Driving health discoveries with non-health data

By Stefan Agamanolis, associate director of strategic research programs, Weatherhead School of Management

In his article last March, xLab faculty co-director Erman Ayday discussed how novel medical insights can be unlocked by bringing together the increasing quantity of digital health data generated and extracted from multiple sources. He also described how building new technological frameworks will be essential to accomplish this sharing and merging of data in a way that preserves individual privacy and respects complex legal and regulatory schemes.

The potential for new health-related discoveries expands even further if these research endeavors could integrate non-health datasets from the myriad of entities that collect it about you. There is a concept in medicine known as “social determinants of health” that describes how your risk for medical conditions and events is impacted by all of the non-medical things happening in your life. For example, your diet and physical activity levels significantly affect your health, and your socioeconomic status impacts your access to healthy food and safe living conditions. Environmental factors like air and water quality can significantly affect your risk for various medical conditions. Social connections and support, or a lack thereof, can impact on your mental and physical health as well. It is broadly accepted that these types of non-medical factors account for the vast majority of your overall health risk, as much as 80 to 90% of it, according to Social Determinants of Health 101 for Health Care: Five Plus Five

However, we still lack a more granular understanding of how the moments making up our daily lives contribute individually or in combination to our medical journeys, and we lack a way to capture social determinant data in a more real-time fashion such that it could be possible to detect a deteriorating risk profile and offer help before you end up needing expensive healthcare services that could have been avoided. In fact, your healthcare provider is likely only getting a very crude and infrequent snapshot of your social determinant profile, if they get one at all. Some providers have adopted standard screening forms but these can be burdensome to the patient and they are likely only administered when you go to a medical appointment.

There must be a better way to capture or derive insights about this important 80% of our lives and use it to reduce health risks and propel well-being. There are many potential starting points. Supermarkets and other stores have insights about what products you use or consume in your household, and using ingredient or supply chain information, we could infer what harmful substances you may have been exposed to. Mobile devices and navigation apps know where you have been, and leveraging other datasets we can make inferences about environmental conditions or exposures you may have endured. The content and tone of your social media posts reflect something about what you are thinking about, your workload, and your mental state. This is just scratching the surface.

There are many potential sources of data, some freely available to the public and others tightly controlled by commercial enterprises, that may individually or collectively offer important hints about medical conditions you may be developing, medical events that you may be at a high risk for, and other health burdens that you may be carrying. But while there is tremendous potential value in consolidating these kinds of data for analysis, the key obstacle to overcome is obviously centered on personal comfort, and fostering trust in the robustness of the technological platform and the visibility and control it provides around how data is manipulated. xLab has been working for several years on an advanced platform to allow individuals to share sensitive information with other parties in a trackable, privacy-preserving fashion.

It may feel inherently uncomfortable to contribute sensitive information about our lives with outside parties to assist in health research and care coordination. But the reality is many tech companies and their affiliates have been gathering, sharing, and selling data like this completely legally for many years based on privacy policies and legal agreements that few of us read when setting up accounts. This data is crunched by proprietary algorithms to tailor our news feeds and target advertising, often for healthcare products and services, in ways that sometimes seem invasive.

As the public becomes more literate and empowered around the use (and misuse) of data they consciously or passively generate, only the organizations and brands with the strongest reputations for safeguarding personal privacy and information security will be in a position to curate the richest datasets and unlock the full potential of these new analytical techniques to drive the future of population health.