Patient Engagement I

Asymmetric Thinking about Return on Investment

Article · March 11, 2016

Lately, we’ve attended many conferences about providing health care to patients with high medical and social needs — people with chronic illnesses who are frequently readmitted to the hospital. It seems as if every presentation refers to “return on investment” (ROI), which is invariably presented as a constraint — as in “Our program kept people out of the hospital, but we just couldn’t get the ROI to work.” Heads nod understandingly, and then participants move on to other topics.

At conferences about providing care for patients with cancer or other acute illnesses, by contrast, we almost never hear the term ROI. Instead, people talk about clinical gains, using understandable and patient-centered terms like “survival.” Though high drug prices are sometimes mentioned, no one ever says the ROI is prohibitive. No one mentions ROI at all.

ROI is the net profit of an investment (the money you got back minus the money you put in) divided by the money you put in. If you invested $100 and got back $110, you gained $10 and the return on your $100 investment was 10%. That’s good, as long as you can’t do even better by putting your $100 somewhere else. When people in health care colloquially say the ROI doesn’t work, though, they’re not saying they could make more money elsewhere; they’re saying they’re losing money. If your ROI equation’s numerator is negative — for example, if you put in $100 and got back $90 — there’s no way the ROI can work.

There is no obvious reason why ROI is more relevant to some clinical situations than to others. So why do we focus so heavily on ROI when the topic is chronic illness but rarely mention it when the topic is cancer? A huge amount of the cancer care we deliver provides such small personal and social gains that, were those gains monetized, the endeavor’s ROI would be deeply negative. And yet we ask, “What’s the ROI of that program that keeps chronically ill patients out of the hospital?” but not “What’s the ROI of treating advanced lung cancer?”

There are at least three reasons for this difference. One is that from the financial perspective of doctors and hospitals, the ROI of treating cancer is favorable. Reimbursements for cancer care are high in part because the political and popular value of cancer care is high, and those values are both revealed and reinforced by a history of largely cost-based fee-for-service pricing explicitly designed to at least meet providers’ costs. We can debate whether this kind of care is a worthy societal expense as compared with other worthy expenses — an exercise that might entail trying to match the financial ROI to a “social ROI” that reflects our values and includes everyone’s costs and benefits. But for now, the main reason the financial ROI is favorable for cancer care is that we have made such care profitable by setting high pay rates for it.

In contrast, the ROI of keeping chronically ill patients out of the hospital under current payment models is often unfavorable — which means you often lose money trying it. The amount of money currently devoted to keeping some patients out of the hospital and in alternative care settings is a fraction of the amount we devote to putting other patients in the hospital. Some might argue that one reason that cancer care is reimbursed so heavily is the presence of the same kind of political pressure that led to prohibiting the Centers for Medicare and Medicaid Services (CMS) from considering cost in coverage determinations. Efforts to help chronically ill patients receive the right level of care do not seem subject to the same pressures.

A second reason is that keeping people out of the hospital is hard — typically requiring care coordination with multiple services. Although treating patients with cancer is also hard, a long history of substantial investment in cancer care has helped hospitals hone their operations. Hospitals don’t have as much experience reducing demand for inpatient care as they do creating and supporting it. Merely implementing new financial incentives can’t make them turn their operations on a dime.

Third, providing cancer care and averting hospitalizations are financed differently. It’s hard to create a favorable ROI for reducing volume in a system dominated by fee-for-service payments for delivering care. Sometimes a favorable ROI is achieved passively when, for example, avoiding care frees up capacity for patients whose care is more profitable. More actively, the avoidance of care can be financed by establishing punishments for delivering avoidable care (penalties for readmissions, for example) or by shifting its cost to the providers themselves (e.g., through capitated or bundled payments).

It might seem that we could make the ROI for appropriate care more favorable if we imposed higher penalties on inappropriate care, just as we could make the ROI for treating cancer less favorable by paying less for cancer treatments. Despite that apparent symmetry, the choice of financing mechanisms — payments versus penalties — determines how much a health care goal will be advanced. If the ROI didn’t work for some form of cancer care — because the payment received was lower than the cost incurred — doctors and hospitals would almost certainly argue for higher payments. But when the ROI doesn’t work for keeping challenging patients with chronic disease out of the hospital, it’s implausible that doctors or hospitals will plead for increased readmission penalties. It would be an unusual health system executive indeed who said, “If CMS just penalized us more for readmissions, we would spend a lot more money on keeping people out of the hospital.” There isn’t any mathematical reason to prefer payment in the form of rewards over payment in the form of avoided penalties, but you can typically generate more advocates for your cause by paying people to follow you than by penalizing them for going the other way.

So when advocates and organizations devoted to keeping people out of the hospital lament their inability to make the ROI work, they should know that the game is thrice rigged against them. In the highly regulated context of health care, the amount and structure of financing are chosen rather than preordained. The ROI is favorable or unfavorable not because of the workings of some invisible hand, but because of choices someone — usually a private or public insurer — has made regarding what amounts will be paid for various types of care and what form payments will take.

What if the financing of cancer care and of efforts to achieve population health goals traded places? Suppose doctors and hospitals were paid for cancer care by capitation or bundles or through penalties for undesired outcomes and were paid directly and adequately to keep people out of the hospital. Oncologists might begin lamenting that although new approaches to cancer care helped patients, they just couldn’t get the ROI to work. And the outlook for population health might become less financially gloomy.

Rewards and penalties have the same ultimate effect on investment income, but they influence thinking in different ways. We might encourage greater effort and innovation in keeping people out of the hospital and coordinating care if we reframed its financing as positive payments for noble work rather than punitive revenue reductions. As U.S. health care financing begins again to shift risks to hospitals and physicians through bundled payments or readmission penalties, the financing of the care for our most challenging patients might be better shifted in the other direction.

Source Information

From the Perelman School of Medicine (D.A.A., R.W.M.) and the Wharton School (D.A.A., M.V.P.), University of Pennsylvania, and the Center for Health Equity Research and Promotion, Philadelphia VA Medical Center (D.A.A.) — both in Philadelphia.


This Perspective article originally appeared in The New England Journal of Medicine.

New Call for Submissions ­to NEJM Catalyst


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »

More From Patient Engagement I

Patient Inducements — High Graft or High Value?

Can health care be more like Uber and Lyft?

Consumer Engagement: New Tools and Capabilities for Health System Marketing

Novant Health is using research and data analytics to identify and engage new consumers, develop more relevant products and services, and support its mission and brand promise.

The Patient as Consumer and the Measurement of Bedside Manner

Many physicians do not believe patient satisfaction is a legitimate pursuit. Yet they must meet consumers of health care where they are — on Internet ratings sites.

Halpern02_clip_still: A/B Testing Health Behavior Interventions

A/B Testing Health Behavior

Health care can’t implement changes overnight like Silicon Valley, but we can still learn from their version of treatment and control.

Social Interventions Can Lower Costs and Improve Outcomes

We can better serve our under-resourced patients by helping them access treatments and social interventions we already know to be effective.

Transparency and Trust — Online Patient Reviews of Physicians

Transparency of validated data about delivery-system performance has the power to change the culture of health care.

Patient-Researchers and Physician-Consultants Confront Chronic Disease

Personal-use sensors and data applications allow patients with Type 1 diabetes and other chronic diseases to take charge of managing their health. This is the future of chronic disease management.

Geisinger’s Refund Promise: Where Things Stand After One Year

Geisinger’s radical refund promise at one year — no evidence of financial catastrophe but a good return on patient trust.

My Favorite Slide: One Size Does Not Fit All in Behavioral Interventions

Behavioral interventions can have very different effects in different portions of the population.

Adding Value by Talking More

Talk is cheap, but it’s also effective, especially when it comes to better patient care.


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »


Population Health — A Bipartisan Agenda…

A long-term agenda can be achieved only when state and federal governments agree on their…

Patient Incentives

51 Articles

The Power of Social Connectedness

We often don’t consider purpose in life and social connectedness when developing health care programs…

Coverage Expansion and Delivery System Reform…

Safety-net health system transformation is threatened by recent health reform proposals that erode coverage gains.

Insights Council

Have a voice. Join other health care leaders effecting change, shaping tomorrow.

Apply Now