Health Care Hotspotting in the United States
The high concentration of health care spending in a small share of the population has generated interest in interventions that can reduce health care spending and improve health care quality by targeting “super-utilizers” of the health care system. In this study, researchers evaluated the impact of a well-known care transition program that provides intensive, time-limited clinical and social assistance to “super-utilizers” of the health care system with complex needs. Researchers found no impact on the rate of six-month hospital readmission.
Health care spending in the United States is heavily concentrated. Five percent of the population accounts for 50 percent of annual spending; one percent accounts for almost one quarter of annual spending. This has generated interest in how to reduce costs and improve the quality of care delivery through interventions targeting patients with very high use of health care services. Also referred to as “super-utilizers,” this small group of individuals with medically and socially complex needs and frequent hospitalizations accounts for a disproportionately large share of health care costs.
There are a number of promising observational studies of interventions targeting super-utilizers. However, regression to the mean—the tendency for patients selected as exceptionally high-cost patients at a moment in time to move closer to average cost over time—may lead observational studies of super-utilizer programs to produce misleading estimates of program effects.
The Camden Coalition of Healthcare Providers (the Coalition) uses real-time data on hospital admissions to identify super-utilizer patients, an approach referred to as “Hotspotting.” Their program—the Camden Core Model— is a time-limited, intensive care transition program that targets super-utilizer patients living in Camden, New Jersey.
In 2012, the median household income in Camden, New Jersey was $25,681 in comparison to $69,667 for the state of New Jersey overall; two-fifths of Camden residents live below the federal poverty line.1 The study population represented less than one-half of one percent of the Camden population but accounted for 11 percent of Camden hospital expenditures. Prior to their index admission, approximately three-quarters of the study population were single, divorced or widowed, one-half had less than a high school degree, and three-fifths reported needing help with mobility. Almost the entire study population (95 percent) was not employed, and 40 percent had been diagnosed with a substance abuse disorder during the index admission. Forty-five percent of the study population was covered by Medicaid as their primary insurer.
The Camden Core Model has received national attention as a promising super-utilizer intervention, and federal funding has expanded versions of the model to other cities. To rigorously investigate their nationally-recognized program in Camden, the Coalition partnered with the researchers in 2014 to design a prospective randomized evaluation.
In this study, researchers evaluated the impact of the Camden Core Model on subsequent health care utilization. Patients who were eligible and consented were randomly divided into two groups: an intervention group that received the program and a control group that was not enrolled in the Camden Core Model and received the usual standard of care.
The time-limited intervention included intensive clinical and social components. Intervention group patients were enrolled while in the hospital. Upon return home, patients were engaged by a multidisciplinary team of registered nurses, social workers, licensed practical nurses, community health workers, and AmeriCorps health coaches. The team worked with each patient to address individualized needs as they transitioned from being in the hospital to being at home. Team support services included accompanying patients to initial primary and specialty care visits, coordinating follow-up care and medication management, conducting blood pressure and blood sugar checks, coaching patients in disease-specific self-care, and helping patients apply for social services and appropriate behavioral health programs. On average, an intervention group patient received approximately eight home visits; median program duration was 92 days.
Eligibility was limited to adults (ages 18-80) living in Camden. Additional criteria for inclusion in the study were: at least two hospital admissions in the last six months at any of four Camden-area hospitals, at least two chronic conditions, and at least two of the following: five or more active outpatient medications, difficulty accessing services, lack of social support, a mental health comorbidity, an active drug habit, or experiencing homelessness.
The intervention had no statistically significant impact on the rate of hospital readmission within six months of hospital discharge. Intervention group patients had a statistically insignificant 0.8 percentage point increase in six-month hospital readmissions (relative to the control group rate of 62 percent). There was also no impact on readmissions over shorter (one month) or longer (one year) time frames, or on mortality.
The results suggest challenges in reducing hospital readmission rates for a medically and socially complex patient population with very high health care utilization rates. While randomized evaluations of some other care transition programs have found that they reduced readmissions by as much as 15-45 percent in the Medicare population, the Camden Core Model served a very different population: younger, with more diverse medical needs, greater social complexity, and much higher health care utilization (prior hospital use was nearly twice that of patients in most previous successful care transition programs).
The study results also underscore the importance of using randomized evaluations to measure the impact of interventions targeting patients with high health care utilization rates. While the results from this study indicate no impact on the rate of hospital readmission, a study comparing readmission rates within the intervention group alone before and after participating in the program would have suggested substantial reductions in readmissions. This is because super-utilizers are typically targeted for intervention during periods of high health care utilization, often when their illnesses are flaring, and many such patients will naturally return to lower rates of utilization. This phenomenon, known as regression to the mean, can lead observational studies to produce misleading estimates of program effects.
Finkelstein, Amy, Annetta Zhou, Sarah Taubman, and Joseph Doyle. 2020. "Health Care Hotspotting - A Randomized, Controlled Trial." New England Journal of Medicine 382 (2):152-62.