Comments on CMMI “AHEAD” (Advancing All-Payer Health Equity Approaches and Development) Model for Healthcare Reform in Rhode Island
Prepared for Linda Ujifusa, Candidate for RI State Senate
By Stephen Kemble, MD, PNHP Board Member and Chair of PNHP Policy Committee
July 20, 2024
Key messages:
Participation in the AHEAD model of global hospital budgets is voluntary for Medicare Advantage and Commercial insurers. Substantial non-participation in the model would greatly limit savings from reductions in hospital billing and collections costs. Absent nearly universal participation in the model, the administrative savings anticipated by the AHEAD model of global hospital budgets will not be realized.
The solution to problems with coordination of care and integrating mental health and substance abuse disorder treatment is not the AHEAD model of introducing more and more complexity into primary care practices, but rather to expand community resources and assure they are readily available to primary care.
Background:
CMS and CMMI are promoting “Value-Based Payment” (VBP) models based on the assumption that excessive cost in U.S. healthcare is due to fee-for-service (FFS) payment which is presumed to reward volume of services driving over-utilization. The evidence for this assumption is scant, as the U.S. always had lower than average utilization compared to other developed countries even when everyone was paid with FFS, and there is extensive evidence that the real problem is excessive administrative cost and profiteering by corporate middle-men seeking to profit from health care. There is indeed abuse of FFS in U.S. healthcare, but it is mostly due to corporate interference and profiteering, and it has been getting worse as doctors are being driven out of independent practice and into corporate employment because of the administrative burdens and costs of recent payment and delivery system reforms. In fact, independent primary care practices are the healthcare players least likely to abuse FFS.
VBP strategies all involve shifting insurance risk onto fiscal intermediaries (insurance plans such as Medicare Advantage or Medicaid Managed Care) and/or onto the providers of care themselves. The assumption is that up-front payment and acceptance of financial risk will control cost and reward “value” instead of “volume” of services in healthcare. However, this assumption is ideological and not based on evidence, and VBP strategies have never achieved the “triple aim” goals of improving quality of care, improving population health, and reducing cost.
The dominant VBP strategy is capitation, or pre-payment of a fixed amount per-member-per-month, so that the capitated entity accepts responsibility for covering needed care for defined members for a fixed up-front payment. This requires defined members, and for Medicare beneficiaries who have not signed up for a Medicare Advantage plan, CMS “attributes” or “aligns” beneficiaries to particular providers based on where they sought care in the past, and then pays a group of doctors or an integrated doctor-hospital system a capitated payment as an “Accountable Care Organization” or ACO. However, the “attribution” or “alignment” process is fraught with errors, so this has been much more difficult than CMS envisioned in their push to apply capitation to beneficiaries who have not signed up as “members” with a Medicare Advantage plan.
Other VBP strategies include bundled payment for an episode of care, such as hip surgery, with a flat payment per episode that includes pre-op care, surgery including both hospital and physician cost, and post-op follow-up care. More limited strategies include withholding part of fee-for-service payment and using it for incentive payments based on outcomes, quality metrics, or total cost of care. All of these create financial incentives to avoid treating sicker, more complicated patients at risk of poorer outcomes.
CMS and CMMI often refer to the incentive under FFS to deliver more “volume” of care, but they ignore the perverse incentives inherent to capitation and other forms of VBP.
Gaming the Risk Pool:
Paying the average cost for a population obviously over-pays for the healthy and under-pays for the sick, creating an incentive to “cherry pick” healthier than average “members” and avoid covering or paying for sicker, higher-risk individuals, or setting up policies to interfere with care of higher risk individuals, driving them to leave the plan. This is called “gaming the risk pool,” and it has been a serious problem with capitated private Medicare Advantage plans since the Medicare Advantage program was introduced in 1997.
Up-coding:
In an effort to reduce the incentive to game risk pools, CMS introduced risk adjustment, which is intended to pay more for higher-risk members and less for healthier ones. The initial risk adjuster was based on demographics, but this was only able to predict individual cost with 1% accuracy and did nothing to deter risk pool gaming. In 2004 CMS introduced the Hierarchical Conditions Categories (HCC) risk adjuster using diagnostic groups. This improved predictive accuracy from 1% to 12%, which is still quite inadequate to deter risk pool gaming. However, the HCC risk adjuster tied payment to diagnoses for the first time and created an opportunity to increase revenue by “up-coding,” or ensuring as many diagnoses as possible are reported on claims, and sometimes fudging diagnoses to make patients appear sicker than they really are.
Skimping on care:
Up-front payment for a defined population means the capitated plan or practice keeps more of their payment if they deliver less care. Although this can deter unnecessary care, research evidence shows that it deters necessary and unnecessary care alike and can impair quality of care and population health.
Profiteering:
A capitated entity is entitled to keep what is not spent from their capitated payment, so it can skim off profit by gaming its risk pool to capture healthier than average members, up-coding to game risk adjustment, and skimping on care. A recent report from PNHP showed that private Medicare Advantage plans are on average paid 22% to 33% more than if the same beneficiaries were in Traditional Medicare, and indeed Medicare Advantage has become the most lucrative health insurance market in the U.S.
This evaluation of the AHEAD model focuses on financial incentives and administrative cost, as these, and not FFS, are the main drivers of excessive cost in U.S. healthcare. The AHEAD model is based on the assumption that FFS is the problem and “value-based payment” is the solution, and it has mixed good and bad strategies when evaluated from the perspective of financial incentives and administrative cost. The goals of the model are 1) controlling growth in the total cost of care, 2) increasing investment in primary care, and 3) improving equity and population health outcomes. These are all laudable goals, but the elements of the model need to be considered for their effects on financial incentives and administrative costs and burdens, as these determine whether the model can achieve its goals.
AHEAD’s Hospital Global Budgets:
Global budgeting of hospitals is essentially a good idea. Based on international comparisons (notably Canada and Scotland), the most cost-effective way to fund hospitals is global operating budgets from a single-payer based on the cost of operations for the range of services offered, and funding capital improvements (new construction, new lines of service) separately based on community need. Fully funding hospitals with global budgets from a single-payer eliminates FFS billing and collections almost entirely, which typically consume around 12-15% of U.S. hospital budgets. Hospital billing and collections in the U.S. are also rife with confusing complexity, gaming of payment systems, and abuse.
Maryland initiated all-payer rate setting since 1971, requiring all payers (Medicare, Medicaid, and commercial insurance plans) to pay the same FFS rates for the same services. This stabilized hospital payment and eliminated cost-shifting (charging higher commercial rates to offset low Medicare and Medicaid rates), and it spread the burden of uncompensated care across all insurers. However, Medicare FFS rates are below average nationally, so all-payer rate setting resulted in higher-than-average Medicare rates for Maryland hospitals. It also reduced gaming of FFS payment methodologies but did little to reduce the administrative cost of billing and collections.
CMS has viewed higher than average Maryland hospital costs as a problem with the financial incentive under FFS to provide too much presumably unnecessary care, and envisioned a solution based on the financial incentive to restrict (presumably unnecessary) care with “value-based payment.” The fundamental flaw in this strategy is that the problem was not financial incentives or over-utilization, but higher than average FFS rates due to the Maryland all-payer rate setting system and failure to reduce the substantial administrative costs borne by hospitals in our multi-payer system.
From 2014-2017, CMS added the “Global Budget Revenue” (GBR) system to all-payer rate setting for Maryland hospitals, but still required hospitals to submit FFS claims, reconciling revenues to the global budget at the end of the year, plus pay-for-performance incentives. GBR did slow hospital spending somewhat, but Maryland was still among the most expensive sites for Medicare hospital payment due to all-payer rate setting.
CMS assumed the “value-based” incentives were not strong enough, and from 2018-2024 Maryland hospitals were put on “Total Cost of Care” (TCOC) budgeting. The TCOC model still used FFS claims reconciled to a budget at the end of the year, but added modifications such as partial capitation of hospital budgets with “members” (remember, capitation requires “members” in order to pay per-member-per-month) “attributed” according to where they sought care, and risk of loss if hospital costs exceeded the capitation rate. TCOC also added more pay-for-performance metrics, and incentives to coordinate with non-hospital providers to improve care of chronic conditions and reduce preventable admissions. This created stronger incentives to restrict care and made the hospitals at least partially accountable financially for care occurring outside their control. And it added administrative costs and burdens, instead of reducing them.
This is bears little resemblance to the much simpler form of global hospital budgeting based on cost of operations (as opposed to capitation based on attributed “members”) used in Canada and Scotland. CMS is manipulating financial incentives instead of making hospital payment simpler and more predictable and minimizing administrative costs, which are the real keys to solving the hospital cost problem.
That being said, there is a way to implement cost-effective hospital global budgets in a multi-payer system. A state could collect funds for hospital care from all insurers operating in that state according to their share of a hospital’s payer mix and use pooled funds to establish hospital global budgets. If all payers were included, almost all hospital FFS billing and collections could be eliminated, achieving the needed savings from reduced administrative cost. The hospitals would be operating within predictable budgets, which could be modified with changed circumstances such as a pandemic, higher pharmaceutical costs, demographic changes, etc. Hospitals should be encouraged to implement front-line quality improvement projects, but financial incentives would be unnecessary since doctors and hospitals are already highly motivated to provide quality care if they have the support to do so, and if they don’t have pervasive interference from non-clinical corporate managers pursuing financial goals.
Presentations to the Rhode Island Governor’s Council on Behavioral Health summarize Hospital Global Budgets under the AHEAD model. If AHEAD really means collecting hospital funds from Medicare FFS, Medicaid, Medicare Advantage, and Commercial payers and then pooling funds to pay hospitals with fixed annual payments that could be adjusted for factors like inflation, changes in people served, and changes in services provided, then AHEAD would likely represent an effective model of hospital global budgeting as described above. However, problems would occur if the CMS-designed methodology still requires FFS claims reconciled to the budget, or involves attributed “members” and per-member-per-month payments that shift insurance risk onto hospitals, or requires hospitals to assume financial risk for care in the community that is beyond the hospital’s control. Also, administrative savings won’t be realized if participation is voluntary for Medicare Advantage and Commercial insurers, as substantial non-participation would mean hospitals could not cut back much on their billing and collections costs.
Primary Care AHEAD Model:
Once again, CMS and CMMI assume excessive cost is due to FFS and a presumed incentive to deliver unnecessary care, even though there has never been evidence for widespread over-utilization of care, especially in primary care practices. CMS assumes lack of care coordination is due to inadequate financial incentives to coordinate care, ignoring pervasive obstacles now in place that interfere with efforts by primary care practices to coordinate care for their patients. Obstacles include narrow and “ghost” networks (large numbers of providers listed as participating that are not available or not accepting the plan’s patients) in many health plans that can make referrals to needed specialists extremely challenging, and narrow formularies and pervasive prior authorization requirements disrupting care at every turn.
The real problem is not lack of financial incentives to coordinate care, but inadequate healthcare resources in many communities. There is a severe shortage of doctors in both primary care and most specialties in many parts of the country, and a hostile practice environment that is driving doctors out of independent practice in primary care and other cognitive specialties such as psychiatry. Those that remain in practice find the complexity and cost of independent practice have become overwhelming. They must deal with billing and collections from insurers that obstruct payment at every turn, and they must buy computer systems and hire staff needed to satisfy complex documentation requirements, maximize diagnostic coding, and report “quality” metrics demanded from Medicare and capitated Medicare Advantage and Medicaid Managed Care plans. So doctors sell their practices to hospitals or insurance plans or private equity firms, only to discover that their corporate owners now expect “return on investment” and require them to serve corporate financial goals at the expense of the best interests of their patients.
The solution is not herding primary care practices into “accountable care organizations” or corporate ownership and then attributing “members” to them so as to pay either the ACO or the corporate owners or the doctors with capitation. A payment system that rewards treating the healthy and avoiding the sick, up-coding diagnoses to game risk adjustment, and persuading patients to avoid needed referrals to specialists is antithetical to traditional medical ethics and leads to “moral injury” and burnout. The administrative costs and burdens required to deal with managing insurance risk and pay-for-performance metrics already leave doctors with inadequate time to give patients the attention they need for high quality care. And in practice, “managing insurance risk” means learning to avoid sicker patients and populations, up-code as much as possible, and restrict care as much as possible.
The AHEAD model for primary care includes a per-member-per-month supplement to their FFS income to cover “advanced care management and behavioral health integration activities.” This is essentially good, although problems with “attribution” to enable per-member payment remain. The perverse incentives with full capitation are largely avoided with a limited capitation payment to cover care coordination for patients with chronic conditions, especially if the patients identified as “members” of the practice have themselves designated that practice as the site of their care and asked for their care to be coordinated by that practice.
The AHEAD model encourages “integration of behavioral health.” There are three potential ways for this to be organized:
Each primary care practice could hire one or more behavioral health specialists such as psychiatrists, psychologists, or social workers within the practice. But this means behavioral health care will be subject to the strengths and weakness of the in-house practitioner(s), who may not have the skills or scope of practice authority to manage a large percentage of the practice’s behavioral health needs.
Primary care practices could be encouraged to screen for mental health problems and then refer those who need specialty care to practitioners in the community. Mental health screening is a required metric for the AHEAD model. However, there are severe shortages of mental health practitioners in much of the country and those that are available often refuse to accept health insurance, especially if the patient has a Medicaid managed care plan. In this context, a primary care practice can identify problems but cannot arrange for them to be treated appropriately.
The most effective strategy is the Collaborative Care Model, which pairs a part-time psychiatrist with a care manager, usually a social worker, who field referrals from primary care. The care manager gathers a full psychosocial history and chart review and presents the case to the psychiatrist, who also has access to the patient’ s chart, and who then gives specific advice to the primary care practitioner on how to manage the patient within the primary care practice. The Collaborative Care team provides follow-up via telephone or video telehealth as long as needed. The Collaborative Care Model has been quite effective in practice, demonstrating results as good or better than referral to community psychiatrists or care by an in-house mental health practitioner. It also greatly leverages the psychiatrist’s time to provide quality care to a much larger population than they could manage seeing all patients directly. This is the most realistic and effective approach to meet behavioral health needs when there is a shortage of psychiatrists.
CMS has developed FFS procedure codes for psychiatric Collaborative Care, to be used by the primary care practice which then must charge the patient a co-pay and reimburse the Collaborative Care team. Most Collaborative Care programs have difficulty funding more than about 80% of their costs using these FFS procedure codes. In Hawaii, unique in the country, the Collaborative Care Model is funded with pooled funds from most insurance plans and paid with a global budget (as recommended above for hospitals), making services available independent of the patient’s insurance status and with no charge to the patient or primary care practice and no need for the primary care practice to hire in-house mental health professionals. Pooled direct funding of the Collaborative Care Model by all payers is about 20-25% more cost effective than using the CMS billing codes because it eliminates having to deal with FFS billing and collections for everyone involved.
The AHEAD model for primary care seeks to align with states’ Medicaid primary care programs and quality priorities. However, the major obstacle to “alignment” with Medicaid is the very low rate of physician participation with Medicaid managed care plans, especially for primary care specialties and psychiatry. Problems with narrow and “ghost” networks are pervasive in Medicaid managed care.
The one notable exception is Connecticut Medicaid, which terminated their Medicaid managed care plan contracts in 2012, took insurance risk back under the state Medicaid department, increased primary care pay to Medicare FFS rates, and added community-based resources for high-risk and special needs patients. They hired one of the former Medicaid managed care plans, a local non-profit plan, as their administrative services only contractor for a small fraction of the administrative cost of full-risk contracting with multiple plans. The result was an immediate drop in Medicaid per-beneficiary cost and around 30% increase in participation by primary care practices over the next few years. Specialists also resumed accepting Medicaid, even without their own pay raise, to support their primary care colleagues now that the managed care plans were out of the picture. The result was improved access to out-patient care, reduced ER visits and hospital admissions and re-admissions, and much lower administrative cost. Six years out Connecticut was paying 14% less per beneficiary than they were in 2012 under the managed care plans. Thereafter per-member Medicaid costs resumed rising, but at a much slower rate than before, and Medicaid as a percentage of the Connecticut state budget has remained stable.
The solution to problems with coordination of care and integrating mental health and substance abuse disorder treatment is not to introduce more and more complexity into primary care practices, but rather to expand community resources and assure they are readily available to primary care:
Improve the practice environment for both primary care and specialty practices with better pay and lower administrative demands, including for psychiatry and other mental health disciplines, to make practice in Rhode Island more attractive than elsewhere in the country and grow the work force (as Connecticut’s Husky Medicaid program has done).
Implement globally budgeted Collaborative Care teams consulting to primary care practices for behavioral health problems and for other specialty consultations (as Hawaii is already doing).
Establish globally budgeted community-based programs for special needs patients and populations and make them readily available to all primary care practices without barriers imposed by multiple plans with narrow provider networks and poor provider participation. This should include programs for the seriously mentally ill and substance abusers, interdisciplinary teams that can work with high-risk patients with various chronic diseases in the home to optimize care and keep them out of the hospital, and readily available community resources to address social needs such as housing, food insecurity, and transportation. These services must be readily accessible to all primary care practices but need not be established within the primary care practices. (Connecticut’s Husky Medicaid program is already doing this with community-based special needs programs.)
References
References and supporting documentation from health policy research and PNHP reports are available on request. Feel free to share this summary and any other reports supplied by PNHP. For more information, please contact Stephen Kemble, MD, as below.
High U.S. Healthcare Cost: Fee-for-Service vs high prices due to high administrative cost:
Papanicolas, Woskie, Jha. Health Care Spending in the United States and Other High-Income Countries. JAMA 03-13-18
Corporatization and Profiteering in Medicare Advantage:
PNHP Report: Our Patients, Their Profits: Quantifying Overpayments in the Medicare Advantage Program. October, 2023.
PNHP Report: Taking Advantage: How Corporate Health Insurers Harm America’s Seniors. May 23, 3034.
Value-Based Payment - Failure to Achieve Program Goals:
MedPAC Report to Congress. The Medicare Advantage Program: Status Report. March, 2024.
Kahn JG, Sullivan KR. Promise vs. Practice: The Actual Financial Performance of Accountable Care Organizations. J Gen Intern Med 37(3); Aug 13, 2021:680–1.
Physician Moral Injury and Burnout:
Sinsky C, Colligan L, et al. Allocation of Physician Time in Ambulatory Practice: A Time and Motion Study in 4 Specialties. Ann Int Med 165(11) (2016):753-832.
Berwick D. Salve Lucrum: The Existential Threat of Greed in US Health Care. JAMA Jan. 30, 2023.
Hartzband P and Groopman J. Physician Burnout, Interrupted. NEJM May 1, 2020.
Hospital Global Budgets and Maryland Experience:
Himmelstein, Jun, Busse et. al. A Comparison of Hospital Administrative Costs in Eight Nations: US Costs Exceed All Others by Far. Health Affairs 33(9) (2014):1586-1594
Brennan TA. Maryland Hospital All-Payer Model: Can It Be Emulated? Health Affairs Forefront. May 31, 2022.
Galarraga J, Pines JM. The Challenging Transformation of Health Care under Maryland’s Global Budgets. Health Affairs Blog Dec. 19, 2017.
Machta R, Peterson G, et. al. Evaluation of the Maryland Total Cost of Care Model: Implementation Report. Mathematica Report. July 2021.
AHEAD Model:
Liao JM, Navathe AS. The AHEAD Model and the Potential to Advance Equity Through Population-Based Care. Health Affairs Forefront Oct. 30, 2023.
States Advancing All-Payer Health Equity Approaches and Development (AHEAD) Model. CMS.gov. 2024.
Narrow and “Ghost” Networks in Medicaid Managed Care:
Ludomirsky AB, Schpero WL, et al. In Medicaid Managed Care Networks, Care is Highly Concentrated Among a Small Percentage of Physicians. Health Affairs 42, No 5 (2022): 760-768.
Dolotina B, Turban J. Phantom Networks Prevent Children and Adolescents from Obtaining the Mental Health Care They Need. Health Affairs 41, No. 7 (2022): 1026-1028.
Zhu JM, Charlesworth CJ et al. Discrepancies Between Reported and Realized Mental Health Care Access in Oregon Medicaid. Health Affairs 41, No 7 (2022)
Connecticut Medicaid – Results of eliminating Medicaid Managed Care Organizations:
Connecticut Department of Social Services. Financial Trends in the Connecticut HUSKY Health Program – Transparency, Sustainability and COVID Impacts. Presentation to the Medical Assistance Program Oversight Committee. January 8, 2021.
Collaborative Care Model for Psychiatry:
About Collaborative Care. AIMS Center, University of Washington Dept. of Psychiatry and Behavioral Sciences.
Blackmore MA, Carleton KE, et al. Comparison of Collaborative Care and Colocation Treatment for Patients with Clinically Significant Depression Symptoms in Primary Care. Psychiatric Services 2018; 69:1184-1187.
Little V, Mallow A, Gatanaga O. Years after CMS Approved New Billing Codes for Collaborative Care, Confusion Persists. Health Affairs Forefront. Nov 22, 2022.