In 2012, UF Health was awarded a 3-year, $600,000 CMS Low Income Pool grant to develop a multidisciplinary clinic (the Care One Clinic) that addressed the unique needs of super-utilizing, high-cost patients. By taking the time to listen to our super-utilizer patients, we identified their often hidden barriers to care. Identifying and addressing these barriers benefited both the patients and our health system. We observed an overall 25% reduction in super-utilizer hospitalizations and a reduction in uninsured hospital admissions that covered the cost of the clinic.
A multidisciplinary approach involving treatment of pain, mental illness, and substance abuse, along with adequate time to listen to patients and earn their trust, are key ingredients to successful super-utilizer care.
It's important to build a financial case to justify the cost of this additional super-utilizer care. In our situation, a reduction in uninsured inpatient days permitted additional insured admissions to generate revenues that exceeded the cost of the clinic; other justifications might include reduced readmissions, reduced drug costs, or reduced uninsured ED use. Such care might also help providers achieve quality and continuity-of-care targets under certain value-based care arrangements.
An analysis by the Agency for Healthcare Research and Quality showed that 5% of U.S. patients account for 59% of U.S. health care costs. Subsequently, there has been a proliferation of programs across the country focusing on the complex needs of super-utilizer patients, many of whom have dual diagnoses of chronic medical conditions and mental illness or substance abuse.
Over the grant period, UF Health super-utilizers represented 6% of the population but accounted for 23% of ED visits and 20% of total clinical costs.
Our super-utilizers fall into two categories: those who visit the ED frequently, and those who have complex post-discharge needs. About 40% of our super-utilizers were uninsured, compared with 12% of patients hospitalized at UF Health. (Over the 3+ years that the clinic has been open, the proportion of its patients that are uninsured has risen to 50% to 70%.)
Our aim was to improve coordination and quality of care for super-utilizer patients in a financially sustainable manner.
Beginning in November 2012, we invited patients in two categories to our clinic. The first were patients who were discharged from the hospital with ≥4 ED visits in the prior 6 months. The second were uninsured and underinsured patients with complex post-discharge needs. The clinic served as an “ambulatory intensive care unit” for close monitoring of patients with a plan for eventual graduation to local primary care clinics. The clinic was open 5 half-days per week. Upon each patient’s initial visit, a social worker completed a needs assessment that identified barriers to care related to housing or transportation, and screened for substance abuse, depression, low health literacy, and insurance eligibility. A clinical pharmacist then reconciled medications, provided patient education, monitored opioid use when needed, and determined affordability of medications. Finally, a hospitalist and an addiction psychiatrist optimized medical management of chronic medical conditions, addressed untreated mental illness, substance abuse, and pain, and provided subspecialty referrals.
Clinic scheduling was planned with the expectation of a 50% no-show rate and a goal to see 5 patients per half-day. These target expectations were largely met. We sought to see ED “frequent flyers” for at least 3 to 5 visits to identify and address the reasons for their repetitive ED use. We sought to move post-discharge patients to appropriate primary care providers as quickly as possible, though uninsured patients usually remained in our care until they obtained public funding.
We found that the most effective way to reduce super-utilizing behavior was to identify what factors in a patient’s life may be contributing to that behavior, and find ways to address them. The Care One Clinic offers a unique opportunity for staff to sit quietly and listen, away from the bustle and distractions of the ED or the inpatient unit.
We discovered some of our patients could not read, they lacked transportation, had uncontrolled pain, or were depressed or lonely. These were matters that patients did not easily disclose, but once we fully understood them, we could begin to design care plans that actually worked.
As we earned their trust, patients shared stories of tremendous grief and pain. They became more than “super-utilizers” in our eyes. They were people who had suffered greatly and, thus, sought solace in the health system, a place where nurses, case managers, and doctors had become like family and a source of support they lacked in the community. Ms. D, who had a history of tampering with her pain pump while hospitalized, described the pain of losing her 3-year-old son in a car accident and how it led to an opiate addiction. Her care plan involved enrollment in structured addiction treatment along with treatment of depression.
Often we could not impact compliance with medications until social factors were addressed. Ms. G, an uncontrolled diabetic who was defensive and paranoid, later revealed prior child abuse and a current abusive relationship. Defining an escape route from the abuse was the crucial first step to decreasing admissions.
Drilling down to the source of the problem was the key for several patients. Mr. F, who presented frequently with hyperglycemia, was found to have poor vision and low health literacy. Discovery of cataracts led to cataract surgery which improved his ability to self-administer insulin, and, subsequently, hospitalizations ceased.
Likewise, our pharmacist often found that patients had not picked up medications, yet reported compliance. One diabetic patient was not refrigerating insulin, and a COPD patient was using empty inhalers. One patient was not taking her medications because she couldn’t see the labels well enough to fill her pillbox correctly. She immediately felt better after the pharmacist began filling her pillbox at each visit. This simple intervention preserves her health while we seek to get her the vision care she needs.
The Care One Clinic is located within UF Health Shands Hospital and is staffed at any given time by a social worker, a hospitalist, an addiction psychiatrist, a clinical pharmacist, and a nurse. Overall, about 15 people share the clinical duties.
Between November 2012 and October 2014, we enrolled 635 patients: 186 super-utilizer and 449 post-discharge patients. Ninety-three percent of our patients had a mental health diagnosis and 66% had a substance use disorder. The super-utilizer group had a 25% reduction in hospitalizations, a 23% reduction in hospital days, and an 11% reduction in ED visits in the 6 months after enrollment compared to the 6 months prior to enrollment (see Figure).
We reduced unfunded hospitalizations for the Care One patient group by 206 during this period, representing about $1.8 million. These savings would have more than covered the $700,000 operating cost of the clinic during the same time period.
Where to Start
Identify super-utilizers and analyze their demographics and medical profiles to define the scope of the problem.
Elements may include payer mix, mental health history, prior substance abuse, and level of health literacy. Adjust the enrollment criteria as necessary to justify the funding. (For example, we used 4 or more ED visits in 6 months as one of our criteria; other providers might be able to justify a program for patients with 3 or more.) Pick an appropriate care team for the needs of the super-utilizer group, assign appropriate space, and create a schedule that allows time to learn about and address patients’ medical and social barriers to care.
Chronic pain and addiction were significant drivers of ED visits in our health system, so the provision of on-site pain and addiction services were important elements of our intervention. However, we needed to identify patients’ underlying issues in order to create sustainable interventions.
- Persuading patients to come to the clinic. Most did not respond to telephone contact, and we found it best to make the initial contact while they are in the hospital or the ED. Using information from the EMR, we identified which ED or inpatients were likely candidates, and then we screened them to confirm and offered them an initial Care One Clinic appointment.
- Addressing transportation. We had hoped to provide transportation vouchers, but legal considerations prevented us from doing so. We identified a third party through United Way that provided our patients with transportation.
Marvin Dewar, MD, JD, CEO and Senior Associate Dean, University of Florida Physicians
Robert Hromas, MD, FACP, Professor and Chair, Dept. of Medicine, UF & Shands
Paul Lipori, Director, Financial Planning and Analysis at UF Health Shands Hospital
Jon Shuster, PhD, Professor, Biostatistician, UF Clinical Research Center
Research reported in this publication was partly supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR001427. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The Care One Clinic was funded by a CMS Low Income Pool Primary Care Award (Medicaid Number 053386600). The contents of this report are solely the responsibility of the authors and do not necessarily represent the official views of CMS.
This case study originally appeared in NEJM Catalyst on August 24, 2016.