本文背景是美国医改想要建立一个叫ACO（Accountable Care Organization, 本文译为“可信赖医疗组织”），这个模式的目标就是要在提高医疗服务质量的同时降低医保支出。

目前试行的ACO模式是建立医师联合执业团体（Physician Group Practice，PGP）。这些团体需要完成美国医疗保险和医疗补助服务中心（CMS）的一些质量指标和开销指标（按小编理解，也就是达到一定医疗质量，同时控制医保支出，实现有结余）。如果完成了指标，作为奖励，CMS会把结余（saving）的80%给他们分红。现在这个ACO模式还在探索中，以上提到的PGP只是一种试行模式，该试行项目从2005年至2009年共试行了4年。本文主要就是通过这四年PGP的成果来探讨ACO模式的可行性。

作者: John K. Iglehart · 译者: Cindy · 阅读原文[en]

合理医疗费用法案（Affordable Care Act, ACA）主要条款之——建立新型医疗服务模式（delivery model）：可信赖医疗组织（Accountable Care Organization，ACO），受到了两党的一致支持。美国国会指导健康与人类服务部门（Department of Health and Human Services，DHHS）建立可信赖的医疗组织（ACO）项目，以求改善提供给医疗保险受益人的医疗质量，同时保留按服务计费的付费模式（fee-for-service payment [1]）。在这个项目下，医疗团体需要承担达成这些目标的义务，同时也能分享医疗保险产生的任何节余（savings）。

One of the few major provisions of the Affordable Care Act (ACA) with solid bipartisan support establishes a new delivery model: the accountable care organization (ACO). Congress directed the Department of Health and Human Services (DHHS) to develop an ACO program to improve the quality of care provided to Medicare beneficiaries and reduce its costs while retaining fee-for-service payment. Under this program, medical groups would have to take responsibility for achieving these goals and would share in any savings derived by Medicare.

尽管人们对ACO的兴趣大量涌现，然而却很少有人关注这项示范项目的结果，该项目由医疗保险和医疗补助服务中心（Center for Medicare and Medicaid Service，CMS）发起，作为医改法中关于ACO条款的范例。在医疗保险的医师联合执业示范中，CMS和十个大型的、有着不同组织架构的多专业团体签订了合同，这些团体包括独立的医师组织、学术人员实践组织（academic faculty practices）、综合服务系统（integrated delivery systems），以及小范围内的医师执业网络（network of small physician practices）。

Despite the burst of interest in ACOs, little attention has been paid to the results of a demonstration project sponsored by the Centers for Medicare and Medicaid Services (CMS) that was the model for the reform law’s ACO provisions. In the Medicare Physician Group Practice (PGP) demonstration, the CMS contracted with 10 large multispecialty groups with diverse organizational structures, including free-standing physician groups, academic faculty practices, integrated delivery systems, and a network of small physician practices.

作为整个医疗保险开销的一部分，按服务计费的医疗支出一直以来保持相对稳定（2009年占总开销4910亿美元中的13%）。然而，这种付费模式因为它固有的缺陷而遭受抨击，它刺激了医师医疗服务数量的增长，却不能保证质量。但是，政策制定者们清楚地记得管理式医疗（managed care）曾遭到的强烈反对，那种按人头收费（capitation [2]）的付费模式促使医生们的医疗服务缩水。因此，医疗保险中的医师付费政策这么多年来一直不变，在可预见的未来里，也没有新的可选方案。

As a share of total Medicare spending, fee-for-service expenditures for physician services have been relatively stable (13% of $491 billion in 2009). However, this payment model has been under attack because of its inherent incentive for increasing the quantity, but not necessarily the quality, of physician-delivered care. But policymakers vividly remember the backlash against managed care, whose capitation payments were seen as an incentive to stint on care, so with no new alternative to fee for service in the offing, Medicare’s physician-payment policy has remained essentially static.

2000年，美国国会下令DHHS试行以激励机制为基础的（incentive-based）医师付费方案，指导医疗保险去鼓励医疗合作以及对更高效的医疗服务程序的投资，并且对为改进医疗保健结果（health care outcomes）做出贡献的医师给予奖励。为此，CMS策划了医师联合执业（Physician Group Practices，PGP）项目，希望以此来检验医疗管理措施（care management initiatives）的有效性，看它是否能通过减少不必要的入院、重入院和急诊出诊来节约开支，同时改进服务质量。

In 2000, Congress tasked the DHHS with testing incentive-based payment methods for physicians, directing Medicare to encourage care coordination and investment in processes for more efficient service delivery and to reward physicians for improving health care outcomes. In response, the CMS designed the PGP project to examine whether care management initiatives could generate cost savings by reducing avoidable hospital admissions, readmissions, and emergency department visits, while improving quality

该示范（demonstration）开始于2005年4月，共有10个大型联合执业团体（PGP）参与（医师人数在232-1291人不等），它们在美国的不同区域开展工作。参与的医师照常获得医保按服务计费的酬劳，同时，如果执业医师们联合为他们所在的团体完成了指定的质量和花费指标（quality and cost targets），这些团体将获得医疗保险节余（Medicare’s savings）的80%作为分红（绩效报酬-performance payments）。为了限定绩效报酬的范围，这些团体必须为医疗保险的 A、B计划产生大于他们目标总支出（target expenditures）2%的节余，以排除随机波动的影响。CMS在同一地理区域建立了一个医疗保险受益者的团体作为参照，在考虑病例混合的基础上（adjusting for case mix），通过对比PGP和参照团体基准年的人均开销，设立支出目标（spending targets）。

The demonstration began in April 2005, with 10 large group practices (ranging from 232 to 1291 physicians) operating in various regions of the country. Participating doctors received their regular Medicare fee-for-service payments, but the groups were also eligible for an 80% share of Medicare’s savings (“performance payments”) if the practitioners collectively achieved specified quality and cost targets for the beneficiaries “attributed” to their group. To qualify for performance payments, groups had to generate savings for Medicare Parts A and B amounting to more than 2% of their target expenditures to rule out the possibility that savings merely represented random fluctuations. The CMS established the spending targets by creating a comparison group of Medicare beneficiaries in the same geographic area and comparing the PGP’s per capita expenditures in its base year with those for the comparison group, adjusting for case mix.

2010年12月9日，美国健康与人类服务部（DHHS）公布了该项目第四年的结果（截止于2009年3月31日）。届时，全部十个医师联合执业团体（PGPs）完成了32项质量目标中的29项，其中大部分是和冠心病、糖尿病、心衰、高血压和预防医疗相关的操作措施。有五个团体在完成他们开销指标的同时，还为医保产生了3870万美元的结余，赢得了3170万美元的绩效报酬。根据美国三角洲国际研究中心（RTI International）——这是一家评估CMS范例的研究机构，PGP将它们的结余归功于许多因素，包括组织架构、对医疗管理项目和重新策划医疗程序的投资、更为密集的诊断编码（more intensive diagnostic coding）以及市场情况的变化。其中威斯康星的Marshfield诊所赢得了全部绩效报酬的一半还多（合计1620万美元），正如它前几年做到的一样（见表）。它的PGP活动的负责人Dr. Theodore Praxel指出，诊所的成功归功于积极推动了多种举措的实施，包括健康信息技术（定点照护提醒point-of-care reminders，实现完全无纸化）、医疗管理项目，以及针对满足指定条件的一类病人而面向医护人员开展的教育和反馈。参与的团体没有因为未完成开销指标而受到经济上的惩罚。

On December 9, 2010, the DHHS reported results from the project’s fourth year (ending March 31, 2009), when all 10 PGPs met at least 29 of the 32 quality goals, most of which were process measures related to coronary artery disease, diabetes, heart failure, hypertension, and preventive care. By also beating their expenditure targets, five PGPs generated Medicare savings of $38.7 million, earning performance payments of $31.7 million. According to RTI International, a research institute that evaluated the demonstration for the CMS, PGPs attributed their savings to many factors, including organizational structure, investments in care management programs and redesigned care processes, more intensive diagnostic coding, and changes in market conditions. One group, the Marshfield Clinic in Wisconsin, earned more than half the total performance payments ($16.2 million), as it had in most previous years (see tableSummary Results of the Physician Group Practice Demonstration, Performance Years 1–4.). Dr. Theodore Praxel, who directed Marshfield’s PGP activities, ascribes the clinic’s success to an aggressive acceleration of “multiple initiatives . . . including health information technology (point-of-care reminders, being completely chartless), care management programs, [and] education and feedback to providers regarding populations of patients with a given condition.” Participating groups were not penalized financially for missing their expenditure targets.

在公布这些结果的同时，CMS行政官Donald Berwick称赞这十家团体是组织医疗服务的领袖，不过他还补充道：“现在我们想提高标准（raise the bar）。我们希望通过支持这些执业团体来证明，如果我们将正确的激励机制落到实处，美国医疗能够完成多少目标。CMS目前正在行动，要将这些医师组织转化进改革法下的ACO项目中去。”

In announcing the results, CMS Administrator Donald Berwick characterized the 10 groups as leaders in organizing care delivery, but he added: “Now we want to raise the bar. We want to support these practices to demonstrate just how much American medicine can achieve if we put the right incentives in place. CMS is currently working to transition these physician groups into the ACO program established under the reform law.”

CMS正在起草条款来指导ACO项目的执行，该项目计划2012年1月开始实施。尽管这项操作高度保密，但毫无疑问地，从PGP示范获得的相关经验已经被CMS分析过了。该项目产生的问题也被健康和人类服务部（DHHS）部长Kathleen Sebelius披露在一篇分析报告中，于2009年3月提交给国会。虽然这篇分析的结论仅基于该项目第二年的情况，但没有任何理由认为第四年的结果会有什么大的不同。

The CMS is drafting regulations that will guide implementation of the ACO program, scheduled to begin January 1, 2012. Although this exercise is highly confidential, the relevant lessons of the PGP demonstration have no doubt been analyzed by the CMS. And issues raised by the project were covered in an evaluation that Secretary of Health and Human Services Kathleen Sebelius sent to Congress in March 2009.Although its conclusions were based on the program’s second year, there’s no reason to believe that the fourth-year results would yield substantially different findings.

美国三角洲国际研究中心（RTI International）探究的问题之一就是，如果没有这个项目，这些参与的组织能否同样取得它们已经获得的这些经济效益。PGP中的4家在项目第二年就获得了节余，人均实际开销为334美元，低于他们的支出指标， RTI指出“这些绩效与示范前期保持一致”，提示即使没有示范项目，这些PGP之前“良好的花费趋势”也将继续保持。

One of the questions probed by RTI was whether the financial results that were achieved might have been obtained by the participating organizations even without the project. Four of the PGPs achieved savings in the second year of the program, with average actual expenditures of $334 per person below their spending targets, but RTI noted that “this performance was almost matched in the predemonstration period” and suggested that the previously “favorable cost trends” at these PGPs “might have continued had the demonstration not occurred.”

这4家在第二年就赢得了绩效报酬的PGP，有的拥有附属的学术医学中心（如Dartmouth-Hitchcock 和密歇根大学），有的是独立的医师组织（如Everett 和Marshfield）。另外5家作为综合服务系统（包括没有附属医学中心的医院系统）一部分的PGP及受附属医院（hospital affiliate）赞助的医师网络（如Middlesex）则没有获得任何绩效报酬。所有地方结余的大部分都来自门诊服务而非住院服务。RTI猜测，医院是对实现结余的潜在威胁，因为医院系统不太可能减少不必要的接诊或者提供低价医疗，不然就会影响他们的住院收入。RTI推断道，那两家参与的学术医学中心“或许可以用私人付费的接诊来弥补医保住院收入的损失”。

The four PGPs earning performance payments in the second year either were affiliated with an academic medical center (Dartmouth–Hitchcock and University of Michigan) or were free-standing physician groups (Everett and Marshfield). No performance payments were earned by the five PGPs that are part of integrated delivery systems (systems that include hospital ownership but are not affiliated with academic medical centers) or by the physician network (Middlesex) that is sponsored by a hospital affiliate. The majority of the savings at all sites occurred in outpatient, not inpatient, services. RTI hypothesized that the presence of a hospital was “a potential deterrent to achieving savings . . . since these systems may be unable to reduce avoidable admissions or use lower cost care substitutes without affecting their inpatient revenue.” The two participating academic medical centers “may be able to replace reductions in Medicare inpatient revenue with private pay admissions,” RTI reasoned.

因为这些PGP保留了按服务计费的付费结构，没有任何注册程序或绑定策略（lock-in feature）能防止受益人到别处求诊。即使这些受益人真的这样做了，他们归结于这些PGP产生的花费仍然需要这些PGP来承担，这进一步增加了PGP控制支出的难度。如果病人在某家PGP中接受了比别处更好的基础医疗服务，他们也被回顾性地（retrospectively）归属于（attributed to）这家组织。这些责任归派（attribution）的细节——是否通知、何时以及如何告知受益人这些PGP属于ACO的一部分，以及CMS应该向ACO提供病人群体的哪些数据——都引发了激烈的争论。

Because the PGP demonstration retained the fee-for-service structure, there was no enrollment process or lock-in feature preventing beneficiaries from seeking care elsewhere. If they did so, the PGP to which they had been attributed remained responsible for the cost of the care they received — which made it more difficult for a group to control its expenditures. Patients were attributed retrospectively to a PGP if they received more of their primary care from the group’s physicians than from anyone else. These details of attribution — whether, when, and how beneficiaries are informed that they’re part of an ACO and what data on the patient population will be provided to the ACO by the CMS — have provoked intense debate.

建立可信赖的医疗组织模式是一项探索中的工作，在起草新的项目条款时，CMS需要解决诸多问题。正如一位怀疑者所说，其中一个挑战就是“这项分红模式（shared savings model）的支持者们需要想出一个不会惹恼任何人的办法（an approach that attempts to upset or dislocate no one）”。既然示范的PGP即使没有完成目标也不会承担任何经济风险，很难说这些ACO的激励措施会对改变医生的行为起到足够有力的作用。如果医保受益人不知道他们被分配给了一所基于改进服务、减少花费措施而运营的医疗组织，他们将不会全力配合医疗管理。在一次一付的医疗付费模式下，病人不必担心出于经济利益的考虑医疗服务会缩水。而正如在PGP范例中那样，如果医疗组织没有及时获得病人医疗消费的趋势（patients’ utilization trends），他们将不会知道自己的表现如何。同时，为了减少巨大的财政赤字，美国国会需要在ACO模式证实其成效之前，采取更广泛激进的手段来减慢医保支出的增长速度。

The ACO model is a work in progress, and the CMS must address many questions in crafting the new program’s regulations.One challenge, as a skeptic described it, is that “proponents of the shared savings model have designed an approach that attempts to upset or dislocate no one.”Since the PGPs in the demonstration faced no financial risk if they missed their targets, it’s hard to say whether the incentives for ACOs will be strong enough to change physicians’ behavior. If beneficiaries are unaware that they’ve been assigned to a medical group that operates on incentives for improving care and reducing costs, they may not be able to become full partners in managing their care. Under the fee-for-service model, patients have not had to worry that financial considerations might lead to stinting on care. And medical groups won’t know how they’re performing if, as in the PGP demonstration, they aren’t provided timely feedback on their patients’ utilization trends. Meanwhile, as it attempts to reduce the vast budget deficit, Congress may need to take more sweeping steps to slow the growth in Medicare spending long before the ACO model can prove whether it is up to meeting these challenges.

[1] 按服务计费（fee-for-service），是指病人可以自由选择就诊的医疗机构，并依据所接受的医疗服务直接付费，事后再向保险公司提出报销申请，这是相对于按人头收费（Capitation）来说的。在这种付费模式下，为了获得更多收入，医疗机构都会积极为病人提供各方面的服务，实现医疗服务的数量增长，但却无法保证质量。病人也必须为这种自由选择的服务付出较高的医疗费用。—-编者按

[2]按人头收费（capitation）,是指医疗组织与一些健康保持机构签订合同，进入健康保持机构的病人会被分配到这些医疗组织，也就是说病人求医时不能自由选择。在这种付费模式下，医疗组织的收入来自于按病人数的定额配给，即对于每个求诊病人，医疗组织的收益是定值，不管他是否接受了医疗服务。因此，医疗机构也就不会积极地去为患者提供服务。—-编者按

来源：《新英格兰医学杂志》2011-1-20  观察

Assessing an ACO Prototype — Medicare’s Physician Group Practice Demonstration. John K. Iglehart.   N Engl J Med 2011; 364:198-200.