Accountable Care Organizations (ACOs) in Medicare fee-for-service aim to “bend the cost curve,” reducing expenditures while not compromising quality of care.1 The Center for Medicare and Medicaid Innovation (CMMI), designed by the Affordable Care Act to test innovative payment models in Medicare and Medicaid, broadly defines ACOs as “groups of doctors, hospitals, and other health care providers, who come together to give coordinated high quality care to Medicare patients.”1
Created in 2012, the Pioneer ACO model is a shared-savings program: if the participating ACOs reduce expenditures below national benchmarks they are allowed to retain a portion of those savings with the Centers for Medicare and Medicaid Services (CMS), but if an organization spends more than the reference, they pay CMS. The model is similar to, but has larger potential risks and rewards, than the Medicare Shared Savings Program where ACOs faced financial incentives to meet quality standards.2 At the end of its 3rd year, CMS boasted $384 million in savings, a boost in average quality, and the Office of the Actuary at CMS certified the program for expansion.2,3,4
After the decision was announced, there was both excitement and no shortage of critics. This program was the first demo test to move beyond “pilot jail,” as one journalist put it, but had CMS jumped the gun? Although the program met the established criteria for expansion, critics point to legitimate flaws in the design and evaluation of the program, such as an imperfect payment system, the resulting dropouts, limited replicability on the national level, and the indecipherable nature of ACOs themselves. The flaws could significantly hinder the success of the program beyond the pilot and warrant an examination.
A flawed payment model based on national benchmarks
13 ACOs dropped out of the program before its conclusion. Those who dropped from the program argued “the model was financially detrimental”5 and called the payment formula “skewed.” Interestingly, eight of the 13 no longer participating organizations in the Pioneer Model have transferred to the MSSP, which has smaller incentives but less risk than the Pioneer program.13 A lack of regional nuance in the payment formula was a legitimate critique of the Pioneer ACO model.5,7. Indeed, the evaluation shows that the surviving ACOs largely operated “in markets that have above-average health spending,” and therefore more opportunity to create savings when compared to national averages.6 For example, the payment model did not account for changes in area wage index, otherwise a component of Medicare reimbursement.
Additionally, only 10 ACOs generated 4% savings in the first year. Although 4% might be significant if scaled nationally, no participant with losses in the first year saved money in the second year, and savings in the 2nd and 3rd years (when payment started to be affected by performance) were as little as a third of the first year’s savings.8 The evidence suggests, therefore, that successful organizations tackled “low-hanging fruit” to cut costs, leaving much less savings to be generated from improved performance in future years without significant changes to the payment structure.5
The model is not widely replicable
The general ACO model assumes a “high degree of cohesion between physicians and their community hospitals,” but critics argue this level of cohesion is “not found in many communities in America.”9 CMMI sought and used ACOs that were well established, highly integrated systems with appropriate experience. Whereas there are other such systems, especially in urban centers, critics stated that “pioneer ACO contract incentives would [not] work for organizations that are less advanced.” 10,9 This critique echoes others who believe the ACO model is only workable in urban centers or densely populated areas and will be difficult to expand throughout the country (although regulators attempt to address this by placing higher thresholds on the number of fee-for-service members in each Pioneer).9 Indeed, the Brookings Institute demonstrated that many of the participants clustered towards lower savings or losses, but with higher quality, largely ended up being ACOs in less densely populated areas.10 Within these same findings were comments that the ACO payment model could adversely affect low income, higher need populations.
The evaluation is not evidence that “ACOs” work
Even though CMS reported significant results at the end of the first 3 years, some critics say that the meaning of the evaluations is “inscrutable.”12 For example, the analysis showed that a vast majority of ACOs had a drop in primary care office visits. The increased number of visits is a “perplexing result,” given the theory that ACOs emphasize low-cost settings, which often involve primary care physicians.12 I would argue part of the problem stems from the conceptualization of an ACO by CMS; ostensibly, their definition provides little guidance as to how the doctors, hospitals, and other healthcare providers are supposed to come together to provide better care. The evaluators and future organizations therefore do not know what ACOs that saved money did that others did not (other than stay in the test); “the ACO ‘treatment’ under investigation is not a prescribed set of activities or interventions.”12 Under these conditions, it is difficult to imagine that evaluators could establish a scientific understanding of how ACOs work.
The certification of the Pioneer ACO Program could be justified by technical criteria, but ignores underlying issues that complicate the reality of the evaluation and the results. CMMI and CMS can find valuable insights from the test, but points raised by the critiques make it difficult to conclude that ACOs will be the boon of improved care coordination and reducing Medicare costs. Attrition in the pilot was a main and concerning challenge for CMS. Although CMS drew participants for the Pioneer program from ACOs already deemed well qualified and could make quick cuts to spending, only the “most efficient” of these organizations or organizations from specific regions could benefit financially and stay in the pilot. Most strikingly, the evaluation does not make clear what exactly these ACOs did to create the savings or maintain quality. In this sense, the model test became a “survival of the fittest” over time, rather than a representative sample of organizations from across the country that could be incentivized to reduce costs through a higher risk shared savings program. In summary, I echo arguments that the program will only be viable in a small number of systems already in place to tackle quick fixes, rather than provide incentives for significant reductions in national expenditures through large participation.4
By: Michael Budros, MPH/MPP Candidate, University of Michigan School of Public Health, Department of Health Management and Policy and the Gerald R. Ford School of Public Policy
1. CMS.gov. 2015. “Pioneer ACO Model Frequently Asked Questions.” Centers for Medicare and Medicaid Services. Accessed online, Sep. 2015: http://innovation.cms.gov/initiatives/Pioneer-ACO-Model/Pioneer-ACO-FAQs.html.
2. Nyweide, David, et al. June 2015. “Association of Pioneer Accountable Care Organizations vs Traditional Medicare Fee for Service With Spending, Utilization, and Patient Experience.” JAMA: 2015; 313(21):2152-2161. Accessed online, Sep. 2015: http://jama.jamanetwork.com/article.aspx?articleid=2290608.
3. McCanne, Don. 2015. “HHS moving forward with Pioneer ACO model” Physicians for a National Health Program. Accessed online, Sep. 2015: http://pnhp.org/blog/2015/05/05/hhs-moving-forward-with-pioneer-aco-model/.
4. Casalino, Lawrence. 2015. “Pioneer Accountable Care Organizations Traversing Rough Country.” JAMA: Editorials. Accessed online, Sep. 2015: http://jama.jamanetwork.com/article.aspx?articleid=2290606.
5. Evans, Melanie. 2014. “Medicare ACOs improve quality, have mixed results on slowing spending, CMS says.” Modern Healthcare. Accessed online, Sept. 2015: http://www.modernhealthcare.com/article/20140916/NEWS/309169938/medicare-acos-improve-quality-have-mixed-results-on-slowing-spending.
6. Evans, Melanie. “Medicare’s Pioneer program down to 19 ACOs after three more exit”. Modern Healthcare. Accessed online, Sept. 2015: http://www.modernhealthcare.com/article/20140925/NEWS/309259938.
7. Dahukey, Aytan and Bhupathy, Vinay. “Pioneer ACOs: Slowed Health Spending, Improved Quality and more Drop Outs.” Healthcare Law Blog: SheppardMullin. Accessed online, Sept. 2015: http://www.sheppardhealthlaw.com/2014/10/articles/centers-for-medicare-and-medicaid-services-cms/pioneer-acos-slowed-health-spending-improved-quality-and-more-drop-outs/.
8. London, Susan. 2015. “Pioneer ACO Model Cuts Medicare Costs, Receives Certification.” Medscape. Accessed online, Sept. 2015: http://www.medscape.com/viewarticle/844195#vp_2.
9. Gamble, Molly. 2013. “ACOs: The Least Agreed-Upon Concept in Healthcare?” Becker’s Hospital Review. Accessed online, Sept. 2015: http://www.beckershospitalreview.com/accountable-care-organizations/acos-the-least-agreed-upon-concept-in-healthcare.html.
10. Evans, Melanie. 2015. “Successful Pioneer ACO journey leaves faint trail for followers.” Modern Healthcare. Accessed online, Sept. 2015: http://www.modernhealthcare.com/article/20150507/NEWS/150509934.
11. Kocot, Lawrence, et al. “A More Complete Picture of Pioneer ACO Results.” Brookings Institute. Accessed online, Sept. 2015: http://www.brookings.edu/blogs/up-front/posts/2014/10/09-pioneer-aco-results-mcclellan#recent_rr/.
12. Sullivan, Kip. “CMS’s latest ACO report is inscrutable”. Physicians for National Health Program. Accessed online, Sept. 2015: http://pnhp.org/blog/2015/06/15/cmss-latest-aco-report-is-inscrutable/.
13. Pham, H. H., Cohen, M., & Conway, P. H. (2014). The Pioneer accountable care organization model: improving quality and lowering costs. JAMA, 312(16), 1635-1636.