In April 2011, the leadership of the Massachusetts General Hospital (MGH) and the Massachusetts General Physicians Organization formed five care redesign teams charged with identifying new, higher-quality and more efficient ways to care for patients. For patients, the promise of care redesign is providing a better experience and doing so at a lower cost. For the institution and the providers, the promise of care redesign is to position MGH to succeed and thrive as health care transitions from the fee-for-service model to an accountable care model. The traditional fee-for-service model has been criticized for leading to fragmentation of care and excess utilization. The proposed future model is integrated payments that will require providers to deliver more integrated care with more thoughtful utilization. Further complicating matters, there will be no clear transition point where all patients transition from the old model to the new one. Rather, we will simultaneously treat some patients with fee-for-service plans and others with accountable care plans. In other words, we will have each foot in a different canoe.
Our team was tasked with redesigning back pain care at our institution. Our instructions were to deliver more integrated patient-centered care – to increase patient affordability while protecting our mission – to develop and drive performance metrics to demonstrate an unparalleled patient experience, outcomes and value, and to improve the reputation of the institution. As we formed our team, we agreed on several core principles. We cannot provide high-quality care if we are not efficient, so we need to make a value case. We need to thrive in whatever payment environment comes our way, and high-quality, patient-centered efficient care will do us well under any scenario. Frontline is where care happens and must be the locus of our work. This undertaking would be the most important activity we would engage in over the next 18 months, and our work product must be sustainable, create habits and improve over time. The leadership of our organization made it clear that this work will be a “must have” experience for future leaders.
The initial scope of our assignment was extremely focused. A review of the average length of stay for nonsurgical back pain patients at MGH revealed that the average length of stay at our institution was six days compared to 3.5 days at other academic medical centers. This excessive length of stay resulted in 300 additional inpatient days per year. The vast majority of these patients are cared for by our hospitalist service. Since our hospitalist service is not regionalized, these patients were admitted to units all over the hospital and frequently occupied beds on surgical floors. Since our inpatient units almost always operate at or near full capacity, there was great interest at our institution in decreasing this length of stay so that elective surgical cases would not have to be canceled due to lack of available beds.
When we looked at our initial focused assignment, we found that only about 200 patients per year are admitted for nonsurgical back pain. We then looked at how many patients come to MGH each year with a primary diagnosis of back pain. We found that we see about 9,500 patients annually with a primary diagnosis of lumbar disk displacement, lumbosacral disc degeneration, lumbosacral spondylosis, lumbar spinal stenosis or lumbago. At that point, we expanded our team’s focus to include all of those patients as well. When it came to building our team, we included two internists, a hospitalist, an emergency medicine specialist, an orthopedic spine surgeon, a neurosurgical spine surgeon, two interventional radiologists, two pain medicine specialists, two physical therapists, a nurse practitioner from occupational health, a social worker and a case manager. We also included an administrative lead, a project manager, a practice improvement consultant and a financial analyst.
With that new expanded focus, we set out to define the patient population and the episode of care. We made a plan to understand the current state by mapping the current care process, identifying care redesign opportunities such as reducing unwarranted variation and reducing overutilization, and incorporating the patient perspective. We also set out metrics to track implementation compliance and to measure success and value.
We started by conducting interviews of all members of the extended team that cares for back pain patients at MGH. Among others, we interviewed primary care physicians, emergency medicine physicians, neurosurgical spine surgeons, orthopedic spine surgeons, physiatrists, pain medicine specialists, interventional radiologists, physical therapists, case managers and social workers. With their help, we mapped current workflows and identified problem areas. We analyzed hospital visit and cost data based on hospital estimates of actual incurred direct costs. In addition, we analyzed professional visits with cost and billing data, and relative value units.
We found that our average back pain patient is 55 years old and that males and females are represented almost evenly. Fewer back pain patients have government insurance than the MGH average (41 percent for back pain versus 51 percent MGH average). Ninety-five percent of our back pain patients are local to Massachusetts. We found that almost all of our back pain care, 98 percent, is provided in an outpatient setting. Although almost all of the care takes place in the outpatient setting, outpatient care only accounted for 45 percent of the cost of back pain care. Two drivers of cost were orthopedic spine surgery, accounting for 34 percent, and neurosurgical spine surgery accounting for 15 percent. Comparing orthopedic spine surgery and neurosurgical spine surgery, the costs per procedure were very similar between services on like procedures. Spinal implants drove the substantial difference in cost between procedures. When our perioperative products committee renegotiated spinal implant pricing, that change generated $4.5 million in annual savings.
When we focused on outpatient care of back pain patients, injections accounted for 35 percent of the cost, physical therapy 21 percent, diagnostic radiology 20 percent and office visits 13 percent. The average back pain patient had 2.8 visits to MGH over two years and received care costing roughly $1,500. Roughly one quarter of our back pain patients received physical therapy, and the average patient had four physical therapy visits over two years. Similarly, one quarter of all back pain patients received some sort of injection, and the average injection patient had about three injections over two years. Just over half of all back pain patients received some form of imaging study of their back, and the average back pain patient underwent 0.7 imaging studies over two years. Lumbar spine X-rays were the most common examination followed by lumbar spine MRIs.
When we examined patient flow through the various services that treat outpatient back pain at MGH, we found that numerous specialties offer the same services. For patients and referring providers, this overlap created great confusion about where to refer patients and when. In many instances, referral patterns were highly variable and based on personal relationships or past experiences. In some instances, treatment for the same condition varied depending on which specialty the patient encountered first. In other instances, patients were referred to specialists for consultations that turned out to be unnecessary. Our surgeons, in particular, were seeing large numbers of patients with back pain and no indications for surgery. We found no standardization of treatment guidelines or of patient-reported outcome measures. We also found that communication and coordination among providers was hit or miss.
Turning to inpatient care of nonsurgical back pain, we found that 90 percent of those patients are admitted from our emergency department. Most of those patients were admitted for a combination of pain management and diagnostic tests. The physicians on our hospitalist service told us that it was unclear which specialty to consult and when for back pain patients given that several different specialties at our institution offer the same services. In some instances, care would be delayed significantly when the wrong specialty was consulted first. In addition, our hospitalist team has fairly high turnover of physician staff because many physicians work as hospitalists for a couple of years before starting a fellowship. This turnover meant that establishing standardized workflow among the hospitalists was a challenge. This challenge was compounded by the fact that the hospitalist service is not regionalized, so their patients are spread out across many different units of the hospital.
Several themes arose as we turned from gathering data and identifying problems to identifying possible improvements. There was clearly a great opportunity for coordination across the continuum of specialties and care. There was also a clear need to reduce unhelpful variability in treatment for a given condition. Another improvement area was decreasing wasteful resource utilization, including unnecessary imaging and unnecessary injections. There was an opportunity to reduce the confusion caused by the significant overlap of services. We could also address our initial assignment and reduce the excessive inpatient length of stay. Finally, we could improve patient education and patient participation in medical decision-making.
Our first step as a group was to co-author standard guidelines for treatment of common painful conditions of the spine. In writing those guidelines, we emphasized consensus and buy-in from all of the different specialties treating back pain at our institution. In addition, we created versions of the guidelines appropriate for primary care providers and for specialists. We made them available in printed form, online and as an app for smart phones. Our next step was to raise funds from the relevant specialties and the hospital to hire a nurse practitioner to staff our “spine line” – a service that patients and referring providers can call for guidance in navigating to the right provider at the right time. We hope that service will decrease confusion and frustration among patients and referring providers and increase the appropriateness of specialist consultations. In addition, we intend to use it as a means to achieve increased compliance with our standard guidelines and to drive down overutilization of imaging and injections. We have tasked that nurse practitioner with helping patients get urgent access if indicated and, in doing so, to decrease emergency department visits for back pain. In addition, for services that numerous specialties provide, she will direct referrals to the relevant specialists according to a predefined, fair rotation. For patients, we have created shared decision-making tools on topics such as when imaging is indicated for back pain, treatment options for common painful spinal conditions, etc. Finally, we have selected common patient-reported outcome measures for use in all of the locations that treat back pain patients in our institution so we can measure and compare outcomes across locations and providers.
For our inpatients with nonsurgical back pain, we have begun preferentially sending them to an observation unit rather than admitting them. In that unit, we set up expedited pain medicine consultations for nonsurgical back pain patients. In addition, we arranged urgent access to outpatient physical therapy for these patients. Indeed, the observation units represent ideal locations for ruling out clinical red flags, managing acute pain, and safely returning patients to home or a post-acute setting. Currently, the hospital projects that we will save roughly 100 bed days on inpatient units and $200,000 in annual costs due to the increased use of observation units.
The introduction of accountable care represents a fundamental shift in the business model of a modern hospital. We feel that it is critical to play a leadership role in preparing our institution for that change and in shaping the solutions that are put in place. Clearly, we still have much to learn about the best ways to care for patients under accountable care, but we have already seen that there are great opportunities to improve care coordination and communication while simultaneously reducing unhelpful variability and waste.