'Hotspotting' Hospital Care Fails to Cut Readmissions in Randomized Trial

Debra L. Beck

January 08, 2020

A multidisciplinary clinical and social intervention program aimed at patients deemed to be "superutilizers" of the healthcare system failed to make a difference in 180-day readmission rates in a randomized, controlled trial.

The study casts some doubt on whether such "hotspotting" might have promise as a way to reduce burgeoning US healthcare costs, as has been hoped. The term refers to a process of predicting, identifying, and addressing the needs of superutilizers to improve their health outcomes and lessen their use of healthcare resources.

Such an intervention, compared to a standard-care control program, showed no benefit in terms of hospital readmission rates at 6 months, the total number of readmissions, or the proportion of patients with more than two readmissions. The study was published December 8 in the New England Journal of Medicine.

The rates of readmission at 6 months, the primary endpoint, were 62.3% for the intervention group and 61.7% for standard care, note the authors, led by Amy Finkelstein, PhD, an economist at the Massachusetts Institute of Technology, Cambridge.

"This nonsignificantly lower rate of readmissions is the substantive finding of our trial, but for me, the most important finding is the second one — that if you just do a pre- and postintervention study, it provides a very misleading picture that suggests the intervention is working," Finkelstein told theheart.org \ Medscape Cardiology.

The second analysis of the intervention, which measured readmissions 6 months before and 6 months after the hotspotting intervention, showed a 38% decline in admissions. That seems favorable for the intervention, but there was a similar decline in the control group. And in both groups, admissions rose sharply in the 6 months before the intervention and fell rapidly afterward.

It's a classic story of regression to the mean, the phenomenon that makes natural variation in repeated data look like real change, Finkelstein observed: unusually large or small measurements tend to revert to values closer to the mean over time.

"You're intervening with people who are, by definition, in unusually poor health or who accrue unusually high healthcare costs. That's what triggers the intervention," she said.

The concern, Finkelstein added, is that because of the natural tendency for trends to revert to the mean, "outcomes that are measured as extreme in one period just naturally become less extreme over time," Finkelstein explained.

Both Clinical and Social Components

The term "hotspotting" became a buzzword after Atul Gawande's 2011 New Yorker article profiled Jeffrey Brenner, MD, formerly a Camden, New Jersey, family physician who was looking for a way to improve the health of his patients, many of whom qualified as high-complexity, high-cost superutilizers, Finkelstein and colleagues write.

Brenner, recognizing the need for a new way for hospitals, providers, and community residents to collaborate, founded the Camden Coalition of Healthcare Providers in 2003, they note.

It developed the Camden Core Model used in the current study, a care-transition program designed to improve health and reduce hospital use by some of the sickest patients in the United States.

The intervention, already rolled out in other cities around the country, included a time-limited but intensive and multidisciplinary intervention with both clinical and social components, the report notes.

It called for a team of registered nurses, social workers, licensed practical nurses, community health workers, and health coaches to coordinate comprehensive follow-up care at home and clinic visits and help patients take advantage of social services and behavioral health programs.

The trialists estimated the cost of delivering the care management program to be about $5000 per patient, 80% of which went to salaries for the staff performing triage and delivering services.

Finkelstein declined to speculate on why the intervention failed to show benefit, but the results were a bit of a surprise to Rishi Wadhera, MD, a cardiologist and investigator at the Smith Center for Outcome Research at Beth Israel Deaconess Hospital, Boston.

"The hotspotting program evaluated by Finkelstein and colleagues was an incredibly intensive care-transition and management program, certainly one that I would have expected to improve outcomes...among medically and socially complex patients," he told theheart.org | Medscape Cardiology. The study highlights how difficult it can be, even with such a program, to reduce admissions in such patients, he said.

But Leah Marcotte, MD, an internist at the University of Washington, Seattle, said the findings are not so surprising, given the double-whammy challenge of managing both medical and social complexity.

"Managing medical complexity, even when patients have good social support and are well resourced, is very hard to do. Managing complex medical conditions when patients do not have stable housing or a reliable way to communicate with the health care team, two barriers specifically called out by the authors in this study, is exceedingly more difficult," Marcotte stated in an email exchange with theheart.org | Medscape Cardiology.

"Addressing social determinants like food insecurity through connecting patients with SNAP [Supplemental Nutrition Assistance Program] benefits as was done in the intervention likely does help, but it is probably insufficient to make large impacts on patients' health," she added.

Test Drive Before Buying?

The trial provides fuel for arguments made by policy experts, including Wadhera, who have derided the government for rolling out new cost-containment policies, including the Hospital Readmissions Reduction Program (HRRP), without first testing them in randomized trials.

"One has to wonder how many nationally mandated programs that aim to improve quality and reduce spending, but haven't been subjected to the same type of rigorous evaluation as the hotspotting program in this study, have simply imposed more administrative complexity, cost, and burden on our healthcare system without meaningfully improving patient care," Wadhera said.

"If the HRRP had been implemented in a staged or randomized manner, we would probably know with more certainty whether the program has been beneficial, ineffective, or harmful to patients."

Wadhera and others have spoken out repeatedly on this issue. At best, "bad" health policies are ineffective at reducing costs. At worse, they put lives at risk.

Recent studies have suggested that the magnitude of reduction in readmissions credited to HRRP might be overstated and that implementation of the policy might be associated with increased mortality among patients admitted with heart failure and pneumonia.

The current study "should inspire policy makers and the Centers for Medicare & Medicaid Services to roll out national programs that aim to improve quality and reduce spending in a randomized manner so that we can understand whether they are truly effective," said Wadhera.

Finkelstein said that although it may seem obvious that a comparison should be made between an intervention arm and a control arm, even if the study is observational in nature, in actual practice this is often hard to do.

"It can be extremely hard to find an appropriate comparison group, especially when you're looking at extremes, such as these very sick, high cost patients. And this is why you do randomized control trials to ensure a proper comparison group," she said.

For example, in a complaint lodged against the Center for Medicare & Medicaid Innovation (CMMI) Bundled Payments for Care Improvement (BPCI), studies of the effectiveness of the program have compared hospitals voluntarily opting into the program with those that did not volunteer, with no attempt to adjust for potential differences between those institutions.

One of the trial's sponsors was J-PAL (Abdul Latif Jameel Poverty Action Lab), a global research center that applies rigorous research and conducts randomized evaluations of policy, health and otherwise. Finkelstein is the J-PAL North America's scientific director.

This study was supported by the National Institute on Aging of the National Institutes of Health, the Health Care Delivery Initiative of J-PAL North American, and the Sloan School of Management of the Massachusetts Institute of Technology. Finkelstein, Wadhera, and Marcotte have disclosed no relevant financial relationships.

N Engl J Med. Published online January 8, 2020. Abstract

Follow Medscape on Facebook, Twitter, Instagram, and YouTube.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.