In the workshop’s final session, George Isham, Senior Advisor at HealthPartners and Senior Fellow at HealthPartners Institute for Education and Research, asked the roundtable members, as well as the workshop’s other participants, to think about what they had heard over the course of the day and to consider the implications of those observations. He started the discussion by agreeing with the statement Matt Stiefel made at the end of the last session about reframing the conversation to make a broader impact on health and health care. “Is health the input? Does that foster engagement? Does it help with the rationale for why we think health is important?” asked Isham. “I heard some examples over the course of the day that made the point that economic development or education might be a better way to frame the discussion to get those factors that produce health into the conversation.”
The discussions of the day, said Isham, led him to think about the systems and the relationships purpose, measurement, data resources, and infrastructure required to capture data, and results and how all of those systems components will form an effective feedback loop to those who are on the ground working to change the system. How these components interact with one another, has not been described adequately in terms of the measurement system, he said. “Maybe there is more work we need in terms of thinking about how our fragmented, multilevel system with multiple stakeholders knits together and works more as a system,” said Isham.
A third point he took from the day’s discussions was that there is a fundamental conundrum between centralized and local data collection. On the one hand, he said, data and data collection systems are hard to develop and fund locally. On the other hand, the point was made throughout the day that metrics and data need to be relevant to the people in a particular community and that data collection systems need to engage local community members. Isham wondered if population health could borrow successful methods for what he called mass customization from industry, where there are standards for creating things that are then modified at the point of delivery to give customers exactly what they need. Assets and resources are built into this type of system to take advantage of the economies of scale of mass production.
Another aspect of this issue of standardization versus customization that struck Isham involved parsimony. “Does parsimony mean that you only get six measures for the country, or does it mean that any one project can pick six measures drawn from a set of 600 in order to engage all of the different perspectives to get the desired results?” asked Isham. Addressing that point, Meg Guerin-Calvert noted one approach to addressing parsimony, which was discussed at one of the tables during the World Café session, would be to identify the locally relevant data that most communities want and then identify gaps in that data that could be filled with community-specific measures. She recounted that several World Café participants noted that high-quality BRFSS data do not go below the larger metropolitan areas in terms of yielding robust and reliable data, and that perhaps be some specific metrics in a set of hundreds would be useful to extend those data to the level of community when needed.
José Montero from Cheshire Medical Center/Dartmouth Hitchcock Keene commented that he found the World Café session to be “an incredible event” that enabled a broader set of ideas about the importance of bottom-up approaches to metrics to emerge from the workshop. He said that although the epidemiologist in him believes the top-down approach is the right way to design metrics, the politician in him realizes that measures need to be designed with input from the community to produce data that respect local values and meet local needs. “That does not mean that you are going to change a measure or completely redesign it, but you will transform it to a level where communities can then use those data,” said Montero. He also wondered if the academic community, himself included, was taking so long to design the perfect measure that change will never happen. What he hoped, based on the workshop’s discussions, was to have some sets of advice on how to modify centralized measures to be useful at the local level and that can enable connections between public health and other sectors of the community. “From the discussions today, this call to action to make data and measures actionable by engaging the
community so they are empowered to take them as their measures and act on them, is something that we need to get to fairly soon,” said Montero.
He also recounted remarks at a prior roundtable workshop about how “health in all policies” language could be interpreted as health imperialism by other fields, and that perhaps the concept of well-being was a more inclusive one. “We need to acknowledge that there are many other tables out there and it is not that we are bringing other people to the health table. We should be building a bigger table where we can talk about these issues with other sectors, and we need to do that soon,” said Montero.
Julie Caplan, who leads California’s Health in All Policies Task Force and thinks about the concept of a culture of health and making health a shared value, said a big question that she would like to see explored has to do with whose job it is to gather data and how to develop cross-sectoral systems for using metrics and indicators. An example of how this comes up in her work, she said, is that she hears from stakeholders that they would like to know how many children walk or bike to school. California, said Caplan, does not have a statewide system for tracking how children get to school, and the schools that do track this do it in a variety of ways. Her team is now forming a multiagency task force to look at this question and identify who holds the data, and if nobody holds the data, to determine how to develop a system that spans education, public health, transportation, land use planning, and other sectors and is useful to all of them. This activity, said Caplan, leads to bigger questions: “How do we finance this kind of work, how do we lead it, and how do we build the relationships to make such a system a reality,” she asked.
James Knickman from the New York State Health Foundation noted that he thought the RWJF’s approach to metrics was sound and that it could be an asset to the population health field. After sitting at a number of tables during the World Café session, he concluded there is a need for collaboration on metrics to make them affordable and common. “We need an affordable health interview survey that can be done at the community level so that we can find out from people what is going on there. We need sensors and other approaches to more efficiently measure physical activity, purchases of healthy food, and those types of activities,” said Knickman. He also said he believes that choosing which outcomes are important can be community-driven using a menu of factors that can be affordably measured. As a final comment, he noted that he had been struggling with the question of whether the field should be pushing forward with new population health approaches now or if it is still in a learning phase that needs to be informed and driven by metrics to ensure that right approaches will be taken. “All of this will take energy at the community level, and if we do it once in 10,000 communities and it is not the right thing, are we going to be able to do it a second time in 10,000
communities?” asked Knickman, who said he vacillates on this issue. As a funder of programs in nine communities in New York State, he said that it is important to learn from all the efforts that others are funding and to develop common methods of assessing the effectiveness of these programs.
Steven Woolf said, “our ideas of what metrics are available are stilted by the ways we are accustomed to collecting data.” Social media, for instance, provides an alternative to the traditional methods of collecting data using surveys and other traditional instruments. “We need to modernize our thinking about the menu of data sources available to us,” said Woolf. As an example, a smartphone app exists in many communities to tell commuters when the next bus will arrive; he wondered if these apps could provide data on access to other public transportation, a domain that can be approximated only through conventional household surveys. Woolf discussed an example he heard recently from a colleague of how technology could be used in a new way. Most ambulances, when idle, sit outside hospital emergency departments or at the fire station, and most 911 call centers know the exact location of each ambulance. “Why not position those vehicles in locations where the highest number of trauma cases are, but also where the greatest medical needs are?” asked Woolf. “Not only would that reduce transport time, but also allow ambulance crews to stabilize people, perhaps without even needing to go to the hospital.” That kind of creative thinking, he said, builds on the existence of new sources of data that are not being used to their fullest potential. “You do not need a traffic survey to tell you which streets are busy because Google Maps now tells you that. Those types of datasets could liberate us to pull metrics that we think are important to our goals rather than being constrained by the traditional ways in which we have collected data,” he said.
Isham noted that there are now apps that can provide wait times at urgent care centers, information that is likely to be useful in other ways. Flores added that collaborating with the technology industry offered many opportunities to change the way data are collected and improve the timeliness with which those data are collected. A participant endorsed the idea that there is great promise and potential in the current technology ecosystem to offer data that can provide information on context about what drives health in communities. She also cautioned that the technology sector is characterized by an extraordinary amount of hype, sometimes without much substance behind it.
Trowbridge stated that population health needs to engage the technology community with regard to its focus on what is called quantified self-movement, the drive to use technology to gather data on individual health parameters. The current emphasis in the technology industry is
on individual health, he said, and thus there has not been much thought about how to use these tools or the data coming from them at a population health level. However, he emphasized the need to recognize that the technology community has almost unlimited potential when it comes to developing tools, but it needs guidance to know what to make. He noted that tools such as the Apple ResearchKit were not developed overnight. “It is going to be difficult to guess the exact right tool to develop because technology moves too fast. Instead, I think it is best to think about what you want to do with a sensor and then tools will evolve rapidly.”
Providing a perspective on the goals of the technology industry, Bhatia said it is not interested in health, but is interested in an irreplaceable, scalable business model with a 10-fold or 100-fold return. “Health care is a $2 trillion beast, and the opportunity is there.” He believes that harnessing the power of the technology community is a challenge the public health community should accept.
Veronica Shepherd called on the population health community to start thinking about which people are not yet at the table, particularly when talking about disparities. “I would like to suggest that there are stakeholders that are doing very hard work creating their own measurements to help shift how people live healthy and well, and they need to be at these conversations on creating shifts in measurements,” said Shepherd. What she has found most important with members of her community is that when she and her colleagues from public health approach them humbly and with respect for the local culture, they learn so much more about how to help people live better. Isham noted that having this workshop in Oakland, rather than in Washington, DC, and using the World Café approach to dialogue gave the conversations a different character than usual.
Steven Teutsch commented that some of the ideas that he heard throughout the day on how communities want to use data were of the “low-hanging fruit” variety, “but the evidence base about what moves the needle in social and environmental health is unbelievably poor.” The reason for that, he said, is the underinvestment in the trans-disciplinary research that would provide communities with information on how to raise high school graduation rates or reengineer the transportation system to reduce disparities and improve access. Studies to produce data that would enable those kind of systems changes are costly and complicated, said Teutsch, and he does not want the nation to look back in 10 or 15 years and say that “We gave it a good a shot, but we acted on an insufficient base of knowledge.” Population health, he said, is going to have to struggle with the issue of deciding when there is enough information to help communities move forward, which in turn, will take an investment in looking at how the interventions that are being taken work
in practice. Isham noted that the next roundtable workshop would be on research and he asked the participants to send examples of where research is needed to the roundtable staff.
In response to Teutsch’s concern that population health may act before it has enough data, Bhatia said the problem is that population health does not have a business model for health, only for sickness. Pharmaceutical companies can take risks with the drugs they develop because they have a business model that accounts for failure. “We are going to have to experiment and evaluate, experiment and evaluate,” said Bhatia, much the way that the technology industry operates. “I think there are principles on how technology operates and startups operate that we should bring into the practice of health,” he added. Mary Pittman from the Public Health Institute agreed there is not a good business model for health and suggested a few components for such a model: equity, policy drivers to improve equity, and measures of the cost of inequities; quality; reorganization of health care to reduce the costs of the current chaotic approach to health; and investments to improve population health with a return on investment metric.
Kelly Worden, responding to Bhatia’s comment about the lack of a business model for health and Teutsch’s concern about acting without a sufficient knowledge base, pointed to the need for process metrics that was mentioned in the morning’s discussions. From her work developing tools that architects and real estate developers can use to assess the health implications of their activities, Worden learned that traditional population health and health care industry metrics are not appropriate for built environment settings. “It might be easier to measure the actual process to determine if we are going down the right path instead of waiting for that ultimate health outcome,” said Worden.
Marthe Gold agreed with earlier comments that the discussion has to be broader than one about a culture of health and that involving other sectors could help attenuate some local problems resulting from resource starvation. She added, “I think the roundtable needs to begin to hear the messages we have been hearing over the last couple of years to change our terminology, maybe even change our name.” Pittman noted that there is a World Happiness Report (Helliwell et al., 2015) that frames these concepts much differently and that can provide lessons for the population health community. A good idea, she added, might be to see what other countries are doing well with respect to well-being and see what might be applicable to the United States. Israel Nieves-Rivera added that hopefulness could be a good concept to add to any expended idea of health.
Gold then suggested that it may be time for a foundational demonstration that would take a set of indicators that are largely viewed as being useful to different sectors, ask many different communities to use
this indicator set, and see what the communities do with them and the resulting data. If the results are good, these indicators could then be taken to scale. Gold also thought there are opportunities for collecting data on social determinants and community health needs under the provisions of the Affordable Care Act and in the community benefit provisions of the tax code. Nieves-Rivera noted that the population health community has not done a good enough job developing performance measures to determine how well interventions meet the needs of a community, and he suggested that the roundtable might want to drill down more on the connection between performance measures and interventions at the community level.
Abigail Kroch voiced her concern that the dialogue about population health metrics is centered largely on clinical measures and data sources, such as the electronic medical record, even though the clinical population is not the general population. In the same way, the population that uses technologies is not the general population or representative of the population that experiences the biggest disparities. “I would caution that as the idea of population health moves into the clinical setting, we are going to be moving away from the populations that need us most,” said Kroch. She suggested that where the field needs to be moving is toward an ability to demonstrate change in communities.
Judith Monroe from CDC commended RWJF’s 20-year commitment to its Culture of Health initiative, but noted as a point of reference that Native American tribes make decisions on a seven-generation timescale. She said she agreed with the idea of looking at ways of using the data coming from smartphone apps and other personal technologies and she supported the idea of a Zillow-like app, mentioned during the World Café discussions, that would parse data by neighborhoods. Monroe then proposed that public health needs marketing metrics, relationship metrics, and measures for unintended consequences, such as the poor health outcomes seen today in Eastern Kentucky that are the consequence of policy decisions made decades earlier.
Thomas LaVeist, commenting on the concern that population health puts too much emphasis on health care in the clinical setting, said the reality is that health care is where the United States allocates significant financial resources and the way those resources are deployed has a disproportionate impact on population health. In his opinion, there has not been enough discussion about how to infuse the health care model with a population health perspective. For example, he noted that the way in which the concepts of personalized medicine are rolling out in the United States is largely pharmaceutically centered, even though there is another piece that has a population health component. “Patients come into the system from a context, a community environment in which they
are living, and there are pieces of data that come with them from a social determinants framework that impact their ability to respond to medical treatment” said LaVeist. The problem today, he explained, is that adding data from population health measures to the electronic medical record will not help the clinicians make decisions relevant to a specific individual because not enough is known yet about the connection between the data generated by population health measures and how a patient will respond to therapy. “That is where we need to start developing protocols,” said LaVeist. “How do we get that information about community context into the health care system? How do we then educate the health care providers about what to do with that information?” On a practical matter, research to address those kinds of questions could draw on the financial resources being devoted to the personalized medicine enterprise.
Isham, in the workshop’s concluding comment, agreed with the idea that those working to improve population health need to develop new ways of tapping into the enormous resources available. “We need to think about how to do that and how to emphasize, ultimately, that action pathway and those interventions that lead to a more appropriate allocation of resources so we can create this better health in our community.” Following that remark, Isham adjourned the workshop.