Article Tools
Care Delivery and Practice Management
Article Tools
The Oncology Care Model: A Critique
Disclosures of potential conflicts of interest provided by the authors are available with the online article at asco.org/edbook.
Abstract
Rapidly increasing national health care expenditures are a major area of concern as threats to the integrity of the health care system. Significant increases in the cost of care for patients with cancer are driven by numerous factors, most importantly the cost of hospital care and escalating pharmaceutical costs. The current fee-for-service system (FFS) has been identified as a potential driver of the increasing cost of care, and multiple stakeholders are interested in replacing FFS with a system that improves the quality of care while at the same time reducing cost. Several models have been piloted, including a Center for Medicare & Medicaid Innovation (CMMI)–sponsored medical home model (COME HOME) for patients with solid tumors that was able to generate savings by integrating a phone triage system, pathways, and seamless patient care 7 days a week to reduce overall cost of care, mostly by decreasing patient admissions to hospitals and referrals to emergency departments. CMMI is now launching a new pilot model, the Oncology Care Model (OCM), which differs from COME HOME in several important ways. It does not abolish FFS but provides an additional payment in 6-month increments for each patient on active cancer treatment. It also allows practices to participate in savings if they can decrease the overall cost of care, to include all chemotherapy and supportive care drugs, and fulfill certain quality metrics. A critical discussion of the proposed model, which is scheduled to start in 2016, will be provided at the 2016 American Society of Clinical Oncology (ASCO) Annual Meeting with practicing oncologists and a Centers for Medicare & Medicaid Services (CMS) representative.
KEY POINTS
The rapidly increasing cost of health care threatens to destabilize the health care system.
New systems are needed, and CMS has proposed a new pilot model, the Oncology Care Model, that provides payments to practices in addition to the traditional FFS if practices can decrease overall cost of care and fulfill quality criteria.
This article and its corresponding Education Session at the 2016 ASCO Annual Meeting are aimed at providing a critical overview of the model by practicing physicians and a CMS representative.
National health care expenditures (HCEs) are a substantial part of the national spending and currently account for 17.1% of the U.S. gross domestic product, significantly more than other developed countries, which typically spend between 8% and 12% of their gross domestic product on health care. Despite high spending, several important health outcomes measures—for example, infant mortality—do not reflect the current level of health care spending, supporting the notion that there is a disconnect between money spent and the quality of care produced by the system. When national health care systems are compared according to country, the United States currently ranks 54th when factors such as efficiency, life expectancy, and health care cost are considered.1
A major concern is the significant growth rate of HCEs that outpace other national financial benchmarks and continue to rise at annual levels of greater than 5%. The “inflation” of health care–related expenditures is mainly driven by two major factors: hospital-related expenditures and drug prices. U.S. expenditures on pharmaceuticals are more than twice that of any other member country of the Organisation for Economic Cooperation and Development.2
Almost no other area of medicine is affected more by the steep increase in the cost of pharmaceuticals than cancer care where drug prices can easily top $10,000 per month per patient.3
There is broad consensus that this trend is not sustainable at the current rate. Private and public stakeholders are discussing mechanisms to curb spending while at the same time maintaining high-quality care. Because physician behavior can be influenced by financial incentives,4 it has been postulated that tying payment directly to the quantity of performed services, the backbone of the FFS system, incentivizes physicians to perform more services to improve financial gain. The flipside of the argument is the Health Maintenance Organization experience of the early 1990s where FFS was replaced by a fixed monthly payment for each patient but not tied to patient encounters or specific services. This led to inadequate access to care with poor patient care and satisfaction.
So how is a rational system of providing the right health care to the right patient at the right time to be designed?
If FFS is abandoned for an improved way to deliver and pay for health care, the following elements have to be met:
A payment system that adequately covers the true cost of care—the current Medicare fee schedule is underfunding the cost of care by 20% to 30%, forcing providers to use other resources to cover the cost of care for Medicare beneficiaries such as increasing their fees for private payers;
A system that allows providers and their staff to care for patients in a variety of ways including phone calls, emails, and other forms of contact in addition to traditional face-to-face encounters;
A system that allows easy and unrestricted access for patients;
A system that ties successful adherence to nationally accepted quality standards and patient satisfaction directly to physician reimbursement;
A system that provides a mechanism for controlling the cost for drug and hospital spending.
Federal legislation now mandates CMS to fund alternate payment models to deliver health care, and CMS has established the Center for Medicare & Medicaid Innovation (CMMI) to explore options. From 2013 to 2015, a $20 million grant was provided to 12 medical oncology practices for the COME HOME project, which used three novel mechanisms to control the cost of care for patients with common cancers:
The systematic use of a phone triage system to determine the acuity of a patient’s complaint and offer them one of three options: advise over phone for minor complaints, a same-day appointment in the medical oncology office for more urgent issues, or a direct referral to the nearest emergency department for the sickest patients;
Extended office hours on evenings and weekends so that patients requiring an assessment or office services such as antiemetics, hydration, antibiotics, or other injections would not have to visit a hospital or emergency department;
The use of treatment pathways that were created by the participating practices to create predictable, cost-effective, but high-quality care.
Although savings varied geographically, as a whole participating COME HOME practices were able to show substantial savings with lower costs largely attributable to avoiding costly hospital admissions and emergency department referrals for patients who could be adequately cared for in the office.5 This example of increasing the value of care by reducing cost while at the same time improving the quality of care and leading to higher patient satisfaction is an encouraging example of positive health care reform.
Unfortunately, CMMI decided to discontinue further funding for COME HOME and has presented a new model aimed at reducing cost of care for medical oncology patients while at the same time attempting to improve the quality of care, the Oncology Care Model (OCM). This model uses two types of additional payments in addition to the usual FFS payments and other payments, for example, for chemotherapy administration: a performance-based incentive as well as bundled monthly payments for patients undergoing chemotherapy. In exchange, oncology practices are obligated to establish systems aimed at providing a higher level of care with additional reporting requirements to CMMI. As detailed further below, practices will receive an additional payment of $160 per beneficiary per month (“the bundle”) for any 6-month period during which a patient is undergoing active chemotherapy treatment in a medical oncology practice. In return, practices agree to strict reporting and quality standards:
Provide patient navigation;
Document a care plan that contains the 13 components as outlined by the Institute of Medicine;
Provide 24-hour-a-day, 7-day-a-week patient access to an appropriate clinician who has real-time access to the practice’s medical records;
Treat patients with therapies consistent with nationally recognized clinical guidelines;
Use data to drive continuous quality improvement;
Use an ONC-certified electronic health record and attest to stage 2 of meaningful use by the end of the third model performance year.
In addition, private payers were invited to apply, extending the potential reach of the project to patients not covered by CMS.
The OCM is often characterized as an episode-of-care or bundled reimbursement model but is in reality a FFS and shared savings model. The shared savings are predicated on a calculated bundle, but the reimbursement application would more honestly be called an Oncology Accountable Care Organization.
There are theoretical advantages to a true bundle for both payers and providers. The payer’s costs are predictable for an individual patient for a set period time and the economic incentive toward high-cost, high-quantity care, inherent in FFS medicine, is replaced by physician accountability for costs and quality they can control. The provider has freedom from payer micromanagement and, ideally, adequate and predictable payments. Then, released from the constraints of practicing to billing codes, practices have the flexibility necessary to seek high-quality care with fewer resources through innovative care redesign. But the key to these advantages is that the bundle has replaced FFS reimbursement.6 Unfortunately, OCM fails to do that.
The Accountable Care Organization model has yet to deliver on its promise of high-quality, low-cost care,7 perhaps because simply adding incentives to decrease costs on top of the existing payment system that incents increased costs does not solve the problem. At the best, it requires a very complicated balancing act of incentives and quality accountability. At its worst, it is the hope that two wrongs make a right. Skeptics might suggest that participants, constrained by continued dependence on FFS reimbursement from capitalizing on a redesign of the way they deliver services, will be best served financially by collecting their FFS billings and management fee, go through the motions of CMS dictates, but avoid real innovative efforts and hope that health care inflation and a bit of luck will result in some additional booty.
However, it is the bundle construct that makes OCM both novel and controversial. Additionally, because one can easily envision that the CMMI pilot model is the precursor of a real bundled reimbursement model, it requires careful scrutiny. The OCM takes a simple approach—all spending on all care provided to the patient with a particular cancer is in a bundle that starts with the initiation of chemotherapy and lasts for 6 months. This construct has raised a number of concerns.
The OCM bundle reimbursement is constructed based on an individual practices’ historical costs when available and on regional or national data adjusted for geographic variation when historical costs are not available. The focus on competing against your own data guarantees that over time opportunities for new cost sharing diminishes. Although CMMI may not choose to ratchet down the reimbursement for a bundle during the pilot model, the history of Medicare’s oldest bundle, dialysis reimbursement, strongly suggests that is exactly what will happen.
The choice of 6 months is arbitrary. In other specialties, CMS and commercial payers have defined an “episode of care” to be a full course of treatment of a specified treatment. In this scenario, adjuvant concurrent fluorouracil (5-FU) and radiation for rectal cancer is a shorter episode and FOLFOX (folinic acid, 5-FU, and oxaliplatin) for colon cancer is a longer episode. In the OCM model, the first patient is paid a 6-month bundle for 5 weeks of treatment and the second patient is paid the same bundle for 24 weeks of therapy. Unless, of course, the second patient has some delays in treatment, then the practice may be paid two 6-month bundles for 27 weeks of therapy. A study commissioned by CMS and performed by the RAND Corporation demonstrated that gaps of this sort in courses of the same therapy are common. One can envision that the longer the bundle the greater the variation of practice costs within a bundle and greater the opportunity for selection bias, either intentionally or inadvertently. Critiques have suggested a sequence of shorter repeating bundles. ASCO has put forward a monthly bundle in its own payment reform proposal.8
The all-inclusive nature of the bundle has also been controversial. Conceptually, the aim of a bundle is to transfer risks associated with high levels of spending from payers to the providers doing the spending. Presumably, the provider then decreases the spending. However, within an all-inclusive bundle, there is spending, or risks, that the provider can control and risks that he cannot. Economists refer to these two categories of financial risk as technical risk and probability risk.
Technical risk is a function of the practice’s clinical skill and efficiency in managing the diagnosis and treatment or other aspects of care that are under the physician's control and that he is able to impact either by controlling utilization or cost, preferably both. Probability risks are random or unpredictable events that we classically buy insurance to mitigate. In theory, a bundle should only include technical risks and the probability risks should remain with the payer. Practically, this can be very difficult.
When ASCO’s payment reform work group attempted to split patient care into oncology related versus other medical care, it quickly became apparent that adjudicating which CPT and ICD-10 codes should be attributed to the oncologist was simply impractical. Furthermore, several pilot projects have now demonstrated that medical oncology homes can dramatically lower emergency department and hospital costs,9-11 broadly indicating that there is substantial technical risk within patient care. CMS has indicated that the probability risk that does exist—a patient with higher than average comorbidities or the patient who is in an automobile accident on the way to their chemotherapy appointment, for example—can be mitigated through risk adjustments based on patient and service characteristics in the first example and through reinsurance or stop-loss provisions in the latter case. Of course, the most important mitigation of risk in the pilot model is the provision of one-sided risk, clearly a carrot that will ultimately be replaced with a stick.
The RAND Corporation estimated that 25% to 40% of the OCM bundle is chemotherapy and supportive care drugs.12 Among key opinion leaders and policymakers, the inclusion of drugs in the oncology bundle seems a foregone conclusion. It is assumed that (1) physicians, not patients, control demand; (2) practices generate substantial revenue from drugs via the buy-and-bill system; (3) if oncologists are at financial risk for the drugs, they will choose the least costly regimens when efficacies are similar; and (4) because oncologists will become more price sensitive, they will seek out better prices through negotiations with manufacturers and group purchasing organizations. Although theoretically compelling, there is reason to question the practical validity of these claims.
If drugs are to be appropriately included in a bundle, analysis should conclude that drug spending is a technical risk. Empirical evidence of this is limited. One often-cited example of drugs being a technical risk is the Zaltrap story.13 Zaltrap was launched in 2012 at twice the price of Avastin, a drug with similar outcomes. Physicians at Memorial Sloan Kettering Cancer Center announced in a New York Times Op-Ed piece that they would not prescribe Zaltrap, and its price was reduced by 50%. Although this may have emboldened oncologists who have increasingly decried high-priced drugs, the average monthly launch prices of cancer drugs since then suggest that the pharmaceutical industry has been unfazed.
Another expectation if drugs are a technical risk is that practices should have leverage over the acquisition prices of pharmaceuticals. Although there are examples of volume discounts for large purchasers and group purchasing organizations, this applies only to generic medications and the infrequent circumstance in which there are multiple infusable competitors with similar efficacy within a therapeutic class. Oncology practices have no control over the price of novel agents, and oral drug prices are negotiated by payers. For these reasons, the prices of new cancer drugs, those that drive costs the most, should be considered a probability risk.
Proponents of drugs in the bundle will argue that the inability to control price is outweighed by the oncologist’s ability to impact utilization. They argue that the 20% to 30% of annual spending on off-label drugs represents tremendous waste in the system, driven by FFS medicine and buy-and-bill drug reimbursement. The evidence for this is sparse. Brooks et al14 reported in an analysis of FFS Medicare data that only 10% of regional variation in spending was a result of drug choice, and that drug costs were only 16% of the total spend. This was dwarfed by the 67% of regional spending variation related to acute hospital care, accounting for 48% of the total spend.
A real-life example of the principles of probability and technical risks as they pertain to oncology may be found in the recently reported United Health Care (UHC) demonstration project. Margins on drugs were replaced by a management fee to remove profit via drug choice, and practices were incentivized to decrease hospital and emergency department services through shared savings and a bundling of inpatient evaluation and management services. Drug costs increased, but the demonstration was determined a success because of impressive decreases in acute-care costs.9 One interpretation of these data is that the probability risks (drug costs) could not be controlled, but the technical risks associated with hospital and emergency department services were.
The risks of a bundle are dependent on its variability. There are three ways to manage the variable risk. One way is to define bundles with enough granularity to minimize the variability between patients and their associated costs. Today, that means developing bundles based on the genetic profiles of cancers. In a fairly short time, non–small cell lung cancer has evolved from one monolithic cancer and four equally efficacious treatment regimens to at least four different cancers with very different therapies and associated personalized costs. Unfortunately, it would be naive to assume that any payment system will be able to keep up with the personalized medicine revolution. It is little wonder that UHC and The University of Texas MD Anderson Cancer Center have chosen head and neck cancer as their first effort at a bundled payment contract. It is a disease for which drug therapy and treatment options are limited. Even then, they have eight distinct treatment/payment bundles.15 Due to the impracticality of managing granular bundles, it is anticipated that the OCM bundles will be fairly generic buckets.
In that circumstance, the other way to manage the risk of the bundle will be for practices to get big. Large volumes of patients can then mitigate the risks associated with any one individual. This would be expected to further the consolidation of oncology into larger groups through mergers, acquisitions, and affiliations both in independent and hospital-based practices. This may prove acceptable to health policy makers but could have important unintended consequences; there is very limited evidence that consolidation improves either quality or access to care.
The third way to mitigate the risk of drugs in the bundle is to move beyond efficiency in drug choice and utilization to the realm of stenting on care. OCM promises to monitor quality of care and to hold practices accountable for it, but the task is not a simple one. It is hypocritical to believe that buy-and-bill drives overutilization and to ignore the incentive to underuse inherent in drugs in a bundle.
An alternative to drugs in the bundle, one that may still assure payers that drugs are being used appropriately, are value-based treatment pathways.16 Evidence-based with the purpose of improving value, quality, and safety, pathways have been shown to reduce variance in drug regimens, doses, schedules, and treatment duration. Several studies have shown savings with the utilization of what must be considered rudimentary pathways and less than stellar pathway compliance.17,18 One project reduced the 12-month average costs for chemotherapy and supportive care drugs for patients with lung cancer by 37% or $9,747.18
Another advantage to pathways over bundles is that, as opposed to bundles developed by payers, pathways can be developed by oncologists through transparent, peer-reviewed, committee-selected regimens that incorporate efficacy, toxicity, and cost considerations. ASCO recently published a policy statement providing direction and recommendations regarding the development and utilization of pathways in clinical practice.19 As pathway development matures, they will incorporate more sophisticated value models that will account not only for drug costs but also the total costs of a treatment regimen to include toxicity, ancillary support services, acute care, and patient productivity loss.
Criticisms of OCM’s reliance on competition against oneself, an arbitrary 6-month episode, and fuzzy details on establishing limits on financial risk are details that can most certainly be overcome if CMMI is introspective in its evaluation and willing to make incremental improvements.
However, the model has two Achilles’ heels that it must consider if OCM is to become a win-win-win for patients, providers, and payers. First, it must divorce itself from the reliance on FFS as its backbone of reimbursement. Only then can it align incentives with innovative solutions that can produce both high-quality and efficient cancer care.
Then it must consider that drugs in a bundle may mitigate cost and variability for payers, but this is only accomplished by putting patients and practices at unacceptable risks. Bundled drugs are structured to require the doctor to run a gauntlet between rational care and rationed care through the use of hidden incentives. Pathways, alternatively, use transparency, consensus, and evidence-based practice benchmarks to assist the doctor.
OCM II should not wait 7 years for completion of the current OCM pilot model. It should be a hybrid model that brings together a true bundle without FFS to reimburse the care that an attentive and skilled physician can control and a robust pathways program that will negate a perceived need to change the way we pay for drugs.
ASCO is to be commended to convene an Education Session at its 2016 Annual Meeting during which CMMI will be able to present its model and have an open discussion with oncology practice representatives from various backgrounds about the model’s feasibility. It is clear that any practices participating in the CMS model will have to institute significant structural changes to fulfill these criteria and only the future will tell if the investment of time and resources will benefit patients while at the same time bending the cost curve for oncology. In reality, the model is not designed to achieve at least the second goal, reducing the cost of care. This is because the main drivers of expensive care are not addressed in the model (i.e., the cost of hospital care and the ever-increasing cost of cancer drugs). In addition, other expensive cost centers for patients with cancer such as biopsies, surgical procedures, or radiation therapy are also not part of the model, making it impossible for medical oncology practices to affect the overall cost of care.
Unless CMS addresses all price centers contributing to the escalating cost of cancer care, its reforms will fail to make good on their promise.
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc.
Honoraria: AstraZeneca, Bristol-Myers Squibb, Celgene, Genentech, Janssen Oncology
Consulting or Advisory Role: AstraZeneca, Celgene, Genentech
Speakers' Bureau: Bristol-Myers Squibb, Celgene, Genentech
Travel, Accommodations, Expenses: Celgene, Genentech
Honoraria: AZOncology, AZOncology
Consulting or Advisory Role: New Century Health
| 1. | Chen S, Wong S. Singapore Beats Hong Kong in Health Efficiency: Southeast Asia. http://www.bloomberg.com/visual-data/best-and-worst/most-efficient-health-care-2014-countries. Accessed March 29, 2016. Google Scholar |
| 2. | Organisation for Economic Cooperation and Development. OECD Health Statistics 2014. How Does the United States Compare? http://www.oecd.org/unitedstates/Briefing-Note-UNITED-STATES-2014.pdf. Accessed March 29, 2016. Google Scholar |
| 3. | Light DW, Kantarjian H. Market spiral pricing of cancer drugs. Cancer. 2013;119:3900-3902. Crossref, Medline, Google Scholar |
| 4. | Lungren MP, Amrhein TJ, Paxton BE, et al. Physician self-referral: frequency of negative findings at MR imaging of the knee as a marker of appropriate utilization. Radiology. 2013;269:810-815. Medline, Google Scholar |
| 5. | Burns J. COME HOME Program Set to Save $33.5M Over 3 Years. http://www.onclive.com/publications/oncology-business-news/2014/August-2014/COME-HOME-Program-Set-to-Save-335M-Over-3-Years. Accessed March 29, 2016. Google Scholar |
| 6. | American Medical Association. Center for Healthcare Quality and Payment Reform: A Guide to Physician Focused Alternative Payment Models. http://www.chqpr.org/downloads/Physician-FocusedAlternativePaymentModels.pdf?utm_source=Distribute+Physician-Focused+APM+Report&utm_campaign=Bundling+Badly&utm_medium=email. Accessed March 29, 2016. Google Scholar |
| 7. | Burns LR, Pauly MV. Accountable care organizations may have difficulty avoiding the failures of integrated delivery networks of the 1990s. Health Aff (Millwood). 2012;31:2407-2416. Accessed March 29, 2016. Crossref, Medline, Google Scholar |
| 8. | American Society of Clinical Oncology. Patient-Centered Oncology Payment. Payment Reform to Support Higher Quality, More Affordable Cancer Care. http://www.chqpr.org/downloads/ASCO_Patient-centered_Oncology_Payment.pdf. Accessed March 29, 2016. Google Scholar |
| 9. | Newcomer LN, Gould B, Page RD, et al. Changing physician incentives for affordable, quality cancer care: results of an episode payment model. J Oncol Pract. 2014;10:322-326. Link, Google Scholar |
| 10. | Association for Value-Based Cancer Care. Value-Based Cancer Care: Oncology Medical Home Shows Significant Cost-Savings, Improved Care Delivery. http://www.valuebasedcancer.com/vbcc-issues/2015/december-2015-vol-6-no-11/26472-oncology-medical-home-shows-significant-cost-savings-improved-care-delivery. Accessed March 29, 2016. Google Scholar |
| 11. | Sprandio JD. Oncology patient-centered medical home. J Oncol Pract. 2012;8:47s-49s (suppl 3). Link, Google Scholar |
| 12. | CMS Alliance to Modernized Healthcare. Specialty Payment Model Opportunities and Assessment: Oncology Design Report. Santa Monica, CA: RAND Corporation; 2014. Google Scholar |
| 13. | Conti RM, Berndt ER. Winners and losers of the Zaltrap price discount. Health Affairs weblog. February 22, 2013. http://healthaffairs.org/blog/2013/02/20/winners-and-losers-from-the-zaltrap-price-discount-unintended-consequences/. Accessed March 29, 2016. Google Scholar |
| 14. | Brooks GA, Li L, Uno H, et al. Acute hospital care is the chief driver of regional spending variation in Medicare patients with advanced cancer. Health Aff (Millwood). 2014;33:1793-1800. Crossref, Medline, Google Scholar |
| 15. | The University of Texas MD Anderson Cancer Center. MD Anderson, UnitedHealthcare Launch New Cancer Care Payment Model. MD Anderson News Release December 16, 2014. http://www.mdanderson.org/newsroom/news-releases/2014/md-anderson-unitedhealthcare-new-cancer-payment.html. Accessed March 29, 2016. Google Scholar |
| 16. | Conti R, Ward JC, Page R, et al. A pathway through the bundle jungle. J Oncol Pract. In press. Google Scholar |
| 17. | Kreys ED, Koeller JM. Documenting the benefits and cost savings of a large multistate cancer pathway program from a payer’s perspective. J Oncol Pract. 2013;9:e241-e247. Link, Google Scholar |
| 18. | Neubauer MA, Hoverman JR, Kolodziej M, et al. Cost effectiveness of evidence-based treatment guidelines for the treatment of non-small-cell lung cancer in the community setting. J Oncol Pract. 2010;6:12-18. Link, Google Scholar |
| 19. | Zon RT, Frame JN, Neuss MN, et al. American Society of Clinical Oncology Policy Statement on Clinical Pathways in Oncology. J Oncol Pract. 2016;12:261-266. Google Scholar |
