In a speech at a health conference in Battle Creek, Michigan, Professor A. Arnold Clark of the state board of health condemned our tendency to shortchange investments in prevention: “But let us see how much we really pay in Michigan for the prevention of disease. How much do you pay in Battle Creek? You have in this city about 45 physicians. I suppose they average an annual income of about $200,000 each. That makes about $9 million which you pay each year to cure you after you get sick.
“Now how much do you pay to prevent your getting sick?” Professor Clark continued. “Probably, not more than $50,000. You have a health officer and you probably pay him about $50,000 a year to stamp out communicable diseases in Battle Creek. Now prevention is better than cure.… Some people seem to think that because they always have lived, all money spent to prevent disease and dying, is money wasted. You have heard of the man who dropped his life insurance because he had kept the thing up 20 years, and never derived any benefit from it yet. That seems too often to be the policy of the city, the state, the nation.”
Professor Clark delivered this speech in 1890. (I modernized the dollar figures, but otherwise the quote has been left intact.) It’s telling that public health experts are still forced to make the same argument today: Prevention is better than cure. Actually, it’s enraging. Because in the 130 years since Professor Clark made his speech, we have gathered overwhelming evidence of the efficacy of prevention and public health: Just look at our life expectancy.
In 1900, the average life expectancy at birth for Americans was 47.3 years. By 2000, life expectancy had reached 76.8. Clearly a dramatic improvement, but let’s be clear about what these numbers mean and don’t mean.
“Life expectancy” is an average across a population. In a population of 5, 1 person might live to 75 while the others live to 91, 70, 66, and 82, yielding an average life span of 76.8. The average blurs the variety. (Stick with me here—I know this is obvious so far.)
But sometimes averages don’t just blur an underlying reality, they obliterate it. I’m amazed, for instance, by how many perfectly smart people seem to believe that the 47.3-year life expectancy in 1900 is synonymous with “most people lived significantly shorter lives back then.” I suppose they picture our ancestors, in their mid-forties, tottering around with canes and false teeth and trying frantically to get their affairs in order. From this perspective, the passage of the Social Security Act in 1935 would have been a cruel joke indeed—Yes, you can start collecting retirement income at 65—twenty years after you’re dead!! [maniacal laugh]
A representative sample of life spans in that era absolutely did not look like this: 46, 48, 56, 39, and 48, averaging out to 47.4 years. Rather, it looked more like this: 61, 70, 75, 31, and 0. At the turn of the century, in 1900, almost 1 out of every 5 children was dead before reaching their fifth birthday.
The natural life span of human beings today is not that different than it was a hundred years ago. What’s different is that we’re saving a lot of people—especially babies and children—from dying too early. On this point, you may have noticed that Clark, in his rant above, emphasized “communicable diseases.” That’s because in his era, about a third of all deaths were due to infectious diseases such as pneumonia, influenza, tuberculosis, and diphtheria. These diseases were disproportionate killers of children. Today, the number of people dying of those infectious diseases has decreased from about 33% in 1900 to less than 3% in 2010.
What’s responsible for that decline? Upstream efforts. Better hygiene, cleaner water, pasteurization, and improved living conditions, as well as the advent of sewage systems and the introduction of antibiotics and vaccines. Yet even in the face of this massive success—and by “massive success,” I mean imagine that every fifth child in your family tree would have died were it not for this work—public health must still plead for resources.
“We under-invest in the services and policies that would keep people healthier so that they would not develop those illnesses or have the injuries or suffer from premature deaths that we know could be avoided,” said John Auerbach, the president of the public health policy group Trust for America’s Health. “It’s tragic.” The group pegged the total national spending on public health specifically at $88.9 billion, just 2.5% of total health care spending in the United States in 2017.
Public health efforts suffer from what is effectively a punishment for success. “In public health, if you do your job, they cut your budget, because no one is getting sick,” said Julie Pavlin, a physician with experience running global health programs and combating infectious diseases for the Army. And her comment gets to the heart of the matter: The fee-for-service model in health care favors reaction over prevention.
“We’ll pay $40,000 a year for the price of insulin, but we won’t pay $1,000 to prevent someone from ever getting diabetes,” said Patrick Conway, a former deputy administrator at CMS (the Centers for Medicare & Medicaid Services). “We should pay for value. Imagine if a car that took more hours to build was more expensive. That would make no sense. Cars would not become better and cheaper if you paid for them that way.”
At one point, I came across a commentary that highlighted American consumers’ world-leading access to MRI scans. We get them quicker and more often, apparently, than anyone else in the world. (USA! USA!) To take pride in that is a bit like bragging that Americans lead the world in receiving airport security pat-downs. I mean, if there’s something to find, it’s true that we’d like to find it quickly, but surely, we’d rather be the nation whose people need the least checking-over. (And as the Gil Welch turtles/rabbits logic implies, we might discover things that don’t need discovering.) What the MRI statistic illustrates is a simpler idea about our fee-for-service system: When you get paid for something, you do more of it. (No doubt we also “lead the world” in dental X-rays. And just imagine if TSA agents were paid by the grope.)
So, in short, reactive efforts succeed when problems happen and they’re fixed. Preventive efforts succeed when nothing happens. Who will pay for what does not happen?
That’s not an unanswerable question. There are people who will pay for what does not happen. (Including you! Presumably all those oil changes are preventing something.) But creating payment models to fund upstream efforts can be almost unbelievably complicated, for reasons we will explore in this chapter.
First, though, we should remind ourselves how easy it should be to pay for upstream efforts. Take the case of Poppy + Rose, a comfort food restaurant in downtown Los Angeles. Diana Yin, the co-owner, monitored customer reviews carefully, and she noticed a customer had complained online about receiving cold waffles for brunch. She did some detective work and discovered that the restaurant’s one waffle maker could not keep up with demand at brunch. So cooks had started making waffles before the rush hit to build up a stash in reserve. A clever workaround—but it led to cold waffles. Nobody likes a cold waffle. So Yin sprung for a second waffle maker.
This is the dream scenario from the perspective of paying for prevention. It’s so simple: The person who paid, Diana Yin, is the one who will reap the rewards. Think of it in terms of “pockets”: The money was spent from one pocket, and it will be returned to that same pocket. And Yin will probably recoup her investment quickly. This same one-pocket logic would apply, of course, to an investment you made in yourself: a certification or a graduate degree. You might spend thousands of dollars today in the hopes of earning many more thousands in the future.
Our story gets more complicated quickly, though. Having one pocket does not guarantee that wise upstream investments will be made. Here’s an example: For decades, caregivers in nursing homes have suffered lower-back injuries from lifting and transferring patients. This is awful for the caregivers, of course, and also costly for their employers, who must contend with lost workdays and workers’ compensation claims for the injuries.
Entrepreneurs created mechanical patient-lifting equipment to solve exactly this problem. For the nursing home leader, though, it was not an obvious investment to make. The machines were very expensive, and they required a whole new set of procedures—staffers had to relearn how to transfer patients using the machinery—and they were slower than the old-fashioned, lower-back-fueled technique. So why embrace the nuisance and cost? It’s easier to stay in the tunnel and accept that, every now and then, someone’s going to get hurt.
Then, in the late 1990s, an evaluation found that if caregivers used specific, research-tested techniques for transferring patients—including the use of that equipment—then nursing homes could cut lost workdays and workers’ compensation claims by two-thirds. As a result, the investment in the equipment could be repaid in less than three years. These findings were publicized within the long-term care industry, and nursing homes increasingly adopted the new procedures, leading to a 35% reduction in lower-back injuries between 2003 and 2009, according to a CDC report.
So here’s our first wrinkle: The nursing homes had the luxury of a single pocket, but the choice of whether to buy the patient-moving machines was murkier than with the waffle maker example. It was hard to assess the investment from within a single nursing home. A broader perspective, with evidence drawn from across the nursing home industry, was required: Hey, this equipment is well worth the money. Even in a simple case like this one, then—where a good payoff awaited an investment—the inertia pushed against prevention.
Now let’s swing to the opposite end of the spectrum: the maddening complexity of creating funding models for social services. The Nurse-Family Partnership (NFP) offers a representative case. The program was founded in the 1970s by David Olds, a recent college graduate who had grown disillusioned with the inner-city day care center where he worked. Many of the preschool kids he worked with were suffering because of their parents’ bad decisions. One child had developed few language skills, speaking mostly with grunts. When Olds interviewed the child’s grandmother, she told him that her daughter (the child’s mother) was an addict who had used drugs throughout her pregnancy. Another boy always seemed restless at nap time. Olds later learned that he was beaten by his mother every time he peed in his sleep.
Olds realized that he could have helped these kids much more if he could have intervened earlier in their lives. He could have best served them, he believed, by helping their mothers. The kind of abuse he was witnessing was motivated more by ignorance than by cruelty. These mothers, simply put, didn’t have the knowledge or skills they needed to be effective mothers. They didn’t have support systems or role models, and they didn’t know what to do with the frustration and anger that comes with raising kids.
The program he created, NFP, matches registered nurses with low-income, first-time pregnant women. The same nurse visits the young woman in her home regularly during her pregnancy and throughout the first two years of the child’s life. The nurse acts as a mentor, helping the mother deal with the tensions of parenting: what to do when children cry, what to do when they don’t sleep, how to get children on a schedule. And the nurse explains the basics: how to breast-feed, swaddle a baby, transition babies to whole foods, brush a child’s teeth, and so on. Beyond the parenting instruction, a crucial part of the work is simply to be a caring human being who’s there to support the mother. To show her how to take care of herself so that she can care for her child. To help her navigate the complexities of working while raising a child. To listen when the pressures of life seem overwhelming.
Three major randomized controlled trials of NFP have been conducted in the United States: in Elmira, New York; Memphis, Tennessee; and Denver, Colorado. The studies have shown that the program consistently improves maternal health, child safety, and well-being. Among the specific impacts were reductions in smoking during pregnancy, preterm births, infant mortality, child abuse and maltreatment, and criminal offenses by the mother, Food Stamp payments, and closely spaced pregnancies (second births within 18 months of the first). That’s quite a laundry list of bad things avoided. One study estimated a return of at least $6.50 for every $1 invested in NFP.
Talk about an easy investment decision! Even if it took you 20 years to earn the $6.50 return, that’s still the equivalent of about 10% annual interest. So, we’d expect that, given the results, NFP would be available for every low-income, first-time mother in the country who wants it. No, far from it. Why?
In our simple waffle-iron case, the person who made the investment received the benefits. One pocket. But in this case, notice how splintered the rewards are. The primary beneficiaries, of course, are the child and the mother, in that order. But they can’t pay. Who else benefits? All the other institutions who would have had to pay for bad outcomes, if it weren’t for NFP. Let’s take three examples:
There are plenty more beneficiaries beyond those three, including ripple effects on health, education, income, and more. Everyone wins!
Let’s say a local health system could be persuaded to fund NFP. It’s an expensive program, costing roughly $10,000 per woman served. Unfortunately, the health system would receive only a small benefit from the investment, since the primary value goes to all the other parties described above. That’s an example of what’s called the “wrong pocket problem”: a situation where the entity that bears the cost of the intervention does not receive the primary benefit. One pocket pays, but the returns are scattered across many pockets.
Ideally, you’d fix that by passing the hat around to all the relevant parties who’d benefit—taking up a collection to fund NFP. But here are the objections you’ll encounter: There’s no precedent for that. There’s no line item in my budget for “chipping in for a program that may pay me back eventually.” And, Let’s say you’re wrong and we don’t save money downstream—does that mean you’ll refund my money? Concerns like these explain why programs like NFP, which could create enormous social benefit, simply can’t get funded at the level they deserve.
But there are experiments underway to fix the wrong pocket problem. A group in South Carolina designed a “pay for success” model that could fund a wide expansion of NFP’s work. Here’s how they set it up: In 2016, NFP received a $30 million infusion of cash to expand its work in the state, and the results of its efforts will be assessed over six years via a randomized control trial. If the work is successful, according to several measures agreed upon in advance, then the state government would be positioned to fund the work permanently. The magic of the arrangement is that the state government doesn’t take a big financial risk upfront, because the trial stage was mostly funded from outside. So if NFP proves a valuable investment, South Carolina will reap the rewards; if it doesn’t, the state isn’t out much.
Conceptually, this arrangement is not difficult to understand, but the intricacies of the deal were exhausting. “We spent three years trying to figure out how to make the rules allow us to do something that everybody in the room understood on day one was the most obvious thing in the world,” said Christian Soura, then head of South Carolina’s Department of Health and Human Services (DHHS). To get a feel for the difficulty, just look at the list of players involved: The NFP team in South Carolina. The DHHS. The Abdul Latif Jameel Poverty Action Lab. The Harvard Kennedy School of Government Performance Lab. The consulting firm Social Finance. The Duke Endowment. BlueCross BlueShield of South Carolina Foundation. (And honestly, many more—just accept this list as a sampler.)
Soura said the negotiations involved answering “How do we figure out how to get all these different governmental funding streams to let us pay for a thing that all of us know we need more of? And that winds up being this Kafkaesque nightmare of navigating these different federal and state funding restrictions on all these different sources of funding.”
The deal has great promise. The initial investment will allow NFP to offer services to an additional 3,200 mothers, supporting them from their pregnancies through the first two years of their children’s lives. Those children will be raised in happier, healthier homes as a result of NFP’s support. The payoff for those mothers and children should be profound.
Perhaps more important, over the long term, is that the deal could break the wrong-pocket curse. If NFP delivers on expectations, then the state and federal government would want to fund the work on an ongoing basis, since the payoff on the investment would be clear. And since there are 49 other states with high-risk populations of mothers who need help, too, the possibilities for expansion are almost limitless. In that context, three years of laborious haggling over the core contract doesn’t look like a poor investment.
We can pay to fix problems once they happen, or we can pay in advance to prevent them. What we need are more business and social entrepreneurs who can figure out how to flip payment models to support the preventive approach. Here’s a trivial example of how that can work: A few years ago, my wife and I flipped to “upstream” pest control. We’d had a problem with spiders, so we called an exterminator. When he visited, he offered us a subscription service. The idea was that they’d visit on a regular basis—not requiring an appointment, just spraying outside our home periodically—using the best strategies they’d learned to keep bugs at bay. At first, we were skeptical—“Are we getting ripped off here?”—but ultimately what won us over was the beautiful vision of removing bugs from our life concerns. So we did it, and we removed one small source of drama from our lives. No longer are we cycling from Infestation to Rescue to Inaction (and repeat). Now it’s just a quiet and mostly invisible routine: maintain, maintain, maintain.
And it occurred to me, in a similar vein, how many of the world’s household repairs are caused by a failure of upstream maintenance. The air-conditioning system breaks prematurely because the air filters weren’t changed regularly. The hot-water heater stops working because it was never flushed.I Toilet problems, gutter problems, roof problems: Aren’t many of these faults preventable? Some of us treat our homes like cars that have never had an oil change.
If someone would “own” this work for you—assuming responsibility for the integrity of your major household appliances and systems—would you pay them a regular, monthly fee? Like, forever? It’s a concept that at least one major business is exploring. “The home service industry really hasn’t changed in the modern era,” said Brandon Ridenour, the CEO of ANGI Homeservices, which includes both the websites HomeAdvisor and Angie’s List. “It’s almost identical to the way it worked 50 years ago. An individual need pops up unexpectedly—it comes up out of the blue and people are left to deal with it reactively. ‘I need a plumber, an electrician, a handyman.’ That starts the process where they use the phone book or ask friends and then use services like ours.”
But Ridenour wondered whether people might be ready for a subscription model, where service is delivered regularly and preventively, without waiting for the moment of crisis. “The extraordinarily wealthy have estate managers,” said Ridenour. “They contract for these services and the services are delivered throughout the year.” Beyoncé does not call the plumber, in other words. Ridenour believed that a lot of the work that estate managers do could be automated—using data sets to predict when maintenance should happen, and using HomeAdvisor’s massive database of contractors to match people to the job. “Could we democratize estate management for the masses?” he asked.
Paying for upstream efforts ultimately boils down to three questions: Where are there costly problems? Who is in the best position to prevent those problems? And, how do you create incentives for them to do so? Ridenour’s argument seems reasonable: HomeAdvisor (or someone like it), not the homeowner, is in the best position to handle maintenance. Some homeowners are handy and some aren’t, but no individual homeowner can leverage the intelligence from thousands of homes about which specific types of preventive maintenance to employ. There is untapped value in the system: If major appliances could be prevented from failing too early, the resulting value could be divvied up between homeowners as savings and HomeAdvisor as profits.
Let’s apply these three questions to health care. Where is there a costly problem? One example is that Medicare spends a fortune paying for hospital visits that could have been prevented (for instance, if a patient’s diabetes had been kept under control). Who is in the best position to prevent those problems? It’s not hospitals—they don’t have a relationship with the patient before the emergency. Nor is it the patient, really, since a patient isn’t a health care expert. (Just as a homeowner isn’t a home maintenance expert.) The people best positioned to prevent those problems are primary care doctors. So how do you create incentives for them to do so? Meet the Accountable Care Organization (ACO), one of the models introduced in the 2010 Affordable Care Act.
Here’s a highly simplified description of one type of ACO (and trust me when I say there is an endless wormhole of complexity that lies beyond): A bunch of primary care doctors can join together to form an ACO, and Medicare says to the ACO: For the population of patients you serve, we know roughly how many hospital visits to expect this year, and how much those visits will cost us. So if you can reduce those visits by managing your patients’ health better, then we’ll share the savings with you.
“Before ACOs, doctors didn’t get paid a dime for keeping patients out of the hospital,” said Farzad Mostashari, the cofounder of Aledade, a company that helps doctors form ACOs. “In this model, it makes sense for the doctors to spend more time sitting with the patient and their family, rather than worrying about how they can see more patients per hour.”
I talked with Jonathan Lilly, a primary care doctor in West Virginia, and he said that the ACO model had transformed his practice. He sees fewer patients in a day—maybe 20, rather than 25 or 30—and spends more time with each. He and his partners have become less reactive and more proactive: They’re monitoring their patients’ blood sugar levels and blood pressure and weight, making sure those diagnostics are trending in the right direction. They’ve also become more accessible: If you want to keep patients from taking their problems to a hospital, you’ve got to be available to them. So they now offer evening and weekend hours, as well as a “fast track” visit in which a patient can show up any day, even without an appointment, and get a guaranteed visit.
“I’d never practiced this way,” said Lilly. “I’ve always wanted to be a family doctor, and always wanted to be the gatekeeper, and do it the right way. And this has allowed me to do it.” It’s working for Lilly and his partners. Their patients are healthier and happier with their care, and they’re going to the hospital less. As a result, Medicare saved money, and they shared the savings with the ACO, which meant Lilly got paid more.
There are other positive innovations in paying for upstream health. There’s an increasing amount of interest in “capitation,” a payment model used by health systems such as Kaiser Permanente, which has over 12 million members. Kaiser Permanente (KP) is unusual in that it’s both an insurer and a provider. If you’re a member, you pay your monthly premium (or your employer does) to KP, and when you get sick, you go to a KP doctor. This structure allows KP to avoid one of the long-standing tensions in the health care industry: Providers (like your doctors) want to bill insurers for as much as they can, while insurers want to pay for as little as possible, so there’s a constant tug-of-war over which procedures will be covered and how they’ll be reimbursed.
KP providers get paid a flat fee, per person served, to take care of all that person’s needs (on a risk-adjusted basis, so they’d get more to take care of an elderly person than a 25-year-old). That’s capitation. KP doctors don’t have an incentive to order an unnecessary MRI scan, because they don’t get paid any more for doing it. Now, why doesn’t capitation lead to cheating people of services? After all, the fewer services provided, the greater the profit to the provider. The corrective is that—as with Andy Grove’s “paired metrics”—they are also held accountable to quality-of-health metrics and patient-satisfaction metrics. So if they allow their patients’ health to deteriorate, or if those patients report being unhappy with the care they receive, the providers will make less money.
Capitation models open the door to upstream interventions, because they make it easier to justify spending money on prevention. At Geisinger Health System, based in Pennsylvania—another integrated system like Kaiser—diabetic patients are invited to use a “Food Farmacy”: basically a grocery store full of healthy food where they can shop and take home bags of food for free. Why would Geisinger give away free food? Because, to a diabetic patient, food is medicine. And for Geisinger, it’s worth paying for healthy food when it saves the patient from downstream complications that might be far more costly.
Our health system is inching toward a model with better incentives. The success of these efforts provides us a chance to reflect on the lessons of this section of the book. To prevent problems, upstream leaders must unite the right people (caregivers, insurers, patients). They must hunt for leverage points and push for systems change (unnecessary hospitalizations, ACOs). They must try to spot problems early (by, say, monitoring blood sugar levels). They must agonize about how to measure success—avoiding both ghost victories and unintended consequences. And finally they must think about the funding stream: how to find someone who’ll pay for prevention.
It’s quite a gauntlet of challenges to endure. It’s slow and painful. And it’s worth it. Because the scale is so great: 1% of the gargantuan $3.5 trillion health care industry is $35 billion—about the same as Nike’s 2018 global revenue. Tiny shifts in large systems can have powerful effects. So, together, by wading our way upstream, we can approach a world where the preservation of health is as valuable as the treatment of disease.
I. Here is an actual thing that happened with one of my relatives. Their dryer stopped working, and they sought advice from family members. A round of troubleshooting ensued but was fruitless. Then one final question was asked: “You’ve cleaned the lint filter, right?” [Silence.] “What’s the lint filter?”