Health Policy$ense

Don’t Throw Cold Water On Health Care’s Hot Spotters

[Original post: Shreya Kangovi and David Grande, Don’t Throw Cold Water On Health Care’s Hot Spotters, Health Affairs Blog, February 11, 2020. https://www.healthaffairs.org/do/10.1377/hblog20200205.342657/full/: Copyright ©2020 Health Affairs by Project HOPE – The People-to-People Health Foundation, Inc.]

The phrase “social determinants of health”—long a concept among population health scientists—first caught fire in the health care community about five years ago. It signaled a newfound interest in the idea that health care ought to broaden its focus to include the social factors—homelessness, hunger, isolation—that lead to poor health and high health care costs. Despite hyper-partisan debates around health care, many agreed on the idea that if health care addressed upstream social problems—instead of just waiting for people to show up in the emergency department—Americans could live better and longer at lower cost.

A few weeks ago, the New England Journal of Medicine published a trial of one very high-profile social care program: Dr. Jeff Brenner’s “hot spotters” in Camden, New Jersey. The trial found that the Camden program—in which a team of doctors, nurses, and social workers provided targeted case management for super utilizers with complex social and medical needs—did not reduce costly hospital readmissions.  

The immediate reaction by many has been to conclude that the patients are too complex, and the problems are outside of what health care can address without massive overhauls of social policies, requiring political will that does not exist. In short, the same people who touted social care programs in health care yesterday appear ready to throw cold water on them today.

This is the wrong conclusion.

While the Camden trial was disappointing, social care programs by health care organizations can be successful even among complex patients in very poorly resourced communities. The best scientific evidence (some described below) shows that programs can reduce hospitalizations and save money, while also improving health and the quality of health care. It is also important to consider outcomes beyond health care use—the trial of the Camden program did not measure how it may have impacted other aspects of health or experiences with health care. 

To understand the takeaways of the Camden trial, it’s useful to retrace the narrative.

Storytelling And Science

Atul Gawande’s 2011 New Yorker piece featuring the Camden program, in many ways, sparked the national conversation about the social determinants of health. The public’s imagination was captured by stories of patients such as Frank Hendricks, a man who spent “as much time in the hospital as out” thanks to his diabetes, cocaine addiction, and unstable housing. Our heads nodded as we read about Dr. Brenner taking over Hendricks’ care, employing a nurse to watch over his blood sugars while a social worker helped him apply for disability insurance and return to church. It came as no surprise that Mr. Hendricks got clean, lost 200 pounds and stayed out of the hospital. We believed.

And so, reporters, health systems, and policy makers told story after story of programs that fed, clothed, and housed the Mr. Hendricks of America. Our belief was reinforced by a particularly tricky way of measuring success: before and after studies. It turns out if you start out with outliers (as many social care programs do) and follow them over time without doing anything at all, their rates of hospitalization drop. 

What Brenner and the Camden Coalition did next was admirable: They embraced real science. Partnering with expert researchers from the Massachusetts Institute of Technology, they conducted a randomized controlled trial—the type of gold-standard study we use to test new drugs or medical devices. They had everything to lose but wanted to know: Is this particular way of providing social care effective at reducing costly rehospitalization?

The irony, of course, is that their precise science is being met with the same storytelling and sweeping conclusions. Incarcerated, addicted, homeless patients such as Mr. Hendricks are now being used as cautionary tales to conclude that helping such people is a near impossible task without—unlikely—overhauls of economic and social policy.

But it’s not impossible, and the same randomized controlled trial science shows us pathways that work. 

Proven Pathways

For example, over the past decade, we’ve found that specially hired and trained community health workers—trusted, empathic laypeople from local communities—can support very complex patients just like Mr. Hendricks. Three separate randomized trials conducted in Philadelphia—the poorest big city in the United States—found that this improves the quality of health care and reduces total hospital days by two-thirds, returning $2.47 annually for every dollar invested by Medicaid. The program is now being implemented nationally by providers, payers, Veterans Affairs, state and local governments, and retailers.

In Chicago, Laura Sadowski and her colleagues hired social workers to coordinate housing and social services for homeless, hospitalized patients. Their trial found a 29 percent reduction in hospitalization and a 24 percent reduction in emergency department visits. David Olds’ Nurse Family Partnership program sends nurses to visit with high-risk pregnant women—unemployed, unmarried, and with fewer than 12 years of education. Two decades of research have demonstrated that this program not only saves the lives of those mothers but their children as well.

Each of these programs would probably be even more effective if coordinated with supportive policies that lead to employment opportunities and a supply of stable housing. But it’s important to know that health care organizations can deploy them with good effect even in the realities of today.

Calling all social care programs ineffective is like saying: “Pills are bad.” Social care programs are neither simple nor homogenous, and there is as much nuance to each one of these programs as there is to any cholesterol medication. There are many reasons that outcomes can differ across programs—patient selection, program theory and design, personnel, among other factors. Fortunately, implementation science is an entire field dedicated to unpacking these types of factors. And when programs—such as the Camden program—use implementation science they can be adapted and refined as part of a cycle of discovery.

One Step Closer

The Camden Coalition program was at the forefront of the modern movement to link social interventions to health care delivery. They have contributed enormously to the science and continue to adapt their own program in response to recent findings. Just as biomedical science has generated newer and better pills for cancer and heart disease, sociobehavioral science will march us toward better solutions to complex problems such as health inequity and inefficient spending. The Camden trial was not a defeat, but a victory because it gets us one step closer to a goal that is hard, but not impossible.