 Something that has been noted throughout this 25th anniversary interview series is that Americans probably don’t have the will or the capacity to do anything different in terms of lowering the cost of care or widening access to care and that institutions are slow to adopt change. Thinking about the coronavirus disease 2019 (COVID-19) pandemic, do you think that will continue to be the case going forward, or do you think there’s a chance that this could be a tipping point for public health?

 I definitely think it’s a tipping point. There’s no question that something material and hopefully positive will come out of this, but I think you’re asking kind of the more nuanced question, which is exactly how much of a difference will that make? So, yes, I do think that we’re at a tipping point; something will change. But no, I do not think that [it will be permanent]. Let’s say we fast forward a year and there’s a very safe and effective set of vaccines, and we get coronavirus cases down to 100 a day, something that seems reasonable. I think that then, people will unfortunately kind of fall back into old habits, and that includes the health care system.

I think that that will only perpetuate some of the very problems that we’ve seen come out during COVID-19. The only place where I think that that’s not going to necessarily be the case is public health infrastructure. Most county and state public health departments are the first ones to get attacked on the budget side, and I think now that we’ve had such a dramatic pandemic, they will not be the first ones to get cut. Now will that last long? I don’t know. But I think that certainly reinvesting in and rethinking public health infrastructure will be a high priority for everyone.

 One change that some people have predicted is an accelerated shift to value-based care and an increased focus on risk. In discussing the health care system, we saw that providers that were only in fee-for-service arrangements really got slammed. Do you think that could also change with an accelerated shift, or could things get in the way of that transition?

 I think it’s going to change but not for the reasons that we think it might change; it’s not because of COVID-19 and what you just cited around [how] fee-for-service providers got slammed and the hardest-hit small practices, etc. I think it’ll change because of Medicare Advantage, and if you look kind of on the complement, the practices that did well were the ones that were in capitated arrangements. You can call them risk- or value-based arrangements, but they were largely like Medicare Advantage, capitated-type arrangements. And I think because Medicare Advantage will grow and because managed Medicaid is so large, that will almost become a force function. It’s not because you’re going to see this rush from fee-for-service into alternative payment models, the way some of us, including people like me, thought we would have 10 years ago, but I guess it’s going to be because the payers are not going to pay for it. And that includes Medicare, because Medicare Advantage, if anything, takes people out of that fee-for-service treadmill. So, if you call Medicare Advantage or Medicaid managed care “value,” then yes, you’re right. I think there will be a kind of an acceleration, but I think of it more as there are some insurance-based structural elements that just make the move away from fee-for-service easier. I’m not sure if it’s value. That’s probably the politically correct way of saying it.

 Speaking of politically correct, as we’re getting closer to the election, one of the big changes the current administration made is with Medicaid work requirements. If there is a change in the administration next year and a new one comes in, what do you think would be a priority in Medicaid? Would it be a huge convoluted process to unroll that?

 Yes, a priority would be to unroll that, but it’s easier. It’s definitely doable. A Biden administration would 100% do it. Getting it out? The states have decided to take it up. Basically, CMS just made it easier to put in work requirements. It’s not like Medicare or CMS actually said that all states must do this, but they basically did what’s called a waiver template and guidance and notification. You would expect a Biden CMS to completely undo that, and maybe the part that would be a little bit harder is how do you get states like Indiana, for example—where [CMS] Administrator [Seema] Verma came from, where she helped to set up some of the Medicaid program—that wanted to do this on their own, anyway, to undo that? You’re going to have to override it with some overwhelming incentives: money, basically. And I don’t know if states would take that. If a state was kind of on the bubble and didn’t want to do work requirements and felt like they were getting, quote, pressure from the Trump administration, maybe that could be undone. My fear is that Biden will definitely make actions to undo it, but if the states have already adopted it, and they don’t feel like it’s been harmful, and the legislatures and governors in those states are the same political party, I don’t know what would motivate them to undo it, so I think it will be harder. But yes, I would expect a Biden administration to not just [undo] work requirements, but to put out a series of those same kinds of guidances and notifications that actually encourage expansion of Medicaid and to do some of the things that I think the Obama administration started, where they did quality measure harmonization, accountable care organizations. You’ll see some of those same familiar themes probably continued.

 In your mind, what have been some of the standout differences between the expansion states and the nonexpansion states in terms of population health in the Medicaid population?

 I think that expansion states realized that they, in addition to the expansion, needed to really embrace what I think most of us would roughly call delivery system reforms. California serves as a model for a lot of those states to some degree, where they focus on what’s called whole-person care or whole person–centered care. It’s what they call, in Medicaid speak, Money Follows the Person—you try to think about what’s the most comprehensive care you need to give to somebody, and then all the parts of Medicaid that are confusing and somewhat siloed and complicated, and you try to reorganize them to make that whole-person care achievable. I think expansion states have tried to do that. Expansion states like Louisiana, for example, really tried to also focus on some of the cost drivers, pharmaceuticals being one of them. As for the nonexpansion states, I’m from Texas, a nonexpansion state, and Texas actually did try to do some delivery system reforms, but [the reforms] haven’t really gotten the traction that they have in other states. Candidly, Medicaid in those nonexpansion states is still kind of what I would think of as the old Medicaid from a decade-plus ago, from before the Affordable Care Act (ACA). It’s much less managed care, or even when it is managed care, many of the providers really still function almost like fee-for-service doctors. There’s an element in the nonexpansion states where there’s, in a way, a lot more innovation that they could embrace, but because they didn’t expand, they don’t know. It’s a Catch-22. They don’t have the burning platform to try to control costs because they didn’t expand, and at the same time, [because] their providers are traditionally the lowest reimbursed, especially like for childcare, they’re not going to be incredibly motivated to do things that take a lot of effort and investment, when they are not receiving the benefits of that investment. So that’s partly why expansion vs nonexpansion becomes a pretty critical issue if the administration changes; the Biden administration will have to do things to get those states to meaningfully expand, or it has consequences for a public option, because then it puts a lot more of a disproportionate burden on a public option to solve all those gaps.

 Right. We see it already happening where the nation kind of splits population health wise—if you live in, say, the Northeast compared with the South. We’re seeing that now with COVID-19, regarding the states that adopted the mitigation measures early on compared with the states in the South that did not. What do you think the administration could do about that sort of issue? Is it just political?

 I like to think it’s beyond political. The very surface answer is yes, it’s political, red and blue. That’s extremely clear. However, just look at the states: As of the time of this taping, we’ve got about 15 states that are seeing record growth in COVID-19, and they’re mostly the red states—you’ve got the Dakotas, Iowa, Alabama, Tennessee, Georgia, South Carolina. So I’d like to actually say that COVID-19 offers a glimpse into what happens when you don’t have that aspect of the health care system. And my friends who, myself included, come from one of those states, Texas, and my friends in those other states I mentioned, they’ve said that now there’s increasing awareness that a lack of a safety net or lack of an insurance mechanism for people is really costing the state overall. So, in a Biden administration, there’s going to be the part that’s political, but if you get past the politics to some degree, and you can illustrate that the states’ budgets will probably actually in the long run save money, or you’d have to see if there’s legislation to [back up] giving those states extra support. And because if the ACA stuff is gone now, what could we do? You know, what could Democrats do? What could a Republican Senate be willing to do? I’d like to say that it could go beyond the blue and red, but it would have to come with dollars. That costs money. We don’t have any money, but that costs money. And then it would also have to come with probably a lot more technical support. One of the things I have learned is that expansion is not easy, and states that don’t have the infrastructure to support some of this are particularly more likely to say “No, thank you.” States will leave money on the table if it feels like it’s going to be overly burdensome. So, I think that a Biden administration would have to offer both money, through Congress, but also support, which would be through CMS.

 When you reference yourself and your friends in health care, how have the conversations with your patients changed from when you started practicing medicine to the past couple of years to now? What are they worried about?

 It’s funny because I’ve kind of done a full circle in my life. When I started, my first clinic out of residency was in a safety net setting, with people who were uninsured or underinsured, and then I spent years going through a more traditional primary care setting, with people with commercial insurance, Medicare, some Medicaid but not as much, and certainly not as many uninsured. Now I am primarily practicing in a federally qualified health center in DC and Maryland, and of the people I see, two-thirds have no insurance and one-third have Medicaid. Here are the things that have not changed: If you are in a lower socioeconomic status, you are always cost sensitive, always. Like the patients 20 years ago who were in Los Angeles, California, in a federally qualified health center—for my patients, cost was everything. [Even to pay] $4 for ibuprofen, it was very common for me to get approached to see if I could get [it for them]—all federally qualified health centers use the 340B pricing program to get drugs and this is where it works, because they would ask me, “Is this something that I can get through the clinic?” and that is exactly the same. Now, I would say, though, that what’s changed especially—put aside that people have lower insurance access, lower socioeconomic status, lower health literacy—I would say even my most high-insured, wealthy, higher-education patients are becoming more aware of cost issues. They may be able to afford everything even out of pocket, but they’re asking up front about costs more.

And then the other conversation that’s interesting to me is that there’s much more awareness, particularly with procedures, about the relationship between quality and volume. It’s interesting to me when I have patients who would say, “Has that doctor done a lot of hip surgeries?” and so you realize that what they’re doing is they’re connecting the dots that volume actually matters in certain things. And then, the other thing I would say is the [evolution of the] internet era. The internet certainly existed when I started out in practice, but with the ability to get information at your fingertips and the frustrations that patients have with how clunky the health care system is, it’s just much more dramatic. Ten years ago, my patients wouldn’t mind being on hold for 20 minutes to figure out how to schedule something, and today there’s just very little tolerance for that. And that actually is across the board—across race, ethnicity, gender; it breaks all barriers. Everybody is squeezed for time; time is the most precious commodity. Health care does a really crappy job of helping people make time efficient: time lost in waiting rooms, time lost on phones, time lost on fax machines. I can’t believe, in clinic the other day, we heard that one of the local CVS locations that we commonly use was not getting our e-prescriptions efficiently, so we now needed to switch to fax. Only in health care would someone say, “The computer’s not working. We need to go back to fax.” So, that’s a long answer to your question, but I think it’s been an interesting journey to see how health care in some ways has changed dramatically and in some ways, it’s still the same.

 From the patient side, especially for the population that you serve in the health centers, is there a downside to telehealth when you think about access and health literacy?

: We have had predominantly all our patients doing telephone, not video, and it’s for the reasons I think you’re referencing. These are people who don’t have high-speed Wi-Fi at home. Most of the time, they’re taking the calls from someplace at work, or the number of screaming children in the background I’ve had on my patients’ calls with me is almost 100%. These are the moms, especially—since I see a lot more women than men—taking care of lots of kids and managing jobs. So, the good news is that telephone access for them to be able to do their care has actually helped us with compliance and adherence. I’ve been able to actually have follow-up visits, where I [otherwise] would not have been able to because people couldn’t get time off, and waiting for half a day in our clinic’s waiting rooms just was not feasible pre–COVID-19. Telephones made things a lot easier, but that will only last as long as we have this parity between phone and video and some in-person visits. Once that shuts off after the public health emergency, we’re going to return to business as usual. And [because] we had so much success with telehealth—we’re talking about this right now internally—even if they don’t pay as much for it, we want to keep some elements of telephone care. The trillion-dollar question is, what percentage of visits really should be virtual? And nobody really knows the answer to that.

 In terms of what you’ve seen over the past decade or two, what have you seen change in terms of chronic disease and also substance use disorder (SUD), which we know is being affected by the pandemic?

 It’s a great question. More than 10 years ago, I worked in the Senate when one of the first—we called them demos back then—chronic disease demos in diabetes was launched, and it was a big failure. It was a failure because it just wasn’t designed very well. So, I think that we’ve done an amazing job in the past decade of understanding how to design effective chronic disease management programs; we learned a lot over the years. We also realized that there is no such thing as a single chronic disease: You don’t take care of the diabetes without also dealing with their mental health and with their new cancer diagnosis, etc. We’ve done a much better job of [understanding] poly chronic disease, treatment interventions, heart failure, readmissions, helping people figure out their medications after discharge—that’s a lot of work that we’ve done. But then to your second point, we have never done with substance use and behavioral health disorders what I mentioned doing in diabetes and heart failure and asthma, for a lot of bizarre reasons, including the financing of behavioral health being carved out, as well as the fact that we never really integrated people like me, primary care doctors, with mental health providers previously. I think that in a telehealth era, there’s a lot of promise, but we have to overcome some of those structural barriers. For example, you cannot [prescribe] buprenorphine, you cannot get someone off opioids, without face-to-face or video visits. I can’t do telephone medication assistance therapy counseling to get someone off opioids. That’s crazy. Why is that? Because of the structural barriers. So, I think we still have a lot of work to do on the SUD/mental health side. I think it’s possible. I am hopeful that this is, again to your first question, coming out of COVID-19. I think with behavioral health and substance use disorders, it is a tipping point, and I hope that can change.

INTERVIEW

To mark the 25th anniversary of the journal, each issue in 2020 will include an interview with a health care thought leader. The October issue features a conversation with Kavita Patel, MD, MS, nonresident fellow at The Brookings Institution and editorial board member of AJMC®.

Could COVID-19 Become a Policy Tipping Point? A Q&A With Kavita Patel, MD, MS

 To develop a strategy to promote life satisfaction with equity for a diverse insured population.

 Cross-sectional survey and claims analysis.

 We conduct an ongoing survey of a stratified random sample of adult plan members. Among other questions, the survey asks about adequacy of physical activity, healthy eating, abstinence from tobacco, limited alcohol consumption, adequate sleep, and whether the respondent takes time to think about the good things that happen to them (hereafter referred to as “healthy thinking”). We assessed the association of demographic characteristics and the 6 behaviors with life satisfaction.

 We found that although all 6 behaviors were positively associated with life satisfaction, healthy thinking was the behavior associated with the greatest difference in life satisfaction between individuals who did and those who did not practice the behavior. We also found that although members insured through Medicaid or who had a psychosocial diagnosis tended to report significantly lower levels of life satisfaction, two-thirds of the opportunity to improve life satisfaction across the member population was among individuals with neither of these attributes.

 We conclude that the most effective strategy to promote both overall life satisfaction and equity will address social determinants for members with unmet social needs, provide the behavioral and mental health services that benefit members with these needs, and promote healthy lifestyles with an emphasis on healthy thinking for the entire population.

Examining our claims and member survey data, we found that taking time to think about the good things that happen is the behavior most strongly associated with life satisfaction. However, adults living in poverty, as indicated by Medicaid enrollment, and members with claims for psychosocial conditions have particularly low levels of life satisfaction. We conclude that the strategy with the greatest apparent potential to promote overall life satisfaction and equity in health plans similar to ours will:

address social determinants and provide services for members with social needs;

provide behavioral and mental health services to members with these needs; and

promote healthy lifestyles with an emphasis on healthy thinking for all members.

 the majority of the individuals insured by HealthPartners are experiencing less-than-ideal levels of life satisfaction.

 Although individuals who are experiencing social conditions that burden health—like poverty, housing instability, and food insecurity—and psychosocial conditions—such as anxiety and depression—are particularly likely to report low levels of life satisfaction,

 we have found that individuals who are not experiencing these conditions can also suffer from low levels of life satisfaction.

 Promoting life satisfaction as a component of well-being is important to us because we share Atul Gawande’s perspective that “well-being is about the reasons one wishes to be alive.”

To identify the optimal strategy to increase life satisfaction for the adults covered by our health insurance plan, we used HealthPartners summary measures of health and well-being data

 to address 3 questions: (1) Which health-promoting behaviors are associated with the largest differences in life satisfaction? (2) In addition to members experiencing poverty and psychosocial conditions, are there groups of members with particular medical conditions that would benefit from programs to improve their life satisfaction? (3) What strategy appears most likely to promote both health equity and overall life satisfaction for our membership? The answer to the first question is important because it guides the focus of our health promotion programs. The answer to the second question is important because it identifies members whose needs might not immediately come to mind. The answer to the third question will guide our overall strategy to increase life satisfaction for each individual member and across the entire member population.

We conducted the analyses as part of our accountable care organization program to promote our mission “to improve health and well-being in partnership with our members, patients, and community.” The HealthPartners Institutional Review Board (IRB) reviewed the project and agreed that the intent is quality improvement, not human subjects research, and the project is thus not subject to further IRB review.

Founded in 1957, HealthPartners is the largest consumer-governed nonprofit health care organization in the nation. HealthPartners provides a full range of health services including insurance, care delivery for medical and dental conditions, and health and well-being programs for 1.8 million members nationwide. The current analysis applies to health plan members who meet the following criteria: (1) continuously enrolled for a full year, with allowance for a 30-day gap in insurance coverage or discontinuance of enrollment due to death; (2) enrolled during 2015, 2016, and/or 2017; and (3) aged 18 years or older without upper limit. Among the members meeting these criteria, we excluded the following: (1) members known to prefer a language that is not English (less than 2% of our adult member population); (2) members having an address of a correctional facility, nursing home, or hospice (less than 0.002% of our members); and (3) members having their name on our health plan, research, or related do-not-contact lists (less than 1% of our members).

In this analysis, we define psychosocial conditions according to the Johns Hopkins ACG System.

 The 3 most frequent psychosocial diagnoses among our members are major depression, anxiety neurosis, and tobacco use. A complete list of the ACG System psychosocial conditions is available in 

). We used coverage by Medicaid as a surrogate indicator of poverty.

Assessing Member Health-Promoting Behaviors With Surveys

As part of our project to track summary measures of health and well-being, each month our Center for Evaluation and Survey Research emails or mails a survey to a stratified random sample of members who meet the inclusion and exclusion criteria described previously. 

 presents the survey questions that are relevant to the current analysis and the definitions of meeting the behavior goals.

The Center contacts nonresponders sequentially by mail for email nonresponders and by telephone for mail nonresponders. We designed the survey sample to account for the differential non­response (ie, by gender, age, and type of insurance product) observed in prior iterations of the survey. We asked survey respondents to self-identify their ethnicity, race, and education. We collected the survey responses between July 1, 2015, and February 28, 2018, from the population defined previously. We weighted the annual samples to reflect the demographic and disease burden profile of the underlying member population in the year collected and then multiplied all survey weights by a constant in order to reconstruct the 2017 member population that met the inclusion criteria.

We assessed life satisfaction with 1 question: “How satisfied are you with your life?” As in Table 1, the possible responses were anchored at 0, “not at all satisfied,” and 10, “extremely satisfied.” In order to prioritize organizational initiatives to maximize life satisfaction across the entire population, it is necessary to identify subpopulations with the lowest mean levels of life satisfaction, the size of those subpopulations, and a way to express the total burden of low life satisfaction for the subpopulations. The first 2 tasks are straightforward, but to estimate the benefit that might accrue from improving life satisfaction at both the individual and population levels, we needed to create the novel concepts of a “life satisfaction gap” and “life satisfaction points.” In this context, the life satisfaction gap is simply 10 minus the score the survey respondent has selected. For example, because the mean life satisfaction score for our reconstructed 2017 member population is 8.06 (

), the mean life satisfaction gap would be 1.94. The potential life satisfaction points to be gained for a group of individuals are simply the sum, for all members in a group, of the individual life satisfaction gaps. For example, if we were interested in comparing the burden of low life satisfaction in 2 populations, one composed of 1000 individuals with an average life satisfaction score of 7.5 and a second composed of 2000 individuals with an average life satisfaction score of 8.0, the life satisfaction points lost for the first population would be 2500, which is 1000 × (10 – 7.5), and the life satisfaction points lost for the second population would be 4000, which is 2000 × (10 – 8.0).

We present descriptive statistics as means, 95% CIs, and frequencies. We limited the analysis to the 15,692 survey respondents who answered the life satisfaction question and all of the health-related behavior questions. Then, we used weighted general linear regression to model the relationship between (1) practicing or not practicing the 6 behaviors and (2) life satisfaction as a continuous variable, adjusted for the covariates that were available to us: time period, disease burden, age, gender, race and ethnicity, education, and insurance product. We subsequently dropped race and ethnicity and education from the equation because their β weights were not statistically significant (

To test whether our β estimates were sensitive to the general linear regression model, we also used a tobit model to generate β estimates. Because the results were not substantively different, we do not report them. In the remaining analysis, we considered a score of 10 as “satisfied” and applied the results to the entire 2017 reconstructed member population. To estimate the potential reduction in the life satisfaction gap associated with each health-related behavior for our member population, we multiplied the parameter estimates from the general regression model described previously by the estimated number of members not meeting the particular health-related behavior in 2017. We used SAS version 9.4 (SAS Institute) to conduct the analyses.

To answer the final 2 questions, identifying medical conditions that are associated with a particularly large life satisfaction gap and the strategy or strategies most likely to promote both health equity and overall life satisfaction, we identified and visually compared the life satisfaction gaps associated with particular demographic groups and medical diagnoses.

The overall response rate for our survey in 2015, 2016, and 2017 was 42%. With the exception of responses to the alcohol question (9.2% missing) and the physical activity questions (8.9% missing), rates of missing responses to the health behavior questions were 1.5% or less. Only 193 respondents did not respond to the life satisfaction question.

A total of 83% of the survey respondents answered the life satisfaction question and all 6 health-related behavior questions. In addition to listing the number of survey respondents by demographic category, Table 2 displays the attributes of the 2017 member population reconstructed to represent these survey respondents. It also lists the mean life satisfaction score and 95% CI for each of the demographic categories. The 50-to-64-year age group is the largest of the 5, and only 9.2% of the members are 65 years and older. The remaining members are distributed relatively evenly over the 18-to-49-year age range. Although there are only slightly more women than men, the members are predominantly non-Hispanic White. Nearly half have a college degree. Slightly more than 5% are insured through Medicaid.

The overall average life satisfaction score is 8.06. With the exception of a significantly higher life satisfaction score for members 65 years and older, there is little variation in life satisfaction with age, and there is no significant variation by gender or race and ethnicity. Individuals who report having a college degree also tend to report a higher level of life satisfaction. Individuals insured through Medicaid report a significantly lower level of life satisfaction than those insured through another product.

Prevalence of Practicing the Health-Related Behaviors

In the reconstructed 2017 member population, just over 89% of the members reported meeting the tobacco goal and 88% reported meeting the alcohol goal (

). Nearly 70% of the members reported that they meet the physical activity goal, and two-thirds reported meeting the sleep goal. Somewhat more than half of the members reported meeting the healthy thinking goal, but only about a quarter reported meeting the healthy eating goal.

Association of health-related behaviors with life satisfaction at the individual level.

 At the individual level and with adjustment for time period, age, gender, disease burden, insurance product, and the other 5 behaviors, practicing each health-promoting behavior was associated with a statistically significant higher level of life satisfaction. Healthy thinking was associated with nearly a full point of higher life satisfaction. Meeting the physical activity goal, the sleep goal, and the tobacco goal were each associated with a higher level of life satisfaction of about one-third of a point. The associations of meeting the fruit and vegetable goal and the alcohol goal with life satisfaction were about half as large, at 0.17 point and 0.14 point, respectively. The impact of these differences between the current average level and the highest possible level of life satisfaction translate into a narrowing of the gap by nearly 50% for the healthy thinking goal to less than 10% for the alcohol goal.

Association of health-related behaviors with life satisfaction at the population level.

 The life satisfaction point gap for our member population is 1,435,000 points. If our entire member population were to meet the healthy thinking goal, the population might experience a decrease in the life satisfaction point gap of more than 300,000 points (Table 3). Meeting the fruit and vegetable goal might reduce the point gap by almost 100,000 points. If everyone met the physical activity and sleep goals, the point gap might be reduced by around 75,000 each, and if all smokers quit, the point gap might be reduced by 24,000. If everyone met the alcohol goal, the point gap might be reduced by about half this amount. If all members adopted all the behaviors that they did not currently practice, the life satisfaction point gap might be reduced by as much as 600,000 points.

Association of Life Satisfaction and the Life Satisfaction Gap With Medical Diagnoses

Members with a dental claim have, on average, the lowest level of life satisfaction (7.48), but because the group is small (11,747), the life satisfaction gap is also small (30,000 points) (

). The combination of a fairly large number (168,079) of members having a claim for a psychosocial problem and a low mean level of life satisfaction (7.49) gives this group of members the second largest life satisfaction gap (422,000 points). Although the life satisfaction gap attributable to musculoskeletal conditions is larger (523,000 points), it is due to the large number of members with conditions in the diagnostic category; unlike members with psychosocial conditions, they do not have particularly low average levels of life satisfaction (7.95).

The Potential Impact of Behavior Adoption on the Life Satisfaction of Selected Subpopulations

 we present, for selected member categories, the number of members, their mean life satisfaction scores, and the life satisfaction gap associated with the member category. With a grand mean life satisfaction score of 8.06, the life satisfaction gap for our member population is nearly 1.5 million points. The subpopulation with the largest life satisfaction gap is members who neither are insured through Medicaid nor have a psychosocial condition. Their gap is nearly 1 million points. By contrast, although the mean life satisfaction score for our members covered by Medicaid or suffering from a psychosocial condition is significantly lower than average, the gap attributable to these members is about half the size of that attributable to members with neither of these attributes. Whereas the gap for members with a psychosocial condition constitutes about 30% of the total, the gap attributable to members covered by Medicaid is less than 10% of the total.

In this analysis we addressed 3 questions that are important for organizations like ours that are trying to increase the life satisfaction of their constituents while promoting equity: (1) Which health-promoting behaviors are associated with the largest differences in life satisfaction? (2) In addition to members experiencing poverty and psychosocial conditions, are there groups of members with particular medical conditions who would benefit from programs to improve their life satisfaction? (3) What strategy appears most likely to promote both health equity and overall life satisfaction for our membership?

Addressing the first question, we found that all 6 behaviors were associated with higher levels of life satisfaction when the target goals were met. However, practicing healthy thinking was far and away associated with the largest difference. Addressing the second question, we found that, aside from a small group of members with dental claims, members covered by Medicaid or having a psychosocial condition were the only ones with remarkably low average levels of life satisfaction. By contrast, we found that promoting healthy thinking skills for our members who were neither poor nor experiencing psychosocial conditions could significantly raise the level of life satisfaction across the entire membership. Based on these data, we conclude that the answer to the third question is multifactorial: We must address social determinants at a community level and social needs for members insured through Medicaid, provide special services for members with behavioral and mental health needs, and promote healthy lifestyles with an emphasis on healthy thinking for the entire population. Together, these will promote equity and an overall improvement in life satisfaction for our member population, in addition to meeting routine medical care needs.

It might not be surprising that healthy thinking—that is, often taking time to think about the good things that happen in one’s life—was the behavior with the biggest difference in life satisfaction between those who did and did not practice a behavior. However, it might be surprising to some readers that the largest opportunity to improve overall life satisfaction lies with the members who neither suffer from poverty nor from mental or behavioral health problems. This is simply due to the size of the group, not a particularly large gap on average.

Although the benefits of taking action based on this analysis are speculative, it has been clearly documented that individuals who practice health-promoting behaviors have high levels of emotional health and well-being.

 Randomized trials of interventions to increase appreciation, thanking people for even the smallest help and favors, and being kind to others have resulted in decreased symptoms of depression and increased expressions of happiness.

 Because healthy thinking increases life satisfaction, and life satisfaction has been found to be negatively associated with subsequent utilization of health services,

 we might expect that a program that promotes healthy thinking also reduces the seeking of health care. Conversely, individuals who have experienced, for example, adverse childhood experiences and historical trauma would be expected to have lower levels of life satisfaction and optimism.

Any intervention directed at a broad population can, paradoxically, increase disparities. Our data demonstrate that the risk for this to occur is present for our own members. Coaching members who are suffering from the social burdens of poverty—for example, the inability to access a healthy diet or experiencing housing insecurity—to adopt healthy thinking strategies is probably unlikely to be effective and could be considered callous and insensitive. The same is true for members who are suffering from more than modest levels of anxiety and depression; they need help at a level commensurate with their burden. On the other hand, focusing on these 2 groups while ignoring the opportunity to improve life satisfaction for individuals who are not particularly suffering, but are also not thriving to the extent they might, is also not the optimal strategy because it would miss fully two-thirds of the opportunity to improve overall life satisfaction for the population.

These findings are reassuring for our organization. Although not driven by data as explicit as those presented in this analysis, HealthPartners has already started to address the social determinants of health as a community business model.

 Experience gained with this strategy underscores the necessity of developing partnerships with community organizations—both governmental and nongovernmental—because the sectors that influence health and well-being are so diverse and the resource needs are so great. These partnerships are much easier to develop when brand promotion is set aside. Although HealthPartners already provides intensive, high-quality case and care management services to patients with mental and behavioral health challenges and promotes higher levels of well-being and life satisfaction through its health promotion programs,

 it continues to seek and develop more effective solutions by addressing both social needs and social determinants. Among these programs are a partnership with Catholic Charities to provide housing and supportive services for the chronically homeless,

 partnerships with a broad array of community stakeholders to promote optimal brain development in all infants,

 and a partnership with a broad array of stakeholders to create physical and social environments that permit and promote physical activity and healthy nutrition for all children and their families.

There are notable limitations to the inferences that can be made from the analysis we have presented. First, caution is always advised when attributing causality to cross-sectional associations; the extent to which reverse causality is present in our data cannot be determined but could be large. The fact that individuals with high levels of life satisfaction are more likely to seek preventive services

 suggests that they may also be more likely to practice health-promoting behaviors. Basing our analysis on a single question creates obvious limitations. We would have liked to have a far richer database, but we have found that the longer survey that would be required is less acceptable to our members. Without established benchmarks to indicate how much of a life satisfaction gap could be plausibly reduced, we explored extreme scenarios in which 100% of each specified gap is closed. Partial improvements would likely yield proportionally smaller benefits. Finally, our analysis is based on the experience of members who are insured through a commercial or Medicaid product offered by a single health care organization in the Midwest United States, are predominantly non-Hispanic White, are highly educated, and prefer English as their language of choice; thus, the findings must be generalized with caution.

We found that, although practicing each of 6 health-promoting behaviors was significantly associated with a higher level of life satisfaction, increasing the number of members who practice healthy thinking appears to have far and away the greatest potential to increase life satisfaction at both the individual and the population levels. With the exception of a small group of members with dental claims, the members most likely to report low levels of life satisfaction were those living in poverty, as indicated by enrollment in Medicaid, or who have a psychosocial condition. These data suggest to us that the most effective strategy to promote both overall life satisfaction and equity for our member population will address our members’ social determinants of health and social needs: It involves striving to reduce poverty and its impacts, provide behavioral and mental health services to benefit the members who need them, and promote lifestyle interventions with an emphasis on healthy thinking for all members.

 HealthPartners (TEK, ML, JOT, KDC, JMG, NPP, SMK), Minneapolis, MN; HealthPartners Institute (TEK, JYZ, MMJ, NPP), Minneapolis, MN.

 The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (TEK, MMJ, JMG, NPP, SMK); acquisition of data (TEK, JYZ); analysis and interpretation of data (TEK, ML, JYZ, MMJ, KDC, NPP); drafting of the manuscript (TEK, MMJ, KDC); critical revision of the manuscript for important intellectual content (TEK, JOT, JYZ, NPP, SMK); statistical analysis (ML, MMJ, KDC); provision of patients or study materials (JYZ); obtaining funding (JOT, NPP); administrative, technical, or logistic support (ML, JOT, JMG, NPP, SMK); and supervision (TEK, JMG, NPP, SMK).

 Thomas E. Kottke, MD, MSPH, HealthPartners, 8170 33rd Ave S, MS 21110X, Minneapolis, MN 55425. Email: Thomas.e.kottke@HealthPartners.com.

1. Van Kessel P, Hughes A. Americans who find meaning in these four areas have higher life satisfaction. Pew Research Center. November 20, 2018. Accessed October 6, 2019. https://www.pewresearch.org/fact-tank/2018/11/20/americans-who-find-meaning-in-these-four-areas-have-higher-life-satisfaction/

2. Kottke TE, Gallagher JM, Lowry M, et al. The health and well-being of an ACO population. 

3. Clark AE, D’Ambrosio C, Ghislandi S. Adaptation to poverty in long-run panel data. 

. 2016;98(3):591-600. doi:10.1162/REST_a_00544

4. Lucas RE, Clark AE, Georgellis Y, Diener E. Unemployment alters the set point for life satisfaction. 

. 2004;15(1):8-13. doi:10.1111/j.0963-7214.2004.01501002.x

5. Headey B, Kelley J, Wearing A. Dimensions of mental health: life satisfaction, positive affect, anxiety and depression. 

Being Mortal: Medicine and What Matters in the End.

7. Kottke TE, Gallagher JM, Rauri S, Tillema JO, Pronk NP, Knudson SM. New summary measures of population health and well-being for implementation by health plans and accountable care organizations. 

8. Kottke TE, Gallagher JM, Lowry M, et al. Validating a method to assess disease burden from insurance claims. 

9. The Johns Hopkins ACG system: state of the art technology and a tradition of excellence in one integrated solution. Johns Hopkins Healthcare Solutions. 2012. Accessed August 28, 2020. https://www.johnshopkinssolutions.com/wp-content/uploads/2014/04/ACG-White-Paper-Applications-Dec-2012.pdf

10. Sapranaviciute-Zabazlajeva L, Luksiene D, Virviciute D, Bobak M, Tamosiunas A. Link between healthy lifestyle and psychological well-being in Lithuanian adults aged 45-72: a cross-sectional study. 

. 2017;7(4):e014240. doi:10.1136/bmjopen-2016-014240

11. Wiese CW, Kuykendall L, Tay L. Get active? a meta-analysis of leisure-time physical activity and subjective well-being. 

. 2018;13(1):57-66. doi:10.1080/17439760.2017.1374436

12. Dale H, Brassington L, King K. The impact of healthy lifestyle interventions on mental health and wellbeing: a systematic review. 

. 2014;19(1):1-26. doi:10.1108/MHRJ-05-2013-0016

13. Seligman MEP, Steen TA, Park N, Peterson C. Positive psychology progress: empirical validation of interventions. 

. 2005;60(5):410-421. doi:10.1037/0003-066X.60.5.410

14. Southwell S, Gould E. A randomised wait list–controlled pre–post–follow-up trial of a gratitude diary with a distressed sample. 

. 2017;12(6):579-593. doi:10.1080/17439760.2016.1221127

15. Hendriks T, Schotanus-Dijkstra M, Hassankhan A, et al. Resilience and well-being in the Caribbean: findings from a randomized controlled trial of a culturally adapted multi-component positive psychology intervention. 

. 2019;15(2):238-253. doi:10.1080/17439760.2019.1590624

16. Goel V, Rosella LC, Fu L, Alberga A. The relationship between life satisfaction and healthcare utilization: a longitudinal study. 

. 2018;55(2):142-150. doi:10.1016/j.amepre.2018.04.004

17. Kööts-Ausmees L, Realo A. The association between life satisfaction and self-reported health status in Europe. 

. 2015;29(6):647-657. doi:10.1002/per.2037

18. Kim ES, Park N, Sun JK, Smith J, Peterson C. Life satisfaction and frequency of doctor visits. 

. 2014;76(1):86-93. doi:10.1097/PSY.0000000000000024

19. Korkeila K, Kivelä SL, Suominen S, et al. Childhood adversities, parent-child relationships and dispositional optimism in adulthood. 

. 2004;39(4):286-292. doi:10.1007/s00127-004-0740-x

20. Caprara GV, Fagnani C, Alessandri G, et al. Human optimal functioning: the genetics of positive orientation towards self, life, and the future. 

. 2009;39(3):277-284. doi:10.1007/s10519-009-9267-y

21. Isham GJ, Zimmerman DJ, Kindig DA, Hornseth GW. HealthPartners adopts community business model to deepen focus on nonclinical factors of health outcomes. 

. 2013;32(8):1446-1452. doi:10.1377/hlthaff.2011.0567

22. Through its “BeWell Program,” HealthPartners advances to Gold Level HealthLead Workplace Health Accreditation. HealthPartners. September 14, 2015. Accessed October 26, 2019.
https://www.healthpartners.com/hp/about/press-releases/09-14-15a.html

23. Collins J. Higher Ground, new homeless shelter, open for first night in St. Paul. Minnesota Public Radio. January 13, 2017. Accessed March 13, 2020. https://www.mprnews.org/story/2017/01/12/higher-ground-shelter-opens

24. Build a baby’s brain. Little Moments Count. Accessed March 13, 2020. https://www.healthpartners.com/littlemomentscount/index.html

25. PowerUp. Lakeview Health. Accessed March 13, 2020. http://www.powerup4kids.org/Home

26. Kim ES, Kubzansky LD, Smith J. Life satisfaction and use of preventive health care services. 

. 2015;34(7):779-782. doi:10.1037/hea0000174

WEB EXCLUSIVE

The authors used health care claims and survey data to identify a strategy that might promote life satisfaction while advancing equity in an insured population.

A Strategy to Promote Population-Wide Life Satisfaction While Advancing Equity

 To evaluate the utility of machine learning (ML) for the management of Medicare beneficiaries at risk of severe respiratory infections in community and postacute settings by (1) identifying individuals in a community setting at risk of infections resulting in emergent hospitalization and (2) matching individuals in a postacute setting to skilled nursing facilities (SNFs) that are likely to reduce the risk of infections.

 Retrospective analysis of claims from 2 million Medicare beneficiaries for 2017-2019.

 In the first analysis, the rate of emergent hospitalization due to respiratory infections was measured among beneficiaries predicted by ML to be at highest risk and compared with the overall average for the population. In the second analysis, the rate of emergent hospitalization due to respiratory infections was compared between beneficiaries who went to an SNF with lower predicted risk of infections using ML and beneficiaries who did not.

 In the community setting, beneficiaries predicted to be at highest risk had significantly increased rates of emergency department visits (13-fold) and hospitalizations (18-fold) due to respiratory infections. In the postacute setting, beneficiaries who received care at top-recommended SNFs had a relative reduction of 37% for emergent care and 36% for inpatient hospitalization due to respiratory infection.

 Precision management through personalized and predictive ML offers the opportunity to reduce the burden of outbreaks of respiratory infections. In the community setting, ML can identify vulnerable subpopulations at highest risk of severe infections. In postacute settings, ML can inform patient choices by matching beneficiaries to SNFs likely to reduce future risk.

The significant findings of this study include:

A machine learning (ML) algorithm accurately predicted risk of 90-day hospitalizations associated with respiratory infections in a large Medicare beneficiary population using only administrative claims data.

An ML algorithm recommended skilled nursing facilities (SNFs) associated with lower rates of hospital readmissions or emergency department visits due to respiratory infections during an SNF stay based on each patient’s medical characteristics and historical performance of SNFs in providing care for similar patients.

During the current coronavirus disease 2019 (COVID-19) pandemic, ML algorithms have the potential to identify high-risk patients in community and postacute settings who warrant preemptive management to reduce the risk of severe respiratory illness associated with COVID-19.

On March 11, 2020, the World Health Organization declared coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), to be a pandemic (ie, a global outbreak occurring on a scale that crosses international boundaries and affects a large number of people). There is much that we do not know about COVID-19. Estimates for lethality range from less than 0.1%

 What we do know is that it is remarkably contagious, with a reproductive number between 2 and 2.5 compared with a little over 1 for influenza.

While scientists work rapidly to develop therapeutics and vaccines, public health officials have resorted to measures designed to slow the spread of the disease. These efforts to “flatten the curve” aim to preserve capacity and prevent the health system from being overwhelmed by demands for limited resources such as intensive care unit beds, ventilators, and trained care providers. Flattening the curve is targeted toward the population at large. Augmenting these efforts with precise and proactive outreach to higher-risk subpopulations would provide significant value. This subgroup, comprising individuals most vulnerable to severe infections, requires careful and targeted action that goes beyond social distancing and other conventional approaches used for the population at large.

Unfortunately, there are not yet enough data about COVID-19 to draw strong conclusions about who is likely to experience severe sequelae from acquiring the disease. Serious complications associated with COVID-19 are not restricted to elderly or frail individuals; nearly 40% of those hospitalized with COVID-19 in the United States are younger than 54 years.

Using the large amount of historical data about other viral and bacterial respiratory infections, however, it is possible to build useful models for predicting vulnerable individuals who might be at highest risk of severe infections of COVID-19 that require emergent hospitalization. This study explores the precision management of patients at risk of severe respiratory infections—more specifically, personalized and predictive management using machine learning (ML) to provide early targeted management to reduce risk. In particular, this study explores 2 hypotheses centered on the potential for precision management to reduce the risk of infections for vulnerable individuals: First, precision medicine can be used to predict patients’ personalized risk of adverse outcomes due to respiratory infections in the community setting, and it offers an opportunity to reduce their burden on the health care system through targeted interventions. Second, patient risk of contracting respiratory infections at skilled nursing facilities (SNFs) for postacute care varies based on personalized characteristics, and this risk can be predicted using precision medicine to match patients to facilities that minimize this risk.

Administrative claims data from roughly 2 million Medicare beneficiaries in community settings between 2017 and 2019 were utilized in this retrospective study. The primary outcome was urgent or emergent inpatient hospital admission with a primary 

International Classification of Diseases, Tenth Revision

 code related to respiratory infections. A commercially available ML algorithm (HEALTH[at]SCALE Corporation; San Jose, California) was trained on administrative claims data from 2017 and 2018 to predict the risk of respiratory infection–related hospital admission within 90 days. For training purposes, feature sets and outcomes were constructed from 2017 and 2018 data such that each 90-day outcome window ended prior to December 31, 2018. This algorithm was then applied to 2018 administrative claims data to predict the risk of respiratory infection–related hospital admission in the first quarter of 2019. Data from 2019 were used only for evaluation (ie, used as a clean held-out test set that had no overlap with the training data). The incidence of respiratory infection–related admissions was compared between the top 0.1% and the top 1% of the cohort with the highest predicted risk and the overall population during a 90-day evaluation period.

In the second analysis, administrative claims data from Medicare beneficiaries receiving postacute care at SNFs were utilized. The primary outcome that was measured was the incidence of inpatient hospital readmission or emergency department (ED) visits associated with respiratory infections during the course of SNF stays or within a week of SNF discharge. A commercially available ML algorithm (HEALTH[at]SCALE; San Jose, California) was trained on data from 2017 and 2018 to predict the risk of outcomes for each SNF based on patient medical characteristics at the time of discharge. For training purposes, feature sets and outcomes were constructed from 2017 and 2018 data such that each outcome window ended prior to December 31, 2018. The algorithm was then applied to postacute SNF stays in 2019 (again used as a clean held-out test set that had no overlap with the training data). The rate of respiratory infections that required ED visits or inpatient hospital care was compared between patients who received care at 1 of 3 top-predicted SNFs (as generated by the ML algorithm) and patients who received care at facilities that were not among the top 3 predicted SNFs.

 shows the characteristics of the Medicare beneficiary population for this analysis. In the community setting, using trailing 12 months of claims data, ML generated a subcohort comprising the top 1% (and also the top 0.1%) of patients estimated to be at greatest risk of severe respiratory infections over the next 90 days. The predictions were compared with the actual outcomes for these patients over this period. Analysis of the claims data for the 90-day evaluation period showed that patients estimated to be at greatest risk were at a 13-fold increased risk of ED visits (13.3% vs 1.0%), an 18-fold increased risk of hospitalization (14.6% vs 0.8%), and a 15-fold increased risk of either ED visits or hospitalizations (16.5% vs 1.1%) due to respiratory infections relative to the population average (

 [A]). The area under the receiver operating characteristic curve was 0.79 or greater for all outcomes (predicting ED visits, hospitalizations, and the composite of the 2 outcomes).

In the postacute setting, patients receiving care at SNFs recommended by ML achieved substantial improvements in outcomes relative to those who received care at nonrecommended SNFs. Specifically, patients receiving care at 1 of the top 3 machine intelligence–based personalized SNF recommendations within their local geography had relative reductions of 37% in ED visits (5.5% vs 8.7%) and 36% in inpatient hospitalizations (3.9% vs 6.1%) due to respiratory infections during the course of SNF stays or within a week of SNF discharge (Table 2 [B]). There was a corresponding 26% relative reduction in all-cause ED visits for SNFs recommended by ML (26% vs 35%).

In this study, the risk of hospitalizations due to respiratory infections in Medicare beneficiaries in community settings and postacute settings (after hospital discharge) was analyzed. In the community setting, ML accurately classified patients who had the highest risk of hospitalization within 90 days due to respiratory infections. In the postacute setting where Medicare patients were discharged to SNFs, patients who went to SNFs recommended by ML had significantly lower rates of hospital readmissions and ED visits associated with respiratory infections compared with patients who went to SNFs that were not recommended for them.

The current pandemic associated with novel coronavirus–infected pneumonia (NCIP) has swept across continents and afflicted more than 13 million people worldwide and has been attributed to more than 570,000 deaths due to highly efficient human-to-human transmission.

 Morbidity and mortality have been particularly high in elderly patients and those with chronic comorbidities due to severe respiratory complications associated with NCIP.

 It may be possible to improve outcomes and reduce the burden on the health care system by identifying vulnerable patient populations and utilizing preemptive and preventive outreach to reduce their risk.

Given that SARS-CoV-2 is a relatively new pathogen associated with severe respiratory infections, the performance of an ML algorithm in identifying high-risk or “vulnerable” patients in the community and postacute settings was evaluated. Based on the current epidemiologic and outcomes data regarding SARS-CoV-2, patients who have been vulnerable to severe complications from viral influenza, community-acquired pneumonia (CAP), and health care–associated pneumonia would likely be vulnerable to SARS-CoV-2, as well.

 In a study examining risk factors for CAP, it was found that 70% of hospitalized patients were 50 years and older and 78% had 1 or more chronic comorbidities including congestive heart failure, chronic obstructive pulmonary disease, diabetes, or immunosuppression

 In a recently published systematic review of several risk models for SARS-CoV-2, similar risk factors related to advanced age and chronic comorbidities were found to be associated with severe SARS-CoV-2 infections.

 Thus, in the absence of large-scale SARS-CoV-2 patient data for training ML models, use of prior patient data associated with patients hospitalized for respiratory infections may suffice.

Among Medicare beneficiaries in community settings, patients who were classified by ML to be in the top 1% of risk for respiratory infection–related hospitalization within 90 days had an 18-fold higher risk than the baseline Medicare population. In the current COVID-19 pandemic, a similar ML algorithm could be utilized to identify patients who are at high risk of developing severe respiratory illnesses due to COVID-19. Targeted programs aimed at these high-risk patients during the COVID-19 pandemic may include reinforcement of guidelines about social distancing, contact precautions, symptom education, home delivery of medications, house calls/televisits, portal engagement, and other preventive outreach to mitigate risk in this cohort.

Besides community patients, patients being discharged from hospitals for postacute care are at severely elevated risk of experiencing substantial mortality or morbidity as a result of infection. Prior research has shown differential performance of SNFs in reducing hospital readmissions and mortality for Medicare beneficiaries in the postacute setting.

 Thus, the choice of SNF destination may influence outcomes in vulnerable Medicare patients in the postacute setting. An ML algorithm was utilized for selecting SNF destinations, and patients who went to SNFs that were recommended by the ML algorithm were shown to have lower rates of ED visits and hospitalizations due to respiratory infections. Such personalized recommendations for matching patients to facilities are especially pertinent given the requirements currently in place for hospitals to provide patients with information about postacute care provider choices. Personalized recommendations for the choice of SNF driven by historic data are intended to augment the decision-making process for caregivers and patients while preserving choice and focusing on optimizing long-term outcomes. Furthermore, we believe that personalized, ranked recommendations for facilities might be more actionable than statistics about facilities derived from the general population. In the current pandemic, use of ML tools for clinical decision support in the postacute setting may similarly help in preventing SARS-CoV-2–related severe illness.

Future work will explore the use of recent data collected from SARS-CoV-2–affected patients to update our model. Limitations of this study include using only administrative claims data, which may not fully characterize patient frailty and medical problems as well as the clinical data available in electronic health records. Medicare beneficiaries are typically older; thus, the cohort in this study does not reflect predictive performance of the model on younger patients who are also susceptible to respiratory infections. Risk of respiratory infections is also influenced by exposure risk due to disease prevalence in the community

 and other socioeconomic factors such as population density, patient contacts, and air quality index that were not evaluated in this study. The proposed matching of patients to facilities for postacute care was not evaluated in prospective observational causal studies or randomized clinical trials and would benefit from follow-up investigation in studies of this nature.

Tiange Zhan, MS, and Dev Goyal, MS, contributed equally to this work and are listed as co–first authors.

 HEALTH[at]SCALE Corporation (TZ, DG, JG, RM, ZE, ZS, MS), San Jose, CA; Massachusetts Institute of Technology (JG), Cambridge, MA; University of Michigan (MS), Ann Arbor, MI.

 Ms Zhan, Mr Goyal, and Mr Mehta are employees of and stock owners in HEALTH[at]SCALE Corporation, which developed the machine learning algorithms described in the paper. Dr Guttag is a board member and consultant for HEALTH[at]SCALE, owns stock in HEALTH[at]SCALE, has attended the Becker’s Hospital Review conference, and reports patents pending filed by HEALTH[at]SCALE and patents received licensed by HEALTH[at]SCALE. Mr Elahi is an employee of HEALTH[at]SCALE and owns stock in HEALTH[at]SCALE. Drs Syed and Saeed are board members and employees of HEALTH[at]SCALE, report patents pending from HEALTH[at]SCALE, and own stock in HEALTH[at]SCALE.

 Concept and design (JG, ZE, ZS, MS); acquisition of data (TZ, DG, RM); analysis and interpretation of data (TZ, DG, RM, MS); drafting of the manuscript (TZ, DG, JG, ZE, ZS, MS); critical revision of the manuscript for important intellectual content (TZ, DG, JG, ZS, MS); statistical analysis (TZ, DG); administrative, technical, or logistic support (RM); and supervision (ZE, ZS, MS).

 Mohammed Saeed, MD, PhD, University of Michigan, 1500 E Medical Center Dr, Ann Arbor, MI 48109. Email: msaeed@med.umich.edu.

1. Ioannidis JPA. A fiasco in the making? as the coronavirus pandemic takes hold, we are making decisions without reliable data. STAT. March 17, 2020. Accessed March 26, 2020. https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/

2. Baud D, Qi X, Nielsen-Saines K, Musso D, Pomar L, Favre G. Real estimates of mortality following COVID-19 infection. 

. 2020;20(7):773. doi:10.1016/S1473-3099(20)30195-X

3. Coronavirus disease 2019 (COVID-19): situation report, 51. World Health Organization. March 11, 2020. Accessed March 12, 2020. https://apps.who.int/iris/bitstream/handle/10665/331475/nCoVsitrep11Mar2020-eng.pdf

4. Mansell W, Flaherty A. Nearly 40% of those hospitalized with coronavirus are younger than 54: CDC. ABC News. March 19, 2020. Accessed March 26, 2020. https://abcnews.go.com/US/40-hospitalized-coronavirus-younger-54-cdc/story?id=69681304

5. Li Q, Guan X, Peng W, et al. Early transmission dynamics in Wuhan, China, of novel coronavirus–infected pneumonia. 

. 2020;382(13):1199-1207. doi:10.1056/NEJMoa2001316

6. Hak E, Wei F, Nordin J, Mullooly J, Poblete S, Nichol KL. Development and validation of a clinical prediction rule for hospitalization due to pneumonia or influenza or death during influenza epidemics among community-dwelling elderly persons. 

. 2004;189(3):450-458. doi:10.1086/381165

7. Bont J, Hak E, Hoes AW, Schipper M, Schellevis FG, Verheij TJM. A prediction rule for elderly primary-care patients with lower respiratory tract infections. 

. 2007;29(5):969-975. doi:10.1183/09031936.00129706

8. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. 

. 1997;336(4):243-250. doi:10.1056/NEJM199701233360402

9. Chalmers JD, Singanayagam A, Akram AR, et al. Severity assessment tools for predicting mortality in hospitalised patients with community-acquired pneumonia. systematic review and meta-analysis. 

. 2010;65(10):878-883. doi:10.1136/thx.2009.133280

10. Millett ER, De Stavola BL, Quint JK, Smeeth L, Thomas SL. Risk factors for hospital admission in the 28 days following a community-acquired pneumonia diagnosis in older adults, and their contribution to increasing hospitalisation rates over time: a cohort study. 

. 2015;5(12):e008737. doi:10.1136/bmjopen-2015-008737

11. Jain S, Self WH, Wunderink RG, et al; CDC EPIC Study Team. Community-acquired pneumonia requiring hospitalization among US adults. 

. 2015;373(5):415-427. doi:10.1056/NEJMoa1500245

12. Wynants L, Van Calster B, Collins GS, et al. Prediction models for diagnosis and prognosis of covid-19 infection: systematic review and critical appraisal. 

13. Neuman MD, Wirtalla C, Werner RM. Association between skilled nursing facility quality indicators and hospital readmissions. 

. 2014;312(15):1542-1551. doi:10.1001/jama.2014.13513

14. Barnes R, Blyth CC, de Klerk N, et al. Geographical disparities in emergency department presentations for acute respiratory infections and risk factors for presenting: a population-based cohort study of Western Australian children. 

. 2019;9(2):e025360. doi:10.1136/bmjopen-2018-025360

TRENDS FROM THE FIELD

This paper evaluates novel machine intelligence to predict patients at risk of severe respiratory infections and recommend postacute care providers likely to reduce infection risk.

Machine Intelligence for Early Targeted Precision Management and Response to Outbreaks of Respiratory Infections

The coronavirus disease 2019 pandemic has created opportunities for hospitals and emergency departments to emerge leaner and more efficient than before.

Key solutions must be applied to optimize operational efficiency, patient safety, and fiscal responsibility.

Limiting artificial variability and understanding the contrast between it and natural variability are necessary for efficient use of hospital capacity.

The coronavirus disease 2019 (COVID-19) pandemic has forced hospital administrators to implement various operational changes that will redefine health care delivery for the foreseeable future. As operational leaders prepare for an eventual return to prepandemic patient volume, we must balance competing priorities: non—COVID-19 elective patient volume that positively affects the revenue cycle and maintaining optimal surge capacity for future peaks of COVID-19–related illness.

Prior to the current pandemic, emergency department (ED) boarding was an unfortunate reality across the United States that led to well-documented increases in mortality, length of stay, and overall cost for admitted patients.

 Overcrowding in the midst of a respiratory pandemic is a dangerous proposition that affects both patient and provider safety. Lessons learned from the pandemic must empower hospitals to emerge leaner than before.

In the midst of the coronavirus disease 2019 (COVID-19) pandemic, health care leaders must work to optimize emergency department and hospital efficiency while maintaining patient access to care.

October 2020