Why prospective data analysis has much potential for good

I’ve noticed an emerging trend in social determinants of health conversations. On the payer side, payers are less interested in “what” constitutes social determinants-related activity and how much is invested in said activities. Rather, payers are increasingly interested in measuring and enabling the effectiveness of SDOH initiatives to improve people’s health status.

Providers, on the other hand, are increasingly curious about exploring how SDOH data could be turned into insights. For instance, I recently spoke with the head of data analytics for a provider who was exploring potential health insights from voter registration data. Providers want to better understand their markets’ health needs so that they can be responsive.

And providers have SDOH data today sitting untapped in their EHRs. Note that Google has applied for a patent on a technology that predicts individualized-at-the-point-of-care clinical outcomes by applying deep learning to records from millions of patients’ EMR data. Using stored, aggregated EHR data from diverse patients, including clinical notes where SDOH data may be lurking, Google’s technology is intended to predict one or more future clinical events and summarize pertinent past medical events related to the predicted event.

The above examples require finding, acquiring and integrating publicly available data, syndicated SDOH, existing payer data and existing provider data. Of course, the devil is in the details. For example, there’s a crazy quilt of state laws pertaining to who can access what data elements of voter registration lists.

The theory is that if health professionals better understand the factors at-play in patients’ lives, we can proactively target our efforts at specified demographics to drive down healthcare costs and improve health outcomes.

Governmental policies and community-based programs determine our physical environments such as the quality of our drinking water, the housing available to us and our community’s transportation system. Policies and programs also play a major role in our social and economic factors—for example, the safety of our community, education available to children and adults, and employment opportunities. Access to quality clinical care is a major driver, as is our health behaviors—for example, the choices that are made about tobacco use, exercise and sexual behavior.

Listed in priority order, four social determinants drive our health outcomes:

• Socio-economic factors (education, employment, income, social and family support, community safety.
• Health behaviors (tobacco use, diet & exercise, alcohol and drug use, sexual activity)
• Clinical Care (access to, and quality of)
• Physical environment (air and water quality, housing, transit)

There are myriad examples of how social determinant data can be transformed into insights that improve health outcomes and reduce costs. For instance, the University of Wisconsin’s Institute for Research on Poverty has shed light on how existing health policies are incapable of responding to health needs of today’s US families. Most U.S. children will not spend their entire childhood living with both biological parents. Further, most children born to unmarried parents will live in complex family situations and experience family fluidity and parental multi-partnered fertility. Even after accounting for differences in resources at birth, father absence and family complexity and fluidity are associated with numerous sub-optimal outcomes as these children grow into adulthood – including greater mental and physical health problems. However, US health policies are designed for relatively static family structures,

So, what can “data to insights” do in this situation? One response (among others) is preventing family complexity in the first place. There is promising research – which requires additional rigorous studies – that indicates a sizeable potential for reducing family complexity by making long-acting reversible contraceptives widely available to women seeking family planning services.

Thanks to RWJF, we know that consumers are open to organizations using their data for forces of good. We can improve outcomes and drive down healthcare costs if we apply insights from their Data for Health Advisory Committee:

Data moves at the speed of trust. People want their data used for important and helpful reasons, while simultaneously being protected from invasions of personal privacy and breaches of their personal information.
Big Data: People recognize the potential to reveal meaningful insights about the livability and health of their community and to support analytics that can inform health practices.

Long Data: People want knowledge based upon tracking health data over time to see patterns, trends, and predict potential. And, they want their health information available when and where it’s needed.

Infrastructure Competition: The atmosphere within the provider and payer local communities of lack of trust and agreement inhibits communities to grow healthier.

Today, most analytics teams are designed for retrospective analysis, not prospective. While the shift to prospective analysis is underway, the need for retrospective analysis continues. As we move to the future, be sure to direct efforts towards forces of good. Our stewardship responsibility to the field we all serve demands that we be worthy of the public’s trust.