Who we are emerges out of the dynamic interplay of what we are (the person), where we are (the situation), what we say we will do (intentions), and what we do (actions). Medicine has built a model of disease with a focus on the person in a hospital or a clinic and their intentions. In this model, intentions serve as proxies of future actions (i.e. I will stop smoking) and census or survey data serves as a proxy for the situation. This model – partly borne out of the constraints in technology – has fortunately been successful in improving health outcomes in the modern world. However, this strategy appears to be coming to the limits of its effectiveness as the major diseases of modernity – cardiovascular diseases, diabetes, hypertension, and neurodegenerative diseases – seem to be recalcitrant models of disease with sporadic and episodic data streams. In order to make a larger impact in the outcomes of these diseases, a more comprehensive and integrative approach will be needed. It will require measuring the person beyond the situation of the hospital or clinic, supplementing intentions with actions, and complementing their conscious descriptions of internal states with subconscious signals from the body (next essay).
The Office of the National Coordinator for Health Information Technology defines patient generated data as “health related data created and recorded by or from patients outside of the clinical setting to help address a health concern.” Due to the decreasing costs of computation, transport, and storage of data, combined with the improved performance and diversity of digital sensors, there has been an explosion of devices that are recording behaviors such as activity levels, posture, location, sleep, and biometric data such as heart rate, heart rate variability, heart rhythm, oxygen saturation, hydration, and blood glucose levels. In addition, social media feeds provide photos and social network maps. In parallel to the movement of patient generated data, sensors are also providing real-time, high-volume community level data. Cities and neighborhoods are replete with sensors providing measurements of weather, pollution, crime, police activity, traffic, and food access. Rather than relying on long lag time data streams such as census and survey data, this sensor data provides near real-time multidimensional feedback of the environment.
If you extrapolate from the predictive successes of other complex adaptive systems, such as the economy or the weather, integration of multimodal data sources should lead to improved predictions in medicine. It is not an enormous stretch of the imagination, for example, to see how GPS data calibrated with blood glucose and hemoglobin A1c levels can not only help substratify the category of a diabetic patient, but also inform the design of interventions (future essay) to alter patterns of maladaptive behaviors for higher risk patients. Similarly, it is not an implausible thesis that neighborhoods with higher endemic violence as measured by real-time police activity data or sensor data recording gunshots, will not only have more emergency room visits for trauma, but also emergency visits for second order vagus nerve mediated effects such as cardiovascular diseases and metabolic diseases.
Knowledge acquisition and growth is often preceded by seeing the world with new perspectives. In science, technologies have facilitated the unveiling of features in the natural world leading to new insights, explanations, and predictions. The telescope revealed worlds far and wide, the microscope uncovered worlds small and near. Each unveiling hidden data streams leading to an adaptive radiation of explanations and predictions. Analogously, the sensorization and digitization of the situation and the person holds a similar promise. . As the computational social scientist Duncan Watts writes, “by rendering the unmeasurable measurable, the technological revolution has the potential to revolutionize our understanding of ourselves and how we interact.” The person with their intentions, actions, and the situation more completely measured, quantified, and integrated has the potential to have a positive impact in our struggle against the deleterious effects of chronic diseases.