptom reporting, especially because some individuals with systemic inflammation did not exhibit noticeable symptoms, while others reported symptoms without actual inflammation—a phenomenon known as the nocebo effect. Impressively, the AI system detected inflammation caused by SARS-CoV-2 infections before symptoms or PCR test confirmation, highlighting its potential for broad application.

Looking ahead, researchers aim to expand these methods to identify inflammation from other viruses like rhinovirus and RSV, using only wearable sensors without the need for invasive blood tests or clinical visits. This advancement could revolutionize infectious disease management by enabling quick, accessible, and non-invasive monitoring, ultimately enhancing patient outcomes and public health responses.