Why We Need A New Platform for Population Health
If we only take away one lesson from the Covid19 crisis in the last few months (and I realize that there are many) it’s that all of our decisions have a higher level of quality and ultimate success when they are based on sound data (of which we clearly need more!) about people and groups, which has been especially true when assessing susceptibility to infection for example. “Population health” as a term (which is sometimes used interchangeably with “public health”) has been around for around 20 years or so now but has been only loosely defined to date. For example, one definition suggests that “Population health is taking responsibility for managing the overall health of a defined population and being accountable for the health outcomes of that defined population.” However, as short as it is, this definition excludes the many additional nuances offered by others such as:
· “Population health involves the health of the community; it implies wellness promotion as well as the treatment of new and chronic illnesses throughout the care continuum. It also implies improving the health of people previously undermanaged, such as the poor in terms of conditions such as diabetes, hypertension and cancer.”
· “Accountability for the health and utilization of health care services of a defined population of individuals across the care continuum, from preventative to acute to post-acute settings.”
· “Individual responsibility for physical, mental, spiritual and social health. When each person takes control of his or her health, it reflects on our families and society as a whole.”
· “Population health refers to addressing the health status of a defined population. A population can be defined in many different ways including demographics, clinical diagnoses, geographic location, Population health management is a clinical discipline that develops, implements and continually refines operational activities that improve the measures of health status for defined populations.”
· “Population health’ is both a means and an end. The goal is to improve the collective health status of the population at large in a given geographic area. That goal can only be accomplished through a combination of (1) behavior change, which has to be promoted in a tailored manner, using an array of appropriate tools and (2) evidence-based medicine focused both on prevention and treatment of injury and disease and on improving function and happiness for the individuals who make up the population.”
· “Understanding (measuring) the health of a defined population (community, covered life population, set of patients) to include all aspects of health (physical, mental, etc.) and to include the underlying determinants of that health (e.g., poverty, housing, nutrition, exercise, pollution) and, most importantly, working to improve the health of that population.”
· “Population health means taking an analytical approach to understand the health needs, disparities and outcomes of the community and to align improvement initiatives.”
Whichever one of the above you prefer (and it may be a combination), the pattern which emerges most from all of these is that we are attempting to gather relevant health data from a large population of people in order to better understand how smaller groups or ‘tribes’ of people may have common indicators of how to become healthy, stay healthy and recover from sickness, whenever it occurs. Put another way, population health research attempts to identify and gather data about specific groups in order to gain new insights about support and care interventions that can and should be made. These groups are often geographic populations such as countries, cities or defined communities, but can also be other groups such as a set of employees, particular ethnic groups, people who live in one or more zip codes, disabled persons, or in more recent times, Prisoners, People who live in nursing homes or People who contracted coronavirus, etc. The ultimate health outcomes of such groups are clearly of high interest and relevance to policymakers in both the public and private sectors. It should be noted, however, that population health is not just the overall health of a population but also includes the distribution of health. Overall health could be quite high in the majority of the population but possibly quite low in the chosen minority population. Ideally, such differences would then be eliminated or at least substantially reduced by better understanding the differences.
So, if the overall goal of population health is to gather data about particular groups, what might be the most common “unmet needs” or problems so as to best glean the best possible insights from assembling the data that we need? Below are some of the most common ones:
- Individuals and groups are complex and we need to understand and consider their full psycho/social/physical condition when making health and wellness interventions.
- We need to be able to capture an individual’s ‘social determinants of health’ data in records and track/update them over time to see how things might have changed.
- We need to enable ways for 3rd parties (including family/community members) to credibly input data into an individual’s health record (perhaps on an ‘as needed’ or continual basis).
- We need an easy-to-use, interoperable and ‘safe’ system or platform for capturing and analyzing data that links to public health databases to add context to individual data.
- We need to develop tools to make actionable recommendations on both an individual and population-wide level that can be easily used by all stakeholders (e.g. providers, payers, government, and others).
There are many others that we might add but even these show that we have many significant challenges which boil down to a number of core questions we could ask such as:
1. Can we clearly and effectively describe the ‘unmet need’ a new ‘open’ and widely usable population health system would need to solve for (even if it only did this progressively well over time)?
2. Which target group of ‘customers’ has the greatest need for a new potential population health system or platform and why? This includes individuals, health providers, Healthcare payers, Drug companies, Health interested academia and others.
3. From which existing sources does useful population Health data exist to be drawn together (and can we integrate it efficiently and effectively)?
4. What are the major gaps in population health data and why (missing or new data not yet captured)?
5. How might we best fill the gaps in existing population health data (Including data needed from, individuals, Healthcare Providers, Payers, Drug companies/Pharmacies, Labs, etc.)?
6. How should we best think about data collection over time (and what stays fixed or stable versus being variable or changing)?
7. How do we best ensure data quality, validity and efficacy as much as possible?
8. How do we design a system to help with the best possible process inter-operability-across all databases, devices, other inputs?
9. How might we best create readily accessible insights/report/graphs/dashboards from a potentially new and better-designed population health platform
10. How might we keep the ‘friction’ of population health data capture to an absolute minimum (and prioritize the collection of the most useful and predictive data over the less useful/predictive)?
These are by no means the only questions and there are many others that should be added (especially around who ‘owns’ such data, legal compliance and privacy, which includes carefully anonymizing collective data and public versus private use for instance). However, this list shows that there is plenty to think about in order to create a result that is likely to be truly valuable. Having said this, one useful way to think about what needs to be achieved in an overall sense is to consider assembling a rounded set of data about individuals each time that we define a target group to analyze or study. By ‘rounded’ we mean a data set that includes all the variables that are likely to have useful value to understand what might be enhancing or detracting from a person’s health. In general, this fall into 3 categories we often use: 1) ‘general background’ data (often basic demography), 2) Clinical or Medical data, and last but not least 3) Behavioral data (often called lifestyle or psycho-social). If we were to assemble the kind of individual data fields under each of these, we may produce a chart such as the one below:
There are many other individual pieces of data we could add to this, of course, and we could go deeper into even these areas too (an area such as “education” or “happiness” for example, could have many sub-sets). However, the goal of such a chart would be to ensure that data is collected in each of these individual areas and both on a once-off basis and then at regular intervals, where it is likely to change, positively or negatively.
What is interesting to me, and to my mind, is perhaps our biggest shortfall in population health data assembly and use is that even using these 90 or so data fields (which could easily double or treble) we barely capture one-third of them today and not even regularly. In fact, we collect about 50% of the Background data here, perhaps 35% of the Clinical/ Medical data and only maybe 15% of the Lifestyle/Behavioral data. This means that we are missing a huge swathe of useful data that with a little effort we could capture and track and make better wellness or health recovery interventions as a result. Perhaps even more significantly, most health experts (including the World Health Organization) would suggest that the Lifestyle/ Behavioral data category could be as much as three to four times more influential on health, and the capacity to thrive, than the other two combined. What is often somewhat pejoratively called ‘social determinants of health’ are generally poorly defined and understood but we know anecdotally they have a significant and long-lasting impact. This includes somewhat ‘harder’ or more easily measured factors such as people’s living conditions and much softer and less measurable ones such as people’s relative feelings of emotional stability or stress at a given time.
Rather than to declare this to be too difficult to tackle, I believe an effective approach to developing new thinking and a system to support it in this area is to measure a chosen population of people in microcosm. This may be a single population of people that have some common experience (covered by Medicaid for example) or with a common medical condition (such as COPD or high blood pressure) in a single geography like a city or even one zip code. This would allow us to apply many of the above additional data fields over a fixed time frame of say one year (using both medically gathered data and perhaps the ‘softer’ data gathered more remotely via tele-encounter perhaps). In addition, we could separately look increasingly at the most likely care interventions we could make with this microcosm population in both a predictive way and then by looking at the results after the fact and making adjustments accordingly. What is potentially most exciting here in prospect is that we can start to learn what data in combination seems to matter the most to help people thrive or factors that directly enhance health or detract from it when not considered.
A more sophisticated approach to population health data gathering will not solve all of the many problems that exist in the current healthcare system, but they will lay the solid foundation for our decision-making about what impact individuals and tribes of people most positively in substantial ways-this is therefore worth considerable time and attention.
Jon Warner is CEO of Silver Moonshots-www.SilverMoonshots.org, a research and mentoring organization for enterprises interested in the 50+ older adult markets with its own aging focused virtual accelerator. He is also Chapter Ambassador for Aging 2.0 and Co-chair of the SBSS “Aging in the Future” conference, in Los Angeles