CMS is doubling down on the bundled payment reimbursement model as the key driver of the transformation from a fee-based to a value- based health care system. On July 25, the agency announced plans to expand the model to include cardiovascular disease (myocardial infarction and coronary bypass grafting) to hospitals in 98 metropolitan areas. Participation is mandatory and begins July 2017. The expanded program builds on the Bundled Payments for Care Improvement Initiative (BPCI) launched in 2013 and the Comprehensive Care for Joint Replacement (CJR) program- already mandatory in 67 regions including 789 hospitals for total hip and knee replacement surgery since April 2016. CJR will be expanded to include surgical treatment of hip fracture. Significantly, the proposal also notes that both the extended CJR program and the cardiovascular bundles can qualify as Advanced Alternative Payment Models (AAPMs) under the Medicare Access and Chip Reauthorization Act (MACRA).
In contrast to physicians enrolled in the more complex Merit-Based Incentive Payment System (MIPS) of MACRA, AAPM participants will receive an annual lump sum bonus of 5% without exposure to risk of MIPS penalties. It's likely we'll see the bundled reimbursement model extended to other surgical as well as medical conditions as CMS drives toward it's goal of 50% of all medicare payment through alternative models by 2018.
What is a bundle?
The current concept of bundled payments really represents an expansion of the diagnosis-related group (DRG) prospective payment program launched in 1983. In effect, the DRG program established a bundled payment to the hospital for all inpatient services provided for specific diagnoses, exclusive of physician services. By establishing a single fixed payment, DRG-based reimbursement shifted financial risk from CMS to the hospital. In this model, physician services are independently reimbursed as professional fees. Importantly, DRGs only address inpatient care.
The new CMS programs including BPCI and CJR expanded the DRG approach by:
- including professional (physician) fees in the target price
- covering the full episode of care from admission up to 90 days after discharge
Total hip replacement surgery provides a great example to illustrate the differences between the blended DRG/ professional fee-for-service (FFS) payment and full value-based bundling reimbursement. Under FFS, medicare pays the hospital a set fee based on the DRG category covering nursing care, hospital room and operating room time. A separate payment is made to the surgeon for "professional services" which typically include office visits up to 90 days postoperatively. In addition, payments are made for professional fees of the anesthesiologist, physical therapist, and any other professionals involved in the care of the patient. Likewise, any post acute care services provided in a rehabilitation facility or at home are billed as distinct charges. Readmission for complications postoperatively may also generate additional fees. Conversely, in a fully bundled model, a single pre-determined payment covers all providers, services and treatments for the entire episode of care - up to 90 days after surgery. Under the proposed CMS rules, the target bundle payment is based on regional and hospital specific historical billing data reduced by a 1.5-3% discount. Providers demonstrating higher quality outcomes are discounted less. In this model, financial risk shifts to the providers rather than the payor (Medicare). If a hospital and it's providers deliver a quality outcome at a cost less than the bundle target, they keep the difference. Conversely, organizations whose costs (charges) exceed the bundle payment bear the loss. In the case of a hip replacement, for example, a hospital with a longer length of stay for the surgery because of a postoperative DVT (blood clot) and a higher post discharge requirement for inpatient rehabilitation will likely be a loser. Theoretically, the bundled payment model rewards health care organizations and providers for coordinated, cost efficient, high quality care.
Why Bundle Payments?
Porter and Kaplan put it very succinctly in their recent paper: "Bundled payments will empower and motivate providers - responsible for the overall treatment of a patient's condition - to coordinate and integrate all the specialists and facilities involved in the care." Contrasting a bundled payment strategy with fee for service or capitation, these authors identified several advantages to bundling including:
- Fosters integrated and innovative multidisciplinary care
- Built-in accountability
- Actual cost measurement and real cost reduction
From the standpoint of CMS, the bundled payment model shifts financial risk responsibility entirely to the providers - hospitals, physicians, and all providers involved in the episode of care. In addition, the program builds in a cost reduction component by discounting the target price by 1-3% from the historical average payment.
Beyond cost reduction to CMS, the bundle incentivizes health care organizations and providers to work together to provide cost-effective care with measurable quality outcomes. In CJR, the quality score incorporates rate of complications, as well as HCAPS (Health care provider and systems) survey scores and patient reported outcome data. To survive and even thrive in this new environment, health care providers and organizations must evaluate and even retool the full continuum of care while measuring cost and outcomes. Managing the patient through the post acute care space after discharge will be huge.
Risks of bundling - for patients and providers
At first glance, the bundling concept sounds sensible - improve care while reducing costs - but what are the risks?
- Reduced access for care or surgery for the sickest patients
Since CMS doesn't factor in the increased costs of caring for the sickest patients, providers and hospitals may "cherry pick" the best surgical candidates and refer more complex patients to larger more distant centers. A recent University of Michigan analysis found that failure to incorporate risk adjustment in the CJR program penalizes hospitals that treat medically complex patients. The study predicted a $827 reduction in reconciliation payment per episode for each standard deviation increase in a hospital's complexity. By the fourth year of the program, the CMS reimbursement target will be based on a regional rather than institution specific data. Hospitals with a more costly complex patient mix will be at a disadvantage. So-called "outliers" - patients whose cost of hospitalization greatly exceeds the usual average - more severely impact hospitals with smaller case volumes. These hospitals could be forced to eliminate services entirely because of intolerable financial losses.
- Bundling for chronic medical conditions is more complex and unproven
Medical problems such as diabetes, heart disease, hypertension, and even chronic obstructive lung disease often co-exist, particularly in the sickest patients. Unlike surgical care which can be defined in episodes of 30, 60, or 90 days, defining the bundle for complex medical patients is more difficult. Patients discharged after an admission for congestive heart failure require ongoing care indefinitely. How the bundle is defined may determine whether needed services are deferred to stay outside the bundle in the interest of improved reimbursement. Patients with multiple comorbid conditions are less likely to follow a straightforward path of care with predictable costs. Not surprisingly, of the 48 DRGs eligible for the BPCI program, more than 70% of awardees selected major joint replacement of the lower extremity while only 25% chose to participate in COPD. At this point, we have little published data on the small experience with non-surgical bundles.
- Smaller community hospitals are at risk
As noted above, bundling reimbursement favors hospitals with higher numbers of patients with a particular DRG bundle as volume tends to mitigate the impact of a high cost outlier on average spending for an episode of care. Smaller hospitals might face significant repayment penalties based on a single outlier. Fortunately CMS is suggesting lower limits on repayment for smaller hospitals. Since the advent of DRGs more than 30 years ago, most hospitals have optimized inpatient care and achieved significant cost reductions. As the bundle model extends the episode to up to 90 days after discharge, the focus on cost saving shifts to the post acute care (PAC) time frame. Since post acute care (PAC) currently accounts for up to 40% of the cost of care in joint replacement and CHF bundles, hospitals will need to partner with quality skilled nursing facilities to control costs and achieve good outcomes across the entire episode of care. A recent study compared Medicare costs for major surgery in high "quality" hospitals with low "quality" hospitals defined by patient satisfaction and surgical mortality. Costs were higher in lower quality hospitals primarily because of greater use of post acute care, accounting for 59.5% of the difference. Smaller, rural institutions may have fewer options and less control of PAC. Information systems to identify bundle candidates, track care across inpatient and post acute settings, and measure outcomes are essential to success in a value-based bundle payment world. Smaller hospitals may be less capable of making the significant investment in health information technology required.
Does bundling really reduce health care costs and improve quality?
The July proposed expansion of CJR and the bundled payment model signals that CMS is convinced it represents the cornerstone of the shift to value based reimbursement. However, a review of the early experience with the BPCI pilots suggests we might curb our enthusiasm until more data rolls in. The report is limited only to the first quarter of the project (fourth quarter, 2013) and compares BPCI participating hospitals with non participants after a baseline period. The participating hospitals differed signifcantly from the "controls." They were larger, more urban and had fewer medicaid admissions. While 48 DRGs were eligible, most (78%) of institutions selected orthopedic procedures (lower extremity total joint replacement and lumbar surgery). For joint replacement, the key findings were:
- BPCI participaants saw a statistically significant but small decline in the length of stay - 4.4 to 4.3 days (vs 4.5 in controls).
- BPCI hospitals decreased discharges to SNFs by 12% with an increase utilization of home health (HH). No change was seen in the comparison group.
- Baseline medicare charges were higher in BPCI patients ($22643 vs $22008) but declined more ($21484 vs $21586), though both groups showed a decline.
- Mortality was the same for both groups and didn't change from baseline.
- Unplanned readmissions were higher in BPCI at baseline and declined for both groups. Emergency room visits rose for BPCI hospitals within 30 days postop.
In sum, for lower extremity total joint replacement, larger hospitals participating in the BPCI program demonstrated small but significant cost reduction, probably primarily as a result of decreased utilization of expensive inpatient post acute rehabilitation. It's unclear whether this resulted in increased emergency room use.
Of note, the results for the lumbar surgery DRG were significantly worse. For BPCI hospitals:
- Medicare charges increased from baseline during the first quarter of the pilot - $37,019 to $41,289.
- Readmission and repeat surgery rates increased during the period
At the time of the report, data on specific medical DRG bundles (COPD, congestive heart failure, etc) were not released.
While more coordinated care along evidence based care pathways seems likely to improve patient outcomes and possibly reduce costs, much more experience and study is mandatory to validate the bundling concept as the driver of really meaningful health care transformation.
Steps to success in the bundled value-based world
My former employer, Cleveland Clinic, recently reported its successful experience as a participant in the BCPI Model 2 for total joint replacement. The Clinic's so-called "Complete Care Program" rests on a foundation of three pillars:
- Care path utilization - following physician-built evidence-based guidelines to ensure consistent, quality care.
- Care coordination - a single point of contact for the patient throughout the episode of care
- Connected care - communication between caregivers and the patient across acute and post acute care settings
Patients in the Complete Care Program experienced a decreased length of stay of about 0.5 days with the 30 day readmission rate nearly reduced by 50%. The numbers of patients discharged directly to home nearly doubled with a concordant decrease in SNF admissions. Compared to historical prices, cost savings of 9.8% ($522,389) for 271 episodes of care were achieved. Physician engagement in the development and implementation of the care path was critical.
Critical Role for Robust Health Information Technology
From multiple perspectives, the crucial role of 21st century health IT cannot be understated:
- Capability to track patients across the full continuum of care. Early recognition of complications or failure to progress as anticipated is crucial to managing the costly post acute time, reducing readmissions and ER visits. Software must facilitate communication between multiple caregivers as well as the patient.
- Early pre-admission risk assessment, even for elective procedures, anticipates patients at higher risk of complications which prolong the length of stay.
- Discharge planning risk assessment to assist in identifying the optimal post acute level of care required for the best outcome
- EMR technology with clinician friendly interfaces and point-of-care clinical decision support along the entire redesigned workflow improves documentation, provider utilization and "buy-in" while ensuring smooth "handoffs" from inpatient to outpatient providers.
- Ability to aggregate data and evaluate the redesigned clinical workflow. Bundling of all services throughout an episode of care must involve a process of continuous quality improvement, evaluating new strategies and systems of care and suggesting areas for process improvement
The recent proposal by CMS moves the bundled payment initiative from infancy to the toddler stage. We have only limited and preliminary experience with the easiest DRGs to bundle - elective joint replacement. Chronic medical conditions will be much more difficult with considerably more clinical variability, beginning with the definition of the episode. Before the toddler reaches adolescence, much work needs to be done.