“While most providers are still making money by doing more, safety-net providers spend or lose more the more that we do,” says Frederick Cerise, President and CEO for Parkland Health and Hospital System.
He describes how several years ago, at a different health care system, one doctor wanted to install hyperbaric oxygen therapy machines to treat foot wounds. Because a high percentage of the patients were uninsured, that would mean paying for each expensive unit. Cerise found the evidence sparse as to whether this therapy was better than simpler approaches at the time. And though a major hospital advisory group approved of the therapy, their rationale included the number of insurance companies paying for it and that it would generate additional volume.
“I thought at the time, this is exactly why we don’t want to do this, and we never did institute that program, which gives you some insight into the mind of the safety net,” he says. With nearly half of patients uninsured and another 30% on Medicaid, they weren’t making money based on volume. It’s in patients’ best interests to cut overuse or potential overuse in the system, and there’s an economic incentive to cut, too.
“It’s a phenomenon in U.S. health care that in many major cities, a disproportionate share of people who are uninsured, on Medicaid, or otherwise financially challenged are concentrated in a publicly supported system of care,” says Cerise. Characteristics of safety-net providers include:
- High demand for care. Despite insurance coverage expansions in the past decade, about 30 million people are still uninsured and have limited options for care. “In most health systems, volume is your friend. In the safety net, volume is your responsibility,” says Cerise. Per case payments are below cost, so there’s no economic incentive to do more and to generate more volume. Instead, in the safety net, it’s about managing volume and demand.
- Complex interplay between health care and social factors. Homelessness or subpar housing, limited food options, transportation challenges, lack of personal or childcare support, and other social issues impact how safety-net patients receive or don’t receive health care.
- Limited ability to generate revenue outside public sources. Less than 10% of Parkland’s revenue comes from commercial payers; the system cannot generate a profit there to pay for its low-income population. “We don’t pursue business strategies based on commercial viability. We will pursue based on the needs of the population,” he says.
- Flexible funding. A significant portion of funding for safety-net providers comes from public sources structured more like capitated payment than fee-for-service payment — they aren’t tied to codes or RVUs. “The pressure then is to figure out how to provide reasonable access to beneficial care regardless of the reimbursement method, and not how to provide more care that’s reimbursed at a level to generate a profit,” he says. “This often drives low-tech solutions.”
“The pressures of the system and the opportunities line up in a way to promote an approach to care that has been long advocated by health care policymakers,” Cerise says. He notes how in the 1990s, Clayton Christianson described a model of disruptive innovation characterized by transformative changes that are lower tech, simpler, cheaper, and more convenient. These disruptive innovations are not pursued by mainstream firms focused on advancing newer technologies, but rather by those targeting a market less interesting to leading companies — in this case, the uninsured. Safety-net providers tend to focus on populations with basic needs struggling to be met.
“More important than access to the latest technology for any of our patients is just simply access to care,” Cerise says. “It should be no surprise that changes to the delivery system that would result in care that’s simpler, cheaper, and more convenient would come out of safety-net providers.”
He provides some examples. Years ago, when safety-net hospitals were faced with the challenge of nowhere to refer uninsured patients after discharge, they developed clinics to accommodate these discharges. Over time, as preventive care and chronic disease management improved, these clinics became the main attraction. In 2018, Parkland will have about 75,000 hospital discharges compared to over 1 million outpatient visits.
In the ’90s, when many hospitals were acquiring physician practices to feed them with inpatients, safety-net providers were developing clinics to do the opposite — reduce admissions. “In a lot of regards, many of the safety-net systems were accountable care organizations before ACOs existed,” Cerise says.
About 5 years ago at Parkland, it was the norm for someone without insurance who had a complicated infection to stay in the hospital for 4 to 6 weeks to complete a course of IV antibiotics. Someone with insurance would be more likely to complete their antibiotic course at home, with the help of a home health nurse or at a freestanding infusion center.
Kavita Bhavan, an infectious disease doctor, developed a program to teach patients how to administer their own IV antibiotics using a low-tech solution: hanging the IV bag from a coat hanger suspended from a hook on the wall, even one already in place to hold a picture frame. Patients with no medical training can safely administer IV antibiotics to themselves that way, counting the drops per minute to establish the correct flow rate.
After 3 years, Parkland saw a savings of over 27,000 bed days, avoiding over $35 million in costs for these patients. And after 5 years, over 3000 patients had gone through this program. “We have the data to show that it’s safer and a simpler and more convenient way for patients to experience their care,” Cerise says.
“As with other disruptive innovations, we expect that over time those changes will migrate into more mainstream systems,” he says. “We’re already seeing signs of that, whereby patients with insurance and the option of having a nurse come to their house to help them with the antibiotics are choosing to deliver the antibiotics themselves.”
Parkland’s outpatient clinics face similar pressures of high demand, which can translate into prolonged wait times. In response, gastroenterologist Deepak Agrawal recognized that many of his GI clinic patients could be treated by their primary care provider if given more guidance.
Because referral questions don’t require a face-to-face visit, Agrawal developed an email consultation program in which specialists and PCPs share an electronic health record. PCPs can email their consult to the pool of GI specialists, and within 72 hours get their answer back. The pilot proved successful. Email consultation now handles over 60% of referrals to Parkland’s GI clinic, and they’re rolling out the program to other clinics, as well. “The patients get the care they need without the hassle of another visit to a specialist,” says Cerise.
No reimbursement model exists for this program. “We don’t get paid to do it that way,” says Cerise. “We use our public support to do what makes sense — in this case, support the email consultation and not be constrained by the rules of the traditional health care system that says you have to have a face-to-face visit to generate a billable encounter.”
To be more efficient, Parkland has had to learn how to better target care, “to provide more care to people who are at higher risk, and less care to people who need less support.” A group at Parkland led by Ruben Amarasingham developed algorithms to predict who would be, among other things, at high risk for readmission after hospital discharge. Amarasingham’s team used clinical and administrative data for patients with congestive heart failure to look for risk factors. What they found was that the number of home addresses someone had in the past year was a higher predictor of readmission than whether someone received good discharge instructions or whether they left with a follow-up appointment in hand.
“When we know that the number of addresses someone had in the past year is a high-risk factor, it focuses your interventions differently,” says Cerise. Using this information to develop targeted care management, Parkland is now seeing readmission rates lower than the national average for this population with limited social supports.
“The challenges in the safety net are no different than the health care problems facing the U.S. health care system as a whole. Although we have some unique circumstances and they require some innovative answers, the overall approach to care at the population level has generalizable lessons that can be useful for the health care system as a whole,” Cerise concludes.
“The pressure is on all of us — not just the safety net, but on everyone — to reexamine how we care for the entire population, with the understanding that improved targeting of resources has to be a consideration if we’re going to ensure care for everyone.”
From the NEJM Catalyst event Essentials of High-Performing Organizations, held at the University of Michigan’s Institute for Healthcare Policy and Innovation, July 25, 2018.