Promising research from a Google Research and University of Cambridge's Institute

Promising research from a Google Research and University of Cambridge's Institute of Metabolic Science suggesting a possible path for making insulin resistance detectable early (way before onset type II diabetes) from a combination of wearables data and routine blood work. An excellent News and Views commentary (https://lnkd.in/gtQeZD3f) by Christopher Hartshorn does a great job of explaining why this could be a big deal: "Blood-based markers and clinical indices can signal metabolic dysfunction, but they are typically measured sparingly and at later disease stages." "One reason why insulin resistance is difficult to detect early is that the body compensates for declining metabolic efficiency remarkably well. Blood-sugar levels can remain in the normal range even as the physiological effort required to maintain them steadily increases. Most clinical assessments rely on occasional measurements taken under standardized conditions. [...]As a result, early metabolic dysfunction is often invisible until compensatory mechanisms begin to fail, at which point interventions become complex and costly." The continuous data from wearables combined with occasional blood snapshots makes the difference: "Continuously collected smart-watch data capture fluctuations in activity, sleep and cardiovascular function over time that, generally, reflect the cumulative demands of metabolic regulation. By integrating these longitudinal signals with demographic characteristics and readily available blood biomarkers, the authors trained a computational model that identifies stable patterns linked to insulin resistance. The model demonstrated improved predictive performance compared with any single data source alone." Even if the measurements are noisy, the accumulation of many multimodal data points over time compensates for the noise. It is still a long way to clinical application, the results need to generalize, be robust and be actionable across a broad range of situations. But it is a promising first step.