Design of the revised thriving together model

ReThink Health, an initiative of the nonprofit Rippel Foundation, first created the Thriving Together Model in the summer of 2020, with subsequent improvements to incorporate additional data, research, and suggestions from several hundred early users.

The TTM is a system dynamics simulation model [14, 15] that helps stewards negotiate tradeoffs as they navigate toward an equitable, thriving future. It is grounded in Elinor Ostrom’s Nobel Prize-winning work on shared stewardship of common resources [16, 17]. The current TTM closely matches the conceptual framework of the Thriving Together Springboard [8] and incorporates concepts such as targeted universalism [18], multisolving [19, 20], race-related stressors [21, 22], and the organizational capability trap [23], all of which have implications for investing in multiracial well-being. In general, the Thriving Together Model brings greater structure, evidence, and creativity to the tasks of negotiating investment priorities and playing out the likely consequences for multiracial well-being over time.

The TTM recognizes that all communities must contend with adversities: some are sudden like natural disasters and acts of violence; others last months or years like pandemics and economic recessions; and still others unfold over decades or longer like chronic diseases and structural racism. Everyone is adversely affected by these hardships, but certain groups are typically harmed more than others, especially members of marginalized racial/ethnic groups. The cumulative toll of these adverse forces ultimately affects one’s length of life (measured by life expectancy at birth) and life evaluation (measured by the Cantril Self-Anchoring Scale) [5]).

The revised TTM encompasses all of these topics. It equips stewards with a simplified but realistic representation of well-being dynamics in a multiracial population, including four main drivers of well-being at a community level, which in turn affect states of individual well-being (i.e., thriving %, suffering %) and life expectancy.

Investment priorities in the TTM involve allocating community assets (both funding and in-kind resources) among the following four drivers of well-being in a population:

• Vital conditions: Investments in this driver help to establish the vital conditions that everyone needs to experience health and well-being [9]. They include a thriving natural world, basic needs for health and safety, humane housing, meaningful work and wealth, lifelong learning, reliable transportation, as well as belonging and civic muscle, which is central to them all. In the TTM, belonging and civic muscle is represented separately because of its unique dynamic effects. The vital conditions framework renders essential concepts often described as “social determinants of health” in a form that is clear, concise, and built for concerted action across all sectors of society.
• Belonging and civic muscle: Investments in this driver enhance both belonging (i.e., feeling part of a community) and civic muscle (i.e., the power of people to work across differences) [4,11,24]. Belonging and civic muscle supports well-being directly, by enhancing social connection, and indirectly, by strengthening collective capacities (such as trusting relationships, power sharing, constructive nonviolence, and civic organizing) that, in turn, enlarge the pool of investable assets (both financial and in-kind) in a community.
• Fairness in system design: Investments in this driver help to assure just and fair treatment for all people as a matter of system design versus systemic exclusion across color, class, gender, ethnicity, and other lines that often divide population subgroups. Fairness in system design explains the pattern of opportunities and outcomes among people in a multiracial and multicultural society.
• Urgent services capacity: Investments in this driver help to meet the changing demand for services that alleviate life crises or urgent needs among those who are struggling and suffering [9]. They include acute care for illness or injury; addiction treatment (including overdose rescue and detoxification); crime response; environmental clean-up; homeless services; as well as unemployment benefits and food assistance. The TTM calculates the adequacy of urgent services over time by comparing urgent services capacity to changing levels of urgent need.

An analysis using an earlier version of the TTM explored how stewards in any given US community could respond to an unjust shock over the course of ten years [13]. Results suggested that an immediate emphasis on belonging and civic muscle coupled with fairness in system design could turbocharge virtuous dynamics and create an equitable path for renewal within a decade.

The revised model used here to study the ELTRR plan does not address short-term responses to a shock, but rather what it takes to move beyond an unacceptable status quo by embracing new priorities for well-being and justice. This can happen by strategically reallocating investment priorities among the four drivers, as in the original model, and by intentionally designing multisector actions to advance multiple goals at once, a practice known as “multisolving” [19, 20].

In addition, the revised model shows how thriving, suffering, and life expectancy vary by race/ethnicity [21, 22]. The updated model estimates the extent to which new priorities have the potential to generate not just more well-being overall, but also more equitable well-being in a community that begins with an entrenched pattern of multiracial inequity.

Our analysis of the ELTRR plan does not attempt to forecast exact outcomes in the future. Rather, it explores relative changes that could unfold over time in American communities if they were able—in cooperation with federal agencies—to move in the direction implied by the ELTRR. Because the ELTRR was designed to be a single unified plan, we regard all 68 programmatic recommendations as a cohesive package of new priorities.

The next section describes the revised model structure in greater detail, along with assumptions about initial conditions, and how we quantified changes to represent priorities described in the ELTRR.

Model structure

Fig 1 shows the overall dynamic structure of the Thriving Together Model. Readers may also refer to the supplement for a complete list of equations and numerical assumptions, including uncertainty ranges for 11 coefficients and time constants. The supplement also contains results from sensitivity tests of single parameters, and from 5,000 random Monte Carlo simulations across all 11 uncertain parameters allowed to vary simultaneously.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Overview of the Thriving Together Model.

https://doi.org/10.1371/journal.pone.0319320.g001

Most of the dynamic structure remains the same as in the original model [13], including the four community-level drivers of well-being, each measured on a scale of 0–100%: vital conditions, belonging and civic muscle (BCM), fairness in system design, and urgent services capacity). Each driver is modeled in the aggregate, not broken out by subtypes. As before, vital conditions are the most important driver of individual thriving, with fairness and BCM also contributing. Suffering can be reduced to some degree when thriving improves (which reduces the amount of urgent need) and also through a greater adequacy of urgent services.

Average life expectancy is improved by greater thriving [25] and by more expansive vital conditions [26, 27], but it is reduced significantly by the extent of suffering [28]. The effect of vital conditions on life expectancy reflects all those mortality risk factors (like smoking, poor diet, physical inactivity, and slow-acting environmental pollutants) that are not immediately reflected in reduced well-being (struggling, suffering) and which show up, by and large, years later in the form of chronic diseases such as heart disease, diabetes, and cancer. More than 80% of lost life-years in the US are due to such noncommunicable diseases [29], with an average progression or latency period we have estimated at 15 years.

Also central to the model is the idea that the total pool of investable assets at any given time can expand in proportion to the community’s level of belonging and civic muscle. This ability to enlarge the asset pool is reflected in the reinforcing “expanding the pie” loop in Fig 1. The model also includes reinforcing feedback loops by which improvements in vital conditions or in fairness tend to boost BCM (and thus, the pool of investable assets) by bringing people in as active contributors to the community and drawing forth more resources through taxes, philanthropy, social entrepreneurship, volunteer efforts, and organized public work.

Estimating well-being differentials by race/ethnicity

A significant new feature of the revised model is its breakout of thriving, suffering, and life expectancy by race/ethnicity, based on an analysis of historical data from 2008 to the present. To examine differentials, we focused primarily on data for 2016–2018 (a relatively recent pre-COVID period with large sample sizes in the survey data) to guide our estimates of starting conditions for the model. Table 1 presents a summary of these data along with a few other metrics that help to interpret the observed patterns of multiracial thriving and suffering.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Selected US averages informing assumed differences in thriving, suffering, and life expectancy by race/ethnicity: Black; Hispanic; Asian; White; American Indian, Alaska Native.

https://doi.org/10.1371/journal.pone.0319320.t001

For thriving and suffering, we analyzed historical trends using data provided by the Gallup organization broken out for Black, Hispanic, Asian, and White subgroups. Unfortunately, Gallup did not report data for American Indian/Alaska Native or Native Hawaiian/Pacific Islanders [30].

In 2016–2018, thriving was lowest for Black Americans and highest for Asian, with Hispanic and White between the two. One might expect these differences in thriving to reflect similar differences in socioeconomic status. Table 1 shows that during the late 2010s, Black Americans were indeed much worse off than White and Asian on poverty [31], while Black and Hispanic groups were worse off than White and Asian on high school graduation [32].

In contrast, Black and Hispanic suffering was lower than White during the 2010s, and Asian suffering was the lowest of all. We found similar directional differences for two other metrics one might associate with suffering: suicide death rate [33] and the prevalence of substance use disorder [34]. In both cases, White Americans were worse off than Black, Hispanic, and Asian; only American Indian/Alaska Native people had higher rates than White on these measures.

For life expectancy, we analyzed annual data from the National Vital Statistics System on average life expectancy at birth broken out by race/ethnicity; 2019 is the first year that included data for Asian and for American Indian/Alaska Native [35]. Life expectancy for Black Americans was consistently four years less than that of White; Hispanic Americans were about three years greater than White; and Asian Americans were about six years greater than White. American Indian/Alaska Native people were lowest of all, about three years less than Black.

Modeling future changes in well-being differentials

These historical differences provide a starting point for the model at Year 0, prior to any ELTRR implementation. Changes over time in life expectancy differentials between racial/ethnic groups are attributed in the model to changes in vital conditions and fairness. In particular, an improvement in vital conditions across the community is presumed to lead to smaller differences in suffering and life expectancy between racial/ethnic groups [27,36]; and an improvement in fairness is presumed to lead to smaller differences in thriving.

The idea that greater fairness in a community reduces racial/ethnic inequity in thriving was supported by statistical analysis. We had data from several national Gallup surveys for 2007–2021 on perceptions of how Black and Hispanic people are treated and their equality of opportunity [37]. Linear regressions revealed that the differences in thriving by race/ethnicity were strongly associated with a composite of three of the Gallup fairness measures: “agree Blacks treated fairly”, “agree Blacks have equal housing opportunity”, and “agree Black children have equal education opportunity”. We used this composite to configure the model’s initial estimate of fairness for the US at 54%. We estimated coefficients for the effects of fairness on thriving for all people of color in the model based on the regression results.

Model initialization

We updated estimates for the initial values of overall thriving, suffering, and life expectancy (for all race/ethnic groups combined), as well as the four community-level well-being drivers (Table 2). In all cases, we used national US data from the late 2010s and early 2020s (as available) to initialize the model.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Initial Values and Sources for Well-Being and Well-Being Drivers.

https://doi.org/10.1371/journal.pone.0319320.t002

Modeling multisolving and gradual investment implementation

Another new feature of the revised model is a multisolving ratio, a measure of investment effectiveness that reflects the extent to which investments in one vital condition have additional benefits for other vital conditions as well. For example, an initiative to locate new affordable housing close to public transportation could help disadvantaged people fare better in at least two ways simultaneously, not just one. We represent such an initiative in the model as an increase in the multisolving ratio (see Supporting Information for calculation procedure).

A final new feature is the use of gradual linear ramps for the implementation of any new investment strategy. In the original model, with a time horizon of only 10 years, the emphasis was on quick implementation, acknowledging that shifting priorities might result in some unintended suffering in the first few years [13]. In the revised model, with a time horizon of 25 years, stewards can proceed more gradually at first to minimize unintended suffering. For this analysis, we have assumed a 5-year period for investment strategy ramp-ups.

Estimating multisolving and asset allocation for the base run

The revised model is initialized in steady-state equilibrium, meaning that all output variables in the base run remain at their initial values. It is important to establish realistic values for all baseline parameters, so that alternative investment strategies (with new parameter assumptions) can be usefully compared. Informed by our fieldwork with stewards across the US, as well as by exploratory analysis of several ELTRR scenarios, we have tentatively settled upon a baseline value of 1.15 for the multisolving ratio, and a typical asset allocation split as follows: 50% to urgent services capacity, 30% to vital conditions, 10% to BCM, and 10% to fairness.

Estimating multisolving and asset allocation under the ELTRR for model testing

We reviewed published descriptions of the 68 programmatic recommendations in the ELTRR [10] and for each one recorded: (a) its primary focus area, (b) any secondary focus areas listed, and (c) whether or not the description included an explicit focus on “equity”, “equitable”, or “fair” implementation. These terms are used frequently in the ELTRR document and describe an emphasis on expanding opportunity for racial/ethnic groups or other marginalized groups, similar to our concept “fairness in system design”, which more precisely differentiates a program’s means from its ends.

Table 3 summarizes the counts of the ELTRR’s recommended investment priorities, with one row for each of the seven primary focus areas and columns indicating the secondary focus areas along with fairness. For example, the first row indicates that there are 10 recommendations in the ELTRR that focus primarily on a Thriving Natural World, with all 10 of those also secondarily benefiting Basic Needs for Health and Safety, 5 benefiting Humane Housing, 1 benefiting Meaningful Work and Wealth, and so forth. The last row gives totals across all 68 recommendations, indicating a large number of secondary benefits: 103 for the first six vital conditions, 37 for belonging and civic muscle, and 38 for fairness.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. ELTRR recommendations category counts by primary and secondary areas.

https://doi.org/10.1371/journal.pone.0319320.t003

We used the information for each of the 68 recommendations to estimate what the adoption of the full package could mean for equitable well-being in communities across America, represented here using national averages. We considered all recommendations together as a package, because that is how they were intended, not as a menu of independent initiatives. Accordingly, we developed an approach to estimate an overall multisolving ratio and an overall allocation of efforts across Vital Conditions, BCM, and fairness (see Supporting Information). The investment allocation to urgent services is considered at a later step because the ELTRR did not address urgent services.

Applying these assumptions to the package of 68 ELTRR recommendations produced an overall multisolving ratio of 1.30, as well as somewhat higher allocations to BCM and fairness.

To get the full four-way split of investment priorities required for the model, we also needed an assumption for the percentage of investment going to urgent services capacity. Table 4 presents two possibilities. One scenario (labeled “ELTRR_Urg50”) is to keep that percentage at 50% as it is in the base run. Another possibility (“ELTRR_Urg35”) is to reduce the urgent services allocation to 35% -- a relative reduction of 30%. A previous analysis of investment scenarios across 39 large urban counties [44] indicates that one could reasonably expect to reduce the need for urgent services by 29% after the first five years of enacting a well-designed portfolio of investments in vital conditions and BCM, with still more reduction after that. If one could reduce the asset allocation to urgent services by 30% over five years with little or no adverse effects, it would free up significant assets to apply the ELTRR approach more fully, perhaps leading to even greater well-being and longevity over time.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Assumptions for three simulation runs. Changes from the base run in ELTRR_Urg50 and ELTRR_Urg35 are ramped up over five years starting in Year 1.

https://doi.org/10.1371/journal.pone.0319320.t004

Plausibility of results

This model-based analysis is not a forecast but rather an exploration of what changes might be possible if federal agencies and communities across the US were able to move in the direction defined by the ELTRR. This study does not attempt to anticipate future events that could damage population well-being (such as economic recession, natural disasters, or violence). A prior paper explored the dynamics of responding to these sorts of severe shocks [13]. Instead, this study reveals why and how strategic investment priorities could influence multiracial well-being over time. Still, it is fair to ask whether the scenario-driven projections of a greatly improved future are plausible.

First, consider the projection of overall thriving, which climbs at the rate of 4–6 percentage points per five-year period from Year 5 to Year 25—an increase of 20 percentage points over 20 years. A study of thriving and suffering across all US counties [48], comparing the period 2008–2012 (midpoint 2010) to 2013–2017 (midpoint 2015) found that ten or more counties improved by 10–15 percentage points over this five-year period. We do not know how long such rapid improvement could continue, but these data show that the model’s rate of ascent is credible at least on a five-year basis.

Next, consider the model’s projection of overall suffering, which declines from about 4% initially down to 1.4% by Year 25. The same county-level study cited above [48] found that during the period 2013–2017, the four counties with the lowest percentage suffering were in the range of 0.4% to 0.7%. This result establishes the plausibility of a very low suffering percentage over at least a five-year period.

Finally, consider the model’s projection of life expectancy, which by Year 25 climbs to 81.5 years overall, and more than 83 years for the Asian and Hispanic subgroups. No state in the US has an overall life expectancy of more than 80 years [49], but there are individual counties with estimated life expectancies as high as 92 years [50]. Moreover, in 2019, many OECD countries had overall life expectancies of 82 years or greater, with Japan (84.4) and Switzerland (84.0) at the top [51].

Significance

Although the Thriving Together Model remains a work-in-progress (with possible extensions discussed below), this exploratory analysis contributes in several ways to our understanding of collective well-being and the dynamics that drive potential improvement over time. Starting from such key concepts as shared stewardship of common resources [16, 17], targeted universalism [18], multisolving [19, 20], race-related stressors [21, 22], and the organizational capability trap [23], we built a quantified, testable simulation model that encompasses them all and closely matches the conceptual framework of the Thriving Together Springboard [6].

Through a series of systematic tests, including sensitivity analyses across multiple uncertain assumptions, we demonstrated the importance of three reinforcing feedback loops (shown in Fig 1) that each build Belonging and Civic Muscle and, in turn, “expand the pie” of investable assets in a community.

We also showed how investments to assure Fairness in System Design along with efforts to establish Vital Conditions are critical when seeking equitable gains among those racial/ethnic groups that are most marginalized and thus have the most to gain.

Moreover, we identified a highly impactful transition path, defined by the ELTRR_Urg35 scenario, in which the investment in urgent services capacity is gradually reduced, over five years, to make way for more investment in all other priorities. This scenario breaks the system out of a “capability trap” [23] (marked by an overreliance on urgent services) and incurs only a relatively small, temporary rise in suffering in the short-term on the way toward persistently better results afterwards. Finding that transition path helps us to see more generally that it may be possible to escape capability traps without confronting an unacceptable worse-before-better tradeoff.

The ELTRR_Urg35 scenario manages this by increasing the multisolving ratio and moving away gradually from a primary emphasis on urgent services. The moderately lower allocation to urgent services capacity (of 15 percentage points or 30% over five years) would not require dramatic closure of urgent care centers or immediate loss of safety net services. Instead, it could be accomplished by adjusting future allocations so they do not automatically replenish urgent service capacity in the same amount as before. Under this scenario, when urgent service contracts come up for renewal, approximately 6% per year could be sunsetted to achieve a 30% reduction over five years; and the sunset rate could be even lower if it were possible to deliver services more efficiently.

Bringing the ELTRR scenario to life

Responding to the severity and unjust consequences of the COVID-19 pandemic, the ELTRR recommendations spell out unprecedented changes to routine business practices across scores of federal agencies – all oriented toward creating the vital conditions that everyone in the US needs to participate, prosper, and reach their full potential. The plan also commits to assure fair and just opportunities across all racial/ethnic groups, among other aspects of social equity. By design, the plan’s recommendations are within the power of each agency to enact immediately. They do not require any new legislative appropriations or new federal authorities. They only require the will and wherewithal of federal staff and partners across the country to see the potential for meaningful improvement in the lives of Americans and to bring these recommendations to life.

To make the ELTRR a reality, collective and harmonized interagency action needs to continue to expand across all parts of the federal government, as well as with state, local, territorial, and tribal governments and civil society partners. Leaders need to empower their staff to define outcomes of their efforts in terms of improvements in vital conditions and to seek collaborations that bear fruit for communities. For example, leaders of the Office of the Assistant Secretary for Health in the Department of Health and Human Services are both guiding development of initiatives and expanding strategic collaborations through an overt focus on “People and Places Thriving” (PPT) [12]. Officials from the other 46 agencies comprising the interagency PPT workgroup are making similar commitments. Together, they are introducing procedures to track progress, ensure active learning, and generate new evidence about the value of working through shared stewardship. The federal PPT plan is influencing numerous whole-of-government initiatives related to environmental justice, resilience, and well-being. Furthermore, state, tribal, territorial, and local leaders are beginning to craft their own complementary action plans seeking to make the most of federal connections and resources.

To realize the brighter future that this study reveals, stewards in every sector and every walk of life must move beyond the familiarity of conventional programs, policies, and investment priorities. Instead, they ought to devise a new generation of place-based and people-centered strategies that unlock the multisolving power of Belonging and Civic Muscle, Fairness, and Vital Conditions. Even now, innovators in communities across the country are beginning to craft community-led investment agendas with these features. See, for example, the recent report on, “Investing in Generational Change” produced through the award-winning work of BeWell Palm Beach County [52]. And they are not alone. Many other multisector stewardship groups are also beginning to negotiate regional well-being portfolios that fit their own aspirations, circumstances, and capacities [53, 54].

As more and more changemakers join a rising movement to thrive together, they may use the investment framework – and compelling results – from this study to inform their negotiations and commitments.

Opportunities and extensions

The Thriving Together Model can be used by anyone who wants to explore what it takes to make a great stride toward multiracial well-being. Results reported here come from a model configured with nationwide data. However, stewards across the US may adjust the model’s starting conditions to better represent their own regions. That kind of regional configuration is possible with appropriate data and resources and has already begun with collaborating organizations representing states (e.g., Delaware) and multi-county regions (e.g., in southwestern Texas).

Although the model could, in theory, represent any number and type of population subgroups, this study reports results across four racial/ethnic groups. Other sociodemographic characteristics could be considered, such as age, educational attainment, or urban/rural residence. However, current data systems do not provide sufficiently detailed information to model multiple, overlapping characteristics. We chose to concentrate on multiracial well-being because it affirms an important, commonly understood idea of what it would mean to move toward an equitable future in which everyone has a fair chance to reach their full potential. Regrettably, we could only study four racial/ethnic groups because current data systems often exclude others, such as American Indian/Alaska Native and Native Hawaiian/Pacific Islander. These groups ought to be included in nationwide and regional data systems.

The ELTRR was designed to be a cohesive, interagency plan with 68 specific recommendations that we modeled as a single package. Future analyses could explore the potential contributions of smaller subsets of initiatives.

We may also consider disaggregating the Vital Conditions that we have so far modeled as a single combined index. Our Vital Conditions Index reflects an equally weighted average across six vital conditions and implicitly assumes that all six conditions move together over time and do not conflict with one another. That is a reasonable assumption supported by a prior disaggregated analysis [44]. Even so, there may be circumstances where it is worthwhile to track each vital condition separately.

Finally, future studies may examine not only the size of relative investments, but also how -- and by whom -- those investments are implemented. We have started to model and explore two such questions in particular:

1. What is the balance of lived/learned expertise? There is a growing consensus that investors ought to rely on a balanced mix of lived and learned expertise [55, 56]. An excessively top-down approach (relying only on “learned expertise”) may promise quicker results but tends to erode Belonging and Civic Muscle over time. Conversely, an exclusively grassroots or bottom-up approach (relying only on local “lived expertise”) may strengthen Belonging and Civic Muscle but could overlook successful policies and practices established in other places.
2. What is the threat of backlash and how strong is the commitment to avert it? US history is filled with instances where progress toward multiracial well-being – or even the suggestion of it -- provokes backlash [57, 58]. One may expect that moves toward greater fairness reflected in the ELTRR recommendations might similarly be undermined by resistance. Stewards using the model may want to open frank discussions about the threat of backlash in their regions and what it takes to build the will to thrive together [59].