hospital-physician collaborations: antitrust and …/media/hogan-lovells/pdf/... · 2017-01-27 ·...

PLEASE DO NOT COPY, DISTRIBUTE OR CITE WITHOUT THE PERMISSION OF

THE AUTHORS

HOSPITAL-PHYSICIAN COLLABORATIONS: ANTITRUST AND HEALTH CARE FRAUD AND ABUSE

CONSIDERATIONS

By Robert F. Leibenluft, Esq., Jonathan L. Diesenhaus, Esq.,

Eric M. Baim, Esq. & Leigh L. Oliver, Esq.

Hogan & Hartson LLP 555 13th Street, NW

Washington, DC 20004 (202) 637-5600

http://www.hhlaw.com

“Accepted for publication in the Health Law Handbook, 2009 Edition. Alice G. Gosfield, Editor, ©Thomson Reuters.

A complete copy of the handbook is available from West by calling 1-800-382-9252 or online at www.west.thomson.com.”

www.hhlaw.com

Part III

ACCOUNTABILITY ANDENFORCEMENT

Chapter 7

Hospital-Physician Collaborations:Antitrust and Health Care Fraudand Abuse Considerationsby Robert F. Leibenluft, Esq., Jonathan L. Diesenhaus,Esq., Eric M. Baim, Esq., and Leigh L. Oliver, Esq.

§ 7:1 Introduction§ 7:2 Analytical framework under antitrust laws§ 7:3 —Does the venture involve agreements which might be

construed as per se o�enses?§ 7:4 —Is there su�cient integration to avoid per se

condemnation so that the joint venture can be analyzedunder the rule of reason, i.e., does the venture have thepotential to achieve substantial e�ciencies?

§ 7:5 —Even if there is substantial integration, are thecompetitive restraints “ancillary” to the venture'sprocompetitive goals?

§ 7:6 —Will the joint venture, on balance, have anticompetitivee�ects?

§ 7:7 Analytical framework under federal health care fraudand abuse laws

§ 7:8 —The civil monetary penalty for the reduction orlimitation of services

§ 7:9 —Anti-kickback statute§ 7:10 —The Stark law§ 7:11 Application of antitrust and fraud and abuse laws to

hospital/physician collaborations§ 7:12 —Infrastructure and return on investment§ 7:13 —Selection, breadth, and removal of physicians

251

§ 7:14 Incentives§ 7:15 —Establishing appropriate incentives under antitrust

law—Basis and amount of the �nancial incentives§ 7:16 — —Individual vs. group performance§ 7:17 —Establishing appropriate incentives under fraud and

abuse laws—Shared savings (“gainsharing”) advisoryopinions

§ 7:18 — —Incentive payment advisory opinion§ 7:19 — —Stark law exception for hospital shared savings and

incentive payments§ 7:20 —Putting it all together§ 7:21 Conclusion

KeyCiteL: Cases and other legal materials listed in KeyCite Scope can beresearched through the KeyCite service on WestlawL. Use KeyCite tocheck citations for form, parallel references, prior and later history, andcomprehensive citator information, including citations to other decisionsand secondary materials.

§ 7:1 Introduction

It is becoming increasingly clear that prevailing structuresfor delivering hospital and physician services in the UnitedStates often frustrate e�orts to improve quality and enhancee�ciency, and that greater clinical collaboration is a prereq-uisite for achieving both. However, federal and parallel statelaws present hurdles, some real and some perceived, to re-structuring delivery models for improving quality and e�-ciency through collaboration among physicians and hospitals.This article addresses two sets of these legal constraints—governing antitrust and health care fraud and abuse laws—and examines the extent to which they can limit e�ectivecollaboration, and how some of these obstacles can beovercome.

Historically, most physicians have practiced in smallgroups, and they continue to do so—currently over 50% of allphysicians are in groups of �ve or less.1 These small prac-tices lack the scale and infrastructure that can optimally

[Section 7:1]1Hing E. Burt CW, Characteristics of O�ce-Based Physicians and

their Medical Practices: United States 2005–2006, National Center for

Health Law Handbook

252

support health information technology, interface e�cientlywith third-party payors, and develop and maintain special-ized expertise. Hospitals, too, are constrained in how theycan improve the quality of care provided in their facilities.With the exception of the oversight hospitals exercise overradiologists, anesthesiologists, pathologists, and otherhospital-based physicians, they typically do not employ norexercise extensive control over the physicians who practicein their facilities. In fact, hospitals are habitually reluctantto impose restrictions on their voluntary medical sta�—physicians upon whom they depend for admissions. Thus,hospitals rarely turn to the blunt instrument of withholdingprivileges to discipline “independent” physicians who resistinitiatives to enhance quality and e�ciency.

The impact of these constraints on quality and cost hasbeen widely chronicled.2 Moreover, the need for morecoordinated care among physicians, and between physiciansand hospitals, is only intensifying as payment systems areincreasingly rewarding care that can most e�ciently beprovided through collaborations. Thus, for example, Medicareand other payors are turning to “pay-for-performance” (P4P)programs to incentivize providers to improve quality. Theseprograms, which are focused primarily on hospital outcomesmeasures, reward hospitals for improvements that largelydepend on the hospitals' ability to change the behavior andclinical practices of their medical sta� over whom they havelimited control. While physicians in individual or small grouppractices can participate in P4P programs, larger groups areoften better equipped to monitor data and implementorganized processes that will enable them to fully implementthe desired changes and take advantage of P4P incentives.3

The importance of collaborations will be even greater ifcertain payment reforms are instituted that rely on paying

Health Statistics, Vital Health Stat. 13 (166), at 4 (2008), available at http://www.cdc.gov/nchs/data/series/sr�13/sr13�166.pdf.

2See generally Institute of Medicine, Crossing the Quality Chasm: ANew Health System for the 21st Century (2001) [hereinafter, CrossingQuality Chasm].

3Jon Christianson, Sheila Leatherman, & Kim Sutherland, RobertWood Johnson Foundation, Paying for Quality: Understanding and Assess-ing Physicians Pay-for-Performance Initiatives, at 4 (2007).

§ 7:1Hospital-Physician Collaborations

253

bundled payments for an entire episode of care furnished bya hospital along with its physicians and other providers.4

These trends may result in an increase in the number ofphysicians who are employed by hospitals, as well as thegrowth in both single and multispecialty group practices.However, history suggests that this increase will be gradualas many physicians resist change, wishing to retain theirautonomy and continue to practice independently in individ-ual or small group practices. Thus, the question ariseswhether there is a practical way in which physicians cancontinue in their independent practices, but collaborate witheach other, and with hospitals, to achieve signi�cant qualityand cost improvements.

Collaborations among independent economic entities arecommon throughout the economy. When they involve agree-ments among competitors, particularly if these include priceor price-related terms, they can raise serious antitrustissues. As described in more detail below, a crucial elementin addressing these antitrust issues is to establish that theagreement is reasonably necessary or “ancillary” to an“e�ciency-enhancing integration of economic activity.” Muchattention has been focused on what should be considered inevaluating the extent of such an integration among healthcare providers, particularly where the providers are notcontracting with health plans on the basis of capitation orsome other form of shared �nancial risk, which is typicallythe case today. The federal antitrust enforcers have statedthat such “clinical integration” programs typically willinvolve: (1) mechanisms to monitor and control health careutilization to control costs and assure quality; (2) selectivelychoosing participating providers; and (3) signi�cant invest-ment of capital, both �nancial and human, in the necessary

4See, e.g., S.M. Shortell and L.P. Casalino, Health Care ReformRequires Accountable Organizations, 300 J.A.M.A. 95 (2008); Gos�eld, ThePrometheus™ Payment Program: A Legal Blueprint, Health LawHandbook (Gos�eld, ed. 2007) (proposing new payment systems using“Evidence-informed Case Rates™” that would compensate all of the provid-ers who furnish services for a speci�c clinical condition); Harold S. Luft,Total Cure, The Antidote to the Health Care Crisis, at 91–92 (2008)(describing a model of payments for episodes of care to “care deliveryteams”).

§ 7:1 Health Law Handbook

254

infrastructure.5 Moreover, some believe that �nancial incen-tives are especially important to provide the mechanisms in(1), above, needed for a successful clinical integrationprogram.6

However, collaborations that involve payments fromhospitals to physicians directly implicate federal health carefraud and abuse laws. These laws generally are designed toguard against the in�uence �nancial incentives may have onphysician referral patterns and clinical judgment. Thecriteria for selecting participating physicians, and how ahospital might provide infrastructure support to the physi-cians, also can raise fraud and abuse issues. In short, clini-cal integration that involves hospital/physician arrange-ments could be viewed with favor under the antitrust lawswhile heightening concerns under the fraud and abuse laws.

The task for providers who wish to navigate these issuesis made more di�cult by the fact that there is little relevantcase law, and both the antitrust enforcers as well as thegovernment o�cials responsible for applying the fraud andabuse statutes have not always given either clear guidanceor encouragement. In their defense, the enforcers themselveshave a di�cult challenge. The antitrust law is very broadlywritten, and of general applicability to all industries, so it iscrucial that whatever guidance the antitrust enforcersprovide is consistent with their approach in other settings.The fraud and abuse laws, in contrast, are very speci�c tothe health care industry, but in some respects their lack of�exibility may make it harder to accommodate productivecollaborations. Both sets of enforcers are also concerned thatwhatever guidance they adopt not “straitjacket” providersand deter innovation, yet at the same time they do not wishto foreclose their ability to prosecute egregious arrangementsthat few would argue constitute acceptable conduct.

5Dep't of Justice & Fed. Trade Comm'n, Statements of AntitrustEnforcement Policy in Health Care (1996), available at http://www.usdoj.gov/atr/public/guidelines/0000.pdf [hereinafter, Health Care PolicyStatements].

6FTC Commissioner Thomas Rosch has said, “Those incentives arewhat makes �nancial integration work. If a clinical integration programincludes a very strong system of rewards and punishment, . . . it could besuccessful.” J. Thomas Rosch, Commissioner, FTC, Remarks at the 2007Antitrust in Health Care Conference, Clinical Integration in Antitrust:Prospects for the Future, at 17 (Sept. 17, 2007), available at http://www.ftc.gov/speeches/rosch/070917clinic.pdf [hereinafter, Rosch Remarks].


255

This article takes a close look at these issues, andexamines how several key aspects of clinical integrationprograms are analyzed under both the antitrust and fraudand abuse laws. It identi�es where hospitals and physiciansretain some latitude to pursue carefully crafted collaborativee�orts under existing law and government enforcementpolicy.7 That said, we also identify some of those areas wheremore �exible guidance and permissive structures would beinstrumental in helping achieve the Institute of Medicine'srecommendation to transform health care quality through,among other things, aligning payment policies and qualityimprovement.8

We begin with some “basics” for those who are not expertsin both antitrust and fraud and abuse law.

§ 7:2 Analytical framework under antitrust lawsThe primary antitrust concern raised by a collaboration

among competing providers is whether it might violate theprohibition in section 1 of the Sherman Act against “[e]verycontract, combination . . . or conspiracy, in restraint oftrade.”1 In decisions spanning more than a century, courtshave identi�ed certain agreements that are viewed as solikely to have an anticompetitive e�ect that they can becondemned without any inquiry into the relevant markets

7Of course, entities considering hospital-physician collaboration willneed to consider a number of additional laws that are outside the scope ofthis paper. For example, tax-exempt organizations such as nonpro�thospitals would need to evaluate—under the Internal Revenue Code andapplicable Internal Revenue Service guidance—any incentive programthat might arguably result in private inurement of net earnings to “insid-ers” or confer impermissible bene�t on private parties. The IRS uses a va-riety of factors to assess whether incentive compensation could jeopardizea hospital's tax-exempt status. See, e.g., I.R.S. Info. Ltr. 2002-0021 (Jan.9, 2002), available at http://www.irs.gov/pub/irs-wd/02-0021.pdf, whichdetails factors that will be considered in determining whether incentive-based compensation arrangements result in private inurement orimpermissible private bene�t.

8Crossing the Quality Chasm, at 18 (recommending that “allpurchasers, both public and private, should carefully reexamine their pay-ment policies . . . to remove barriers that currently impede qualityimprovement, and to build in stronger incentives for quality enhance-ment”).

[Section 7:2]115 U.S.C.A. § 1.


256

that might be a�ected by the agreement, and whether theagreement might have procompetitive aspects that, on bal-ance, could redeem the conduct. When agreements aredeemed per se unlawful, the parties are not given an op-portunity to defend their conduct by claiming that theirprices are “reasonable” or that their conduct causes goode�ects. Examples of per se conduct include price �xing andmarket allocation agreements (e.g., competitors agreeingwith each other to divide up a market so that they do notcompete with each other in certain geographic areas or withrespect to certain products or services).

In contrast, the federal antitrust agencies have observedthat if otherwise competing entities join together in an“e�ciency-enhancing integration of economic activity” and“enter into an agreement that is reasonably related to theintegration and reasonably necessary to achieve its procom-petitive bene�ts,” the agreement is analyzed under the so-called “rule of reason.”2 Analysis under the rule of reasonrequires a much more extensive inquiry into the actual com-petitive e�ects of the venture, including market power, andantitrust plainti�s �nd it much more di�cult to prevail in achallenge under the rule of reason.3

Thus, the key antitrust issues raised by a joint ventureamong health care providers essentially involve addressingthe following four questions.4

2U.S. Dep't of Justice & Fed. Trade Comm'n, Antitrust Guidelinesfor Collaborations Among Competitors (2000), at 8, available at http://www.ftc.gov/os/2000/04/ftcdojguidelines.pdf [hereinafter CollaborationGuidelines]. This article refers often to the Collaboration Guidelines, andto the industry-speci�c Health Care Policy Statements, because they re�ecthow the federal antitrust agencies interpret and intend to apply theantitrust laws. While not binding on the courts, conduct that is consistentwith these guidelines will likely be di�cult to challenge successfully.

3This is a somewhat simpli�ed summary of the relevant law that isevolving and rather complex. For example, recent decisions suggest adeparture from a simple dichotomy between per se and rule of reasonconduct, and there have been various formulations regarding how theseissues should be addressed by the courts. The two-step framework shouldsu�ce, however, for purposes of the analysis herein.

4A full discussion of all the antitrust issues raised by joint venturesis beyond the scope of this article. To simplify matters, this discussionfocuses only on the antitrust issues that arise under Sherman Act § 1regarding agreements that competitors may enter into with each other


257

§ 7:3 Analytical framework under antitrust laws—Does the venture involve agreements whichmight be construed as per se o�enses?

If the venture does not involve an agreement amongcompetitors relating to the prices they will charge, thecustomers they will serve, or other matters that arguablycould be condemned as per se illegal, then the venture willbe analyzed under the rule of reason. While this does notmean there are no potential antitrust concerns, the antitrustrisks are much less serious, as a challenge will depend onthe ability of a complainant to demonstrate that the parties

through the formation and operation of a joint venture. The formation of ajoint venture under certain circumstances also may be subject to analysisunder Clayton Act § 7 which governs mergers and acquisitions. That anal-ysis is essentially the same, however, as the analysis described here of anintegrated competitor collaboration under the rule of reason. Thus, a jointventure that survives antitrust scrutiny under Sherman Act § 1 will likelyalso pass muster under Clayton Act § 7 review. Joint ventures that do notinvolve competitors also can raise antitrust issues, but these will almostalways be evaluated under the rule of reason and, unless they involveproviders who have substantial market shares (i.e., more than 35%–40%),are unlikely to be problematic. If such ventures do involve a dominantprovider, they are not necessarily illegal, but a more extensive antitrustanalysis would be required to address potential issues such as monopoliza-tion and tying.


258

have market power1 and that their conduct, on balance, willhave an anticompetitive e�ect.2

§ 7:4 Analytical framework under antitrust laws—Isthere su�cient integration to avoid per secondemnation so that the joint venture can beanalyzed under the rule of reason, i.e., does theventure have the potential to achievesubstantial e�ciencies?

Assuming that the venture does involve agreements that,standing alone, might be per se illegal, the next question iswhether such agreements are not “naked” restraints, butrather are related to what the federal antitrust agencieshave called “an e�ciency-enhancing integration of economicactivity.” The agencies comment that in such arrangements,

participants collaborate to perform or cause to be performed(by a joint venture entity . . .) one or more business functions,such as production, distribution, marketing, purchasing orR&D, and thereby bene�t or potentially bene�t, consumers byexpanding output, reducing price, or enhancing quality, ser-vice or innovation. Participants in an e�ciency-enhancingintegration typically combine, by contract or otherwise, signif-

[Section 7:3]1There is no bright-line market power test. The federal antitrust

agencies have established a “safety zone” for joint ventures generallywhere they have indicated that “absent extraordinary circumstances” they“will not challenge a competitor collaboration when the market shares ofthe collaboration and its participants collectively account for no more thantwenty percent of each relevant market in which competition may be af-fected.” Collaboration Guidelines, at § 4.2. The federal agencies haveestablished a safety zone for �nancially integrated physician networks of30% share for nonexclusive networks and 20% share for exclusivenetworks. Health Care Policy Statements, at 79–80. However, these are“safety zones” and therefore conservatively drawn, and parties may havegreater market shares (particularly if the venture is nonexclusive) withoutnecessarily having market power. Moreover, even if the collaboration andits participants have a substantially larger market share, it is possiblethat they may lack market power because entry into the market bycompetitors is relatively easy. Unfortunately, de�ning the relevant prod-uct and geographic markets and assigning market shares can be a di�cultand complex task, even if reliable data are available, which often may notbe the case.

2See § 7:6 for the analysis with respect to anticompetitive e�ect of ajoint venture.


259

icant capital, technology or other complementary assets toachieve procompetitive bene�ts that the participants could notachieve separately. The mere coordination of decisions onprice, output, customers, territories, and the like is not integra-tion, and cost savings without integration are not a basis foravoiding per se condemnation.1

In the context of health care collaborations, the federalHealth Care Policy Statements focus on whether the col-laboration involves su�cient “�nancial” or “clinicalintegration.” Joint price negotiations that are ancillary to�nancially or clinically integrated physician organizations,therefore, will not be condemned as per se unlawful horizon-tal price-�xing agreements, but rather should be consideredunder a rule of reason analysis.2 The Health Care PolicyStatements recognize that there are many di�erent arrange-ments that have the potential to achieve substantiale�ciencies. On one end, there are group practices that arefully integrated where, among other things, doctors sharerevenues and losses, contribute to salaries of support sta�and administrative personnel, develop practice protocols,and maintain joint insurance coverage. On the other end ofthe spectrum are ventures that have little or no �nancialrisk sharing, but instead implement clinical programs thatintegrate physician and hospital care to improve quality ande�ciency across a broad network of providers. The determi-nation of whether a venture is either �nancially or clinicallyintegrated is a critical step in assessing a venture's legalityunder the antitrust laws.

Guidance from the antitrust agencies is fairly straightfor-ward with respect to the elements necessary to meet the safeharbors of �nancial integration because the agencies aremore con�dent that such arrangements have the potential toachieve signi�cant e�ciencies. Financial integration exists ifproviders provide services at a shared capitated rate, orwhere payment is subject to a substantial �nancial with-holding depending on whether group performance goals aremet. In �nancially integrated ventures, participants areclearly incentivized to cooperate in controlling costs andimproving quality by managing the provision of services

[Section 7:4]1Collaboration Guidelines, at § 3.2.2Health Care Policy Statements, at 3 n.27.


260

because a failure to achieve the goals across the ventureresults in lost revenue for all participants.

The available guidance from the agencies with respect toclinical integration is not as simple. In 2002, the FTC issuedan Advisory Opinion to address the proposed clinical integra-tion program of MedSouth, an independent practice associa-tion (IPA) in south Denver, Colorado.3 The FTC ultimatelyconcluded that MedSouth's proposed clinical integration ofphysician practices had the potential to generate substantial�scal, administrative and quality-related e�ciencies andthus, as long as MedSouth was able to implement theprograms and achieve the e�ciency goals laid out in its pro-posal, the program would be analyzed under a rule of reasonanalysis.4 Nearly �ve years after the initial MedSouthopinion,5 the FTC issued an Advisory Opinion to the GreaterRochester Independent Physician Association (GRIPA) stat-ing that it had no intention to issue a challenge with respectto GRIPA's nonexclusive physician network joint venture.The FTC recognized some key elements of GRIPA's program,including: in-network referrals; monitoring and measuringindividual and group performance against benchmarks; aWeb-based network that allows physicians to share patientinformation and order lab tests and prescriptions electroni-cally; a peer review system to identify underperformingmembers; and substantial investment by physicians of time

3Letter from Je�rey W. Brennan, Assistant Dir. Health Care Servs.& Prods., FTC, to John J. Miles, Principal, Ober/Kaler, 2002 WL 463290(Feb. 19, 2002), available at http://www.ftc.gov/bc/adops/medsouth.htm[hereinafter MedSouth Opinion].

4MedSouth Opinion at *8. Some of the speci�c indicia of integrationin the MedSouth proposal included: a single clinical resource managementtool, computer and data systems, that facilitates increased communica-tions and cooperation among physicians regarding treatment and practicepatterns; required �nancial investment by member physicians in thehardware necessary to participate in the Web-based clinical computersystem; and clinical practice protocols and ongoing monitoring of physi-cian performance.

5In June 2007, the FTC sta� issued a second opinion with respect toMedSouth. Letter from Markus Meier, Assistant Dir. Health Care Servs.& Prods., FTC, to John J. Miles, Principal, Ober/Kaler (June 18, 2007),available at http://www.ftc.gov/bc/adops/070618medsouth.pdf. In thisfollow-up letter, the FTC sta� concluded that MedSouth's responses torequests from the FTC for up-to-date information on the status of its clini-cal integration program supported its 2002 opinion.


261

and e�ort to collaborate to develop and oversee implementa-tion of program practice guidelines.

In 2007, the American Hospital Association (AHA), in ane�ort to provide more complete guidance to its memberhospitals and health care providers, released its ProposedGuidance on Establishing Clinical Integration Programs.6

The document addresses practical considerations in develop-ing a clinical integration program and also expands on thelegal analysis provided in the Health Care Statements andFTC Advisory Opinions.

§ 7:5 Analytical framework under antitrust laws—Even if there is substantial integration, are thecompetitive restraints “ancillary” to theventure's procompetitive goals?

If a joint venture involves substantial integration, thenext step is to examine the agreements made in connectionwith the joint venture to determine whether they are “ancil-lary,” that is, related and reasonably necessary to achievethe procompetitive goals of the venture. For example, twohospitals might decide to form a joint venture to buy a mobilelithotripsy unit that would provide services at each of thehospitals half the time. Agreements regarding the purchase,operation, and sale of the technical component of thelithotripsy services (including the price for such services) areall likely to be viewed as ancillary to the joint venture.However, an agreement regarding the prices for servicesthat are not furnished through the joint venture, for examplethe daily hospital room charge, prices for cardiac services, orthe prices that independent urologists would charge for theirprofessional services,1 are likely to be viewed as not ancillaryto the joint venture and will be condemned as per se illegal.

Thus, this step of the analysis requires consideration ofthe subject of any agreements that the parties make with

6Thomas B. Leary et al., Guidance for Clinical Integration WorkingPaper (Hogan & Hartson LLP for the American Hospital Association)(2007), available at http://www.aha.org/aha/content/2007/pdf/070417clinicalintegration.pdf.

[Section 7:5]1Decision and Order, Matter of Urological Stone Surgeons, Inc., 125

F.T.C. 513, 1998 WL 34077361 (1998), available at http://www.ftc.gov/os/caselist/c3791.shtm.


262

each other to ensure that they are related and reasonablynecessary to achieving the legitimate goals of the jointventure.2

§ 7:6 Analytical framework under antitrust laws—Will the joint venture, on balance, haveanticompetitive e�ects?

If the competitive restraints are ancillary to the jointventure's procompetitive goals, then it is necessary todetermine under the rule of reason whether, on balance, theventure will be anticompetitive. This assessment involves (1)a further examination of the nature of the relevant agree-ments and the type of competitive harm and bene�ts thatmay result; and (2) an assessment of competitive conditionsin the market, including whether the parties to the jointventure will have market power. These two inquiries may besomewhat interrelated. For example, if an agreement on itsface looks as if it very likely will have signi�cant anticompeti-tive e�ects and few procompetitive bene�ts, it might be chal-lenged without a detailed analysis of the market. Conversely,in other situations, it may be possible to conclude that theagreement is unlikely to have substantial anticompetitive ef-fects, whether or not the parties have market power, andtherefore the antitrust review also could be concludedwithout an extensive market power assessment.

In analyzing the e�ect of the agreement, considerationwill be given to the rationale for the restraint and whetherits goals could be accomplished by less restrictivealternatives. For example, are the joint venture partiespermitted to compete with the collaboration and are they al-lowed to independently set prices for the output of the jointventure? Similarly, is the venture nonexclusive so that theparties can compete with each other with respect to servicesthat are not furnished through the venture? The extent towhich members of the venture can compete to sell their ser-vices outside of the venture may impact the e�cienciesgenerated through the venture but may also reduce theventure's market power, thereby balancing the possibility ofgreater anticompetitive harm. Other anticompetitive con-

2For a discussion of ancillarity in the context of clinical integration,see Robert F. Leibenluft & Theresa E. Weir, Clinical Integration: Assess-ing the Antitrust Issues, Health Law Handbook 32 (Gos�eld, ed. 2004).


263

cerns include whether the collaboration can facilitate collu-sion among members of the venture, for example, throughsharing competitively sensitive information and whether theduration of the venture is reasonably related to the timerequired to achieve and sustain the e�ciency-enhancinggoals of the venture.

§ 7:7 Analytical framework under federal health carefraud and abuse laws1

Several federal health care fraud and abuse laws have asigni�cant impact on the ability of hospitals to form �nancialrelationships with physicians. These laws creates additionalhurdles to be overcome in order for hospitals and physiciansto achieve the bona �de integration that can be essential tosurviving the antitrust scrutiny outlined above. The threefederal fraud and abuse laws that are most likely implicatedin relation to �nancial incentives under clinical integrationmodels are the federal civil monetary penalty (CMP) for ahospital's payment to a physician to induce reductions orlimitations of services,2 the Anti-kickback statute (AKS),3

and the physician self-referral (Stark) law.4

These laws attempt, in di�erent ways, to address a concern

[Section 7:7]1There are other laws that would need to be considered in addition

to the federal health care program fraud and abuse laws discussed here.For example, many states have fraud and abuse laws, such as anti-kickback and self-referral prohibitions, which apply only in connectionwith items or services paid for through the state's medical assistance ben-e�t. In addition, other state laws, often referred to as “all-payor” laws, ap-ply regardless of whether federal health care program patients areinvolved. Finally, a number of states have what are known as “fee split”laws which generally prohibit a licensed health care provider from divid-ing professional fees with another person or entity, especially when thatother person or entity has referred a patient. Although these laws di�erfrom state to state regarding the types of providers and payors that arecovered, with distinctions among the laws as to the scope of the generalprohibition and relevant exceptions, the principles underlying theseprohibitions are generally the same as the federal health care programfraud and abuse laws, so this analysis of these three laws provides a goodframework for discussion.

242 U.S.C.A. § 1320a-7a(b)(1).342 U.S.C.A. § 1320a-7b(b)(2)(D), (E).442 U.S.C.A. § 1395nn; 42 C.F.R. § 411, Subpart J.


264

that �nancial incentives to physicians can, among otherthings, induce referrals that are not based on what is bestfor the patient, encourage unnecessary medical procedures,in�uence physicians to reduce or limit services, subsidizemarginal providers, or undercut “honest competition.”5 Eachlaw is implemented largely independently by di�erent agen-cies and, by design or e�ect, each of these statutes can cre-ate signi�cant tensions around �nancial incentives paidthrough hospital-physician clinical integration models. Asdiscussed below, other aspects that are important to a clini-cal integration program, such as criteria for excludingproviders or funding for infrastructure, also can raiseconcerns under the fraud and abuse laws.

§ 7:8 Analytical framework under federal health carefraud and abuse laws—The civil monetarypenalty for the reduction or limitation ofservices

Of the fraud and abuse laws we will review, the CMP lawis probably the most straightforward, and the one that pre-sents the most obvious obstacle to creating patient- orprocedure-speci�c �nancial incentives for physicians toreduce hospital costs. The law simply provides that a civilmonetary penalty may be imposed against a hospital that“knowingly makes a payment, directly or indirectly, to aphysician as an inducement to reduce or limit services” andagainst a physician who “knowingly accepts receipt of [such]payment.”1 Penalties for violation of the statute, which isenforced by the Department of Health and Human ServicesO�ce of Inspector General (OIG), include a civil moneypenalty of not more than $2,000 for each violation, additional

5See, e.g., Issues Related to Physician “Self-Referrals”: Hearingsbefore the Subcommittee on Health and the Subcommittee on Oversight ofthe House Committee on Ways and Means, 101st Cong., 1st Sess. 27–28(1989) (Statement of Rep. Stark, House Comm. on Ways & Means). AsRep. Stark explains in his statement to the subcommittee, one of theconcerns from physician self-referrals is that “[h]onest competition, isundercut. To maintain market share, suppliers are being forced to compete— not on price or quality — but on the ‘cut’ they give physicians.”

[Section 7:8]142 U.S.C.A. § 1320a-7a(b)(1), (2).


265

assessments, and/or exclusion from participation in federalhealth care programs.2

The prohibition is very broad. The CMP law is violatedwhenever the hospital intends that the payment induce thephysician to “reduce or limit” services and need not be tiedto an actual reduction in the care provided to the patient.The OIG in a 1999 Special Advisory Bulletin addressed theapplication of the CMP law to “gainsharing” arrangements,where hospitals share a percentage of cost-savings realizedwhen physicians adopt hospital-prescribed protocols thatrequire them to standardize the use of certain products orthe approach to treating certain conditions in ways that areintended to reduce the hospitals' costs. In its Special Advi-sory Bulletin, the OIG made clear its view that gainsharingarrangements “appropriately structured . . . may o�er sig-ni�cant bene�ts where there is no adverse impact on thequality of care received by patients” but are nonethelessprohibited by the CMP law, irrespective of medical necessity.3

That proof of actual adverse e�ects on particular patients isnot required under the CMP law was emphasized in an openletter by the OIG in response to criticism of its Special Advi-sory Bulletin.4

The impetus for the CMP law makes somewhat under-

242 U.S.C.A. § 1320a-7a. “Federal health care programs” includeMedicare, Medicaid, Tricare, and other payments (with the exception ofthe Federal Employees Health bene�t Program) arising out of federalfunds. See 42 U.S.C.A. § 1320a-7b(f) (de�ning “federal health careprograms”).

3Gainsharing Arrangements and CMPs for Hospital Payments toPhysicians to Reduce or Limit Services to Bene�ciaries, 64 Fed. Reg.37985, 37985 (July 14, 1999) [hereinafter Gainsharing SAB] (“[A]nyhospital incentive plan that encourages physicians through payments toreduce or limit clinical services directly or indirectly violates the statute);see also OIG Advisory Op. 07-21 (Dec. 28, 2007) (“[W]hether current medi-cal practice re�ects necessity or prudence is irrelevant for the purpose ofthe CMP”).

4D. McCarty Thornton & Kevin G. McAnaney, Dep't of Health &Human Servs., Recent Commentary Distorts HHS IG's GainsharingBulletin (1999), available at http://oig.hhs.gov/fraud/docs/alertsandbulletins/bnagain.htm (“The authors of the commentary contend that the ‘plainlanguage’ of the statute requires both that a violator have the speci�cintent to induce treating physicians to withhold medically necessary ser-vices from their patients and that the incentive plan actually cause areduction or limitation of medically necessary services. In our view, thisinterpretation is plainly wrong. Simply put, the language of the statute


266

standable, if not more acceptable, its sweeping impact on�nancial incentives under clinical integration programs:policymakers were concerned that the Medicare diagnosticrelated group (DRG) prospective payment system (PPS) thatwas adopted in 1983 had drastically changed the �nancialincentives for hospitals.5 Thus, the General Accounting Of-�ce (GAO) commented that where under the prior system—cost reimbursement—hospitals had incentives to encouragephysicians to “admit more patients, leave them in thehospital longer, and use more services while they werethere,” under PPS, the incentives could lead hospitals to,among other things, “underprovide services [and] dischargepatients too early.”6 Although hospitals have an incentive toreduce cost under PPS, physicians are paid separately underMedicare Part B (on a fee-for-service basis) and have noincentive to either save hospital costs or reduce the utiliza-tion of their services. Therefore, some hospitals had designedphysician incentive plans to align the physician interestswith those of the hospital. At Congress' request, the GAOreviewed the incentive plans. In its report, the GAOconcluded that, despite including some features designed toreduce �nancial incentives to give substandard care, “nocombination of features can guarantee that a plan will notbe subject to abuse.”7 In short, the CMP law was enacted inresponse to these physician incentive plans which, in theeyes of Congress, “may create a con�ict of interests that maylimit the ability of the physician to exercise independentprofessional judgment in the best interest of his or herpatients.”8

The CMP law gives rise to the most direct constraints on�nancial and clinical integration between physicians and

refers to ‘services,’ not ‘medically necessary services,’ and requires a show-ing of an intent to induce a reduction of services, not an actual reduction”)(emphasis in original).

5The prospective payment system was enacted as part of the SocialSecurity Amendments of 1983, Pub. L. No. 98-21, 51 U.S.L.W. 203 (1983),and was phased in during �scal years 1984–1987.

6U.S. General Accounting O�ce, Physician Incentive Payments byHospitals Could Lead to Abuse, GAO/HRD-86-103, at 9 (July 22, 1986).

7U.S. General Accounting O�ce, Physician Incentive Payments byHospitals Could Lead to Abuse, GAO/HRD-86-103, at 23.

8Gainsharing SAB, at 37986 (quoting H.R. Rep. No. 99-727, at 441(1986)).


267

hospital where the goal of integration is e�ciency. That said,through the OIG's advisory opinion process,9 gainsharingprograms have emerged that, although the programs wouldconstitute an inducement to reduce or limit services tofederal health care program bene�ciaries in violation of theCMP law, the OIG has been persuaded that the programscontained su�cient safeguards intended to protect againstinappropriate reduction in services such that the OIG woulddecline to impose sanctions. As a result, the CMP law hurdle,while a high one, is not insurmountable for those pursuingenhanced provider collaboration.

§ 7:9 Analytical framework under federal health carefraud and abuse laws—Anti-kickback statute

The AKS prohibits, among other things, giving or receiv-ing any bene�t or “remuneration” in exchange for, or toinduce, the referral of any patients for, or the purchase,lease, order, or recommendation of, any facility, item, or ser-vice for which payment may be made under Medicare,Medicaid, and most other federal health care programs.1 TheAKS covers “any remuneration” whether “in cash or inkind.”2 The OIG is responsible for interpreting the AKS and,along with the Department of Justice (DOJ), enforcing it.Penalties for violation of this felony statute can includesubstantial criminal �nes and imprisonment, possible exclu-sion from participation in federal health care programs, andcivil monetary penalties.3

The OIG and a majority of courts have adopted the posi-

9The OIG, in consultation with the DOJ, issues written advisoryopinions with regard to the application of the AKS and other OIG healthcare fraud and abuse sanctions, including the CMP law. However, the ad-visory opinions may legally be relied upon only by the requestor. See 42U.S.C.A. § 1320a-7d(b); 42 C.F.R. Part 1008.

[Section 7:9]142 U.S.C.A. § 1320a-7b(b)(2)(D), (E).242 U.S.C.A. § 1320a-7b(b)(1), (2).342 U.S.C.A. § 1320a-7 (exclusion from federal health care programs);

§ 1320a-7a (civil monetary penalties of up to $50,000 per act plus threetimes the remuneration); 42 U.S.C.A. § 1320a-7b(b) (imprisonment of upto �ve years or criminal �nes of $25,000 or both); 18 U.S.C.A. § 3571(augmenting penalties: $250,000 per violation for individuals and $500,000per violation for entities).


268

tion that conduct can violate the AKS if “one purpose” (asopposed to a sole or primary purpose) of the remuneration isto induce a recommendation or referral for covered items orservices reimbursable under a federal health care program,though it can be a defense if the purpose is “incidental” ornot “material.”4 A number of courts have also held that mere“encouragement” to obtain, or the “hope or expectation” ofobtaining, program-related business does not, in and of itself,constitute an “inducement.”5

The statute is quite broad and “encompass[es] many harm-less or e�cient arrangements.”6 Accordingly, Congress andthe OIG have created a series of statutory “exceptions”7 andregulatory “safe harbors.”8 These exceptions and safe harborsidentify the criteria of speci�c payment practices that do notviolate the AKS and could protect, for example, paying physi-cians as employees (under the safe harbor for employment)9

or for a physician's medical directorship position (under thepersonal services and management contracts safe harbors).10

According to the OIG, one function of these safe harbors is“to permit physicians to freely engage in business practicesand arrangements that encourage competition, innovation,and economy.”11 Although an arrangement that �ts into oneor more of these exceptions or safe harbors is immune fromprosecution, arrangements that do not �t squarely within anexception or safe harbor do not necessarily violate the AKS,but certainly may be subject to scrutiny and challenge.12

The OIG's general approach to arrangements that mighttechnically violate the AKS is to assess them on a case-by-case basis under “the many factors which are part of thedecision-making process regarding case selection for investi-

4See 56 Fed. Reg. 35952, 35958 (July 29, 1991); see also U.S. v.Greber, 760 F.2d 68, 72 (3d Cir. 1985).

5See Hanlester Network v. Shalala, 51 F.3d 1390, 1398 (9th Cir.1995); U.S. v. McClatchey, 217 F.3d 823, 835 (10th Cir. 2000).

654 Fed. Reg. 3088, 3088 (Jan. 23, 1989).7See 42 U.S.C.A. § 1320a-7b(b)(3).8See 42 C.F.R. § 1001.952.942 C.F.R. § 1001.952(i).

1042 C.F.R.§ 1001.952(d).1154 Fed. Reg. 3088 (Jan. 23, 1989); 42 C.F.R. § 1001.952.1242 U.S.C.A. § 1320a-7b(b)(3)(E); 56 Fed. Reg. at 35954.


269

gation and prosecution.”13 The OIG frequently identi�esthese factors as increased risk of overutilization, increasedprogram costs, adverse impacts on patient freedom of choice,and unfair competition.14

Clinical integration models implicate the AKS in at leasttwo respects. First, the structure of the clinically integratedventure implicates the AKS whenever anything of value�ows from a person or entity that receives referrals to aperson or entity in a position to make such referrals. Integra-tion ranging from the formation a physician-hospital organi-zation or similar joint venture, or a hospital employment ofa physician to help design a clinical integration program,must be analyzed under the AKS to determine whether theirstructure provides payment or other remuneration to thephysician to induce or reward his or her referrals. Second,the payments themselves made under an otherwise innocu-ous structure implicate the AKS where one purpose of thosepayments is to induce or reward referrals of health careitems or services. Incentive payment and shared saving ar-rangements often cannot satisfy a safe harbor. The personalservices and management contracts safe harbor, for one, willnot protect such payments because hospital payments to thephysicians are often not set in advance and do not �t theexisting paradigm of fair market value fees being measuredin terms of hours of service provided.15

§ 7:10 Analytical framework under federal healthcare fraud and abuse laws—The Stark law

Subject to certain exceptions, the Stark law prohibitsphysician referrals for certain “designated health services”to an entity with which the physician or the physician's im-mediate family member has �nancial relationship (either anownership or investment interest or compensation arrange-

1356 Fed. Reg. at 35954. An advisory opinion process is available forthe AKS. See Gainsharing SAB, at 37986 (quoting H.R. Rep. No. 99-727,at 441 (1986)).

14See 56 Fed. Reg. at 35954.15See 73 Fed. Reg. 69726, 69798 (Nov. 19, 2008) (“[P]roperly

structured arrangements involving physician participation in an incentivepayment or shared savings program may meet the requirements of one ormore of the existing [Stark] exceptions for compensation arrangements”).


270

ment), if such services may be reimbursed by Medicare.1

“Designated health services” (DHS) under the Stark lawinclude inpatient and outpatient hospital services andcertain ancillary services for which physicians may referpatients to hospitals.2 Unlike the safe harbors under theAKS—which are su�cient, but not necessary, to protectagainst an alleged violation—if a physician has a �nancialrelationship that falls within the Stark law, that �nancialrelationship must meet one of the exceptions for any refer-rals to be lawful. A physician is subject to a civil monetarypenalty if he or she knowingly makes a noncompliant refer-ral, as is the entity, such as a hospital, that knowingly makesa Medicare claim for services provided pursuant to anoncompliant referral. In addition, a hospital is liable forany reimbursement related to services ordered by the self-referring physician, regardless of whether the hospital knewthat the referral was noncompliant. If the government canshow that a hospital knew or should have known that theclaim should not have been billed, then additional penaltiesmay be available under the civil False Claims Act.3

As with the AKS, the Stark law presents a challenge forhospital-physician clinical integration with respect to anyarrangement that will result in a �nancial relationship be-

[Section 7:10]142 U.S.C.A. § 1395nn; 42 C.F.R. § 411, Subpart J. Although the

Stark law does not apply to Medicaid referrals and billing directly, underthe Medicaid statute arrangements that would violate this law mayprevent a state Medicaid program from receiving federal matching fundsfor services furnished to Medicaid patients pursuant to such arrangements.See 42 U.S.C.A. § 1396b(s).

242 C.F.R. § 411.351. “Designated health services” are de�ned asclinical laboratory services; physical therapy services; occupationaltherapy services; radiology services; radiation therapy services and sup-plies; durable medical equipment and supplies; parenteral and enteralnutrients, equipment and supplies; prosthetics, orthotics, and prostheticdevices and supplies; home health services; outpatient prescription drugs;and inpatient and outpatient hospital services. 42 U.S.C.A. § 1395nn.

331 U.S.C.A. § 3729 (providing for $5,500 to $11,000 per claim plusthree times the amount of the claim). The applicability of the FCA toStark law violations adds another twist to enforcement in this areabecause the FCA can be enforced by whistleblowers in addition or as analternative to government enforcement. See, e.g., U.S. ex rel. Thompson v.Columbia/HCA Healthcare Corp., 20 F. Supp. 2d 1017, 1047 (S.D. Tex.1998).


271

tween the hospital and the physician either through owner-ship (with joint physician-hospital owners) or throughcompensation (from incentive payments or shared savings).Although a number of the relevant Stark law exceptions(e.g., those for fair market value services,4 personal services,5

employment compensation,6 or indirect compensationarrangements)7 might be applicable to purchases of physi-cian services, the elements of those exceptions are di�cult toapply in the context of hospital-physician collaborations thato�er incentive payments to encourage hospital savings, thatmay then be shared with the physicians, or which rewardquality of care.

The analysis of any clinical integration model betweenhospitals and physicians under the Stark law requires acareful consideration of the speci�c type of �nancial relation-ship that is created. The analysis may vary depending onwhether it is a �nancial relationship created through owner-ship (e.g., with hospital or physician owners of a PHO entity),through compensation (e.g., bonus payments), or somecombination thereof. Likewise, it may vary depending onwhether the �nancial relationship is a direct one betweenthe parties or an indirect one with some entity interposedbetween the hospital and physician. To understand the ob-stacle that the Stark law presents, however, we can lookgenerally at some of the elements that appear in some, butnot all, of the potential exceptions noted above.

Fair market value for services. Establishing whether pay-ments for services are “fair market value” is a challengewhere the “services” are not easily translated into an hourlywage (the typical analytical framework for what constitutesfair market value). For example, what is the fair marketvalue for the e�ort required to meet a clinically based perfor-mance outcome measure required under a clinical integra-tion incentive payment program? Likewise, under a sharedsaving program, what should be the share of the savings inorder to re�ect the fair market value of the e�ort needed toachieve those savings? Where the savings result from prod-

442 C.F.R. § 411.357(l).542 C.F.R. § 411.357(d).642 C.F.R. § 411.357(c).742 C.F.R. § 411.357(p).


272

uct substitution of a lower cost item for a higher cost item,does the fact of a greater di�erential between the high andlow cost items justify the higher payment to the physician?Often in shared savings programs every physician thatparticipates is paid an equal, or per-capita, amount of thesavings. As a result, those who did not contribute the samelevel of e�ort may be getting the same payment.

Set in advance. Several Stark law exceptions require thatcompensation be “set in advance.” Under the Stark law,compensation can be considered “set in advance” if the ag-gregate compensation, a time-based or per-unit of service-based (whether per-use or per-service) amount, or a speci�cformula for calculating the compensation is set in an agree-ment between the parties before the furnishing of the itemsor services for which the compensation is to be paid.8 In2007, responding in part to perceived abuses regarding leasepayments based on a percentage of the revenues generatedby use of the space or equipment, CMS had proposed thatpercentage-based compensation arrangements could only be(1) used to pay for “personally performed physician services”;and (2) based on the revenues directly resulting from thephysician services rather than based on some other factor“such as a percentage of the savings by a hospitaldepartment.”9 Were CMS to have adopted such an interpre-tation, compensation paid under shared savings programwould not have met an exception that required compensa-tion to be “set in advance.” However, CMS ultimatelyprohibited the use of percentage-based compensationformulae only in the determination of rental charges forspace and equipment leases. As a result, CMS's current posi-tion on whether percentage-based compensation for otherthan personally performed physician services is somewhatunclear. CMS has noted that its regulations do not prohibitpercentage-based compensation arrangements for nonprofes-sional services like management or billing services but, indoing so, CMS referenced a proposed new exception for incen-

842 C.F.R. § 411.354(d)(1). “The formula for determining thecompensation must be set forth in su�cient detail so that it can beobjectively veri�ed, and the formula may not be changed or modi�ed dur-ing the course of the agreement in any manner that takes into account thevolume or value of referrals or other business generated by the referringphysician.” 42 C.F.R. § 411.354(d)(1).

972 Fed. Reg. 38122, 38184 (July 12, 2007).


273

tive payment and shared savings programs.10 As discussedfurther in § 7:19 below, as of the date of this writing, suchan exception has not been promulgated.

Payment not based on volume or value of referrals. Anumber of the potentially applicable exceptions include therequirement that payments are not determined in a mannerthat takes into account the volume or value of any referralsor other business generated by the referring physician to theDHS entity (e.g., the hospital). It may be possible to structurea program that arguably does not “take into account” thevolume or value of any referrals or other business generatedto the physician where the program includes the safeguardsthat have been included in arrangements already blessed bythe OIG through its advisory opinion process.11 For example,the payments could be distributed on a per-capita basis toall participating physicians and not be directly related to thesavings attributed to the individual physician or, in the caseof an incentive payment program, to his or her compliancewith the performance criteria. In such a case, an argumentcould be made that the referrals made by an individualphysician are not related to the cost savings achieved in ashared savings program (e.g., because that particular physi-cian might not have used the lower cost products) or thepayments made pursuant to an incentive payment program(e.g., because that particular physician might not haveachieved the clinical outcome targets). On the other hand,every time there is a referral there is at least an opportunityto earn a payment. The larger the pool of physicians in theprogram, the more attenuated is the link between each indi-vidual physician's actions and the payment achieved underthe program.

§ 7:11 Application of antitrust and fraud and abuselaws to hospital/physician collaborations

There are substantial di�erences in the nature of theantitrust and fraud and abuse laws as they apply to hospital-physician collaborations. As noted above, in the context ofprovider collaborations, the principal antitrust law is the

1073 Fed. Reg. 48434, 48711 (Aug. 19, 2008) (referring to 73 Fed. Reg.38502, 38548 (July 7, 2008)).

11See § 7:17.


274

simple Sherman Act § 1 prohibition against unreasonablerestraints of trade that applies to joint ventures of all kindsand in all industries. Although the federal antitrust agencieshave issued guidelines from time to time explaining howthey interpret the antitrust laws, they emphasize that theyare not “regulators” and are reluctant to be too speci�c intheir guidance out of concern that it would deter entitiesfrom trying innovative solutions. In contrast, the fraud andabuse laws were passed to address speci�c concerns regard-ing arrangements among health care providers who arereimbursed under certain federal programs. The lawsthemselves are more detailed and are enforced by individu-als at CMS and OIG who do view themselves as regulatorsand are accustomed to drawing more “bright lines” in regula-tions regarding what conduct is or is not lawful. In bothcases, of course, the courts are the ultimate arbiter as towhether conduct is unlawful, but there have been very fewcourt decisions on provider collaborations. For this reason,the guidance and actions taken by the antitrust and fraudand abuse enforcers (through advisory opinions, consents,and speeches) is given particular weight.

As outlined in the preceding sections, antitrust lawrequires that competitor agreements be ancillary to anintegrated venture's procompetitive goals, subject to a ruleof reason analysis. To the antitrust enforcers, the extent of aventure's infrastructure and its funding, its rules concerningwho can participate in the venture and the size of theventure, and the existence and nature of �nancial incentivesare key features in assessing the extent of clinical integra-tion, and whether a venture is likely to raise seriousantitrust concerns. Such factors are also closely scrutinizedfor compliance with the fraud and abuse laws, but with dif-ferent goals in mind. Under these laws, the government isconcerned that the arrangement may disguise an induce-ment in return for referral, or in the case of the CMP law,involve a payment that could be attributed to a reduction initems or services made to federal health care programbene�ciaries.

This section discusses the extent to which the antitrustand fraud and abuse laws, as re�ected in the actions takenby the government enforcers, may create tensions with re-spect to the venture's infrastructure, and in decisions regard-ing participation criteria and the breadth of the provider


275

panel. Section 7:14, below, presents a comparable analysiswith respect to the venture's �nancial incentives.

§ 7:12 Application of antitrust and fraud and abuselaws to hospital/physician collaborations—Infrastructure and return on investment

Clinical integration programs often require substantialinvestment in infrastructure. Costly infrastructure mayinclude, for example, an interconnected electronic health re-cords (EHR) system, including the hardware and softwarenecessary to deploy such a system and the contracts or sta�necessary to program and maintain it. EHR programsprovide physicians and hospitals with real-time access topatient records, as well as up-to-date clinical software, whichcan improve health outcomes.1 Another example of infra-structure is nursing sta� to implement clinical programs,including data and chart reviews.

Under the antitrust analysis, a substantial infrastructurecan help establish that the venture is real and may have thepotential for signi�cant e�ciencies—that is, it is not a sham.The source of funds to pay for infrastructure costs can berelevant to the antitrust analysis because such funding mayre�ect elements of �nancial integration. Physician invest-ment in both startup and maintenance costs for infrastruc-ture evidences actual �nancial integration across physicians,although a collaboration could be economically integratedeven without such contributions. For example, a venturemay utilize, expand, or improve a participating hospital'sexisting EHR system, and sta� could �nd this to be morepractical and e�cient than seeking investments from itsphysicians for an entirely new system.

The source of infrastructure �nancing, in particular therelative return and “return on investment,” may haveimplications under the fraud and abuse laws. Fraud andabuse laws look at these types of ventures to determinewhether the structure is aimed at improving quality and ef-

[Section 7:12]1See eHealthInitiative, eHealth Initiative Blueprint, Building

Consensus for Common Action, 11 (2007), available at http://www.ehealthinitiative.org/blueprint/eHiBlueprint-BuildingConsensusForCommonAction.pdf.


276

�ciency, or is merely a vehicle through which hospital money�ows to participating physicians, directly or as a means ofreducing the cost of operating a physician practice, as areward for past referrals, and/or an incentive to maintainthose referrals. In this respect, the fraud and abuse analy-sis—like the threshold antitrust analysis—is focused onconsidering whether the venture is no more than a sham.However, consistent with their more regulatory approach,the fraud and abuse enforcers have been more speci�c inoutlining particular requirements in connection withinfrastructure. For example, in its Special Fraud Alert onJoint Venture Arrangements, the OIG outlined some of thecharacteristics of suspect joint ventures, many of which arenot intended “to raise investment capital legitimately tostart a business, but to lock up a stream of referrals fromthe physician investors and to compensate them indirectlyfor these referrals.”2 The OIG identi�ed a nonexhaustive listof examples of questionable features, “which separately ortaken together may result in a business arrangement thatviolates the [AKS].”

A number of these questionable features are relevantconsiderations when structuring a clinical integrated entityor arrangement that comports with the antitrust laws. Forexample, the OIG expresses concerns over business struc-tures that are best characterized as only a “shell”; in otherwords, is the structure indicative of a legitimate joint ventureor merely a way to disguise kickbacks? The OIG alsoscrutinizes closely the terms of �nancing and pro�t distribu-tions and questions joint venture situations where theamount of capital invested by the physicians are “dispropor-tionately small and the returns on investment . . . dispropor-tionately large when compared to a typical investment in anew business enterprise.” Other considerations are whetherphysicians are investing only a nominal amount or are paid“extraordinary returns on the investment in comparison withthe risk involved.”

In achieving clinical integration, some types of dispropor-tionate investment on the part of one entity would not beunexpected. As noted above, it may be more practical andeconomically e�cient for the hospital to expand its existing

2See Special Fraud Alert: Joint Venture Arrangements, 59 Fed. Reg.65373, 65374 (Dec. 19, 1994).


277

EHR infrastructure to physician o�ces—a feature thatwould be looked upon favorably under the antitrust analysisif it facilitated ways in which the physicians could workinterdependently with each other. However, the OIG has alongstanding concern that the provision of free or reducedprice goods or services to an existing or potential referralsource may be used as a vehicle to disguise or confer anunlawful payment for referrals of federal health careprogram business, and the provision of free technology hasbeen no exception. In a 1994 Special Fraud Alert on Arrange-ments for the Provision of Clinical Lab Services, the OIGidenti�ed as among the examples of inducements o�ered byclinical laboratories to providers which may implicate theAKS the “provision of computers or fax machines, unlesssuch equipment is integral to, and exclusively used for, per-formance of the outside laboratory's work.”3

Recent developments indicate that the enforcers may beapplying somewhat more �exibility to their fraud and abuseanalysis of unequal contributions to mutually bene�cialinfrastructure investments. In 2006, CMS and the OIGpromulgated exceptions to the Stark law and AKS prohibi-tions to permit certain donations of software and equipmentfrom hospitals to physicians for establishing electronicprescribing and EHR capabilities. The OIG drafted an AKSsafe harbor for the donation of interoperable electronichealth records software or information technology and train-ing services to “protect bene�cial arrangements that wouldeliminate perceived barriers to the adoption of electronichealth records without creating undue risk that the arrange-ments might be used to induce or reward the generation ofFederal health care program business.”4 CMS created aparallel Stark law exception at the same time.5 Mostimportant for this discussion is the recognition on the part ofthe OIG and CMS that “certain transfers of health informa-tion technology between parties with actual or potentialreferral relationships may further the important nationalpolicy of promoting widespread adoption of health informa-

3See 59 Fed. Reg. at 65374.4See 71 Fed. Reg. 45110 (Aug. 8, 2006) (codi�ed at 42 C.F.R.

§ 1001.952(y)).5See 71 Fed. Reg. 45140 (Aug. 8, 2006) (codi�ed at 42 C.F.R.

§ 411.257(w)).


278

tion technology to improve patient safety, quality of care,and e�ciency in the delivery of health care.”6 The agencieshave since abandoned their earlier “exclusivity” require-ment, and instead have adopted an approach that allows forthe donation of EHR software packages that are “necessaryand used predominantly” to create, maintain, transmit, orreceive EHR but may include other software that is function-ally directly related to care and treatment of individualpatients (e.g., patient administration, scheduling functions,billing, and clinical support software).7 Among the types ofdonations that would not be protected are those for personal,nonmedical purposes or to o�ce sta�.

Although the breadth of the Stark law and AKS EHRexceptions suggests that the law can be applied so as toadapt to changes in health care delivery in the interest ofquality care and e�ciency, the speci�city of those exceptionsand the steps required to develop them also demonstratethat the fraud and abuse laws have been far more rigid inapplication than the analysis of venture integration underthe antitrust laws. Thus, it was only after years of discus-sion, a GAO study of barriers to the development and deploy-ment of EHR systems,8 and the appointment of a WhiteHouse EHR “czar,” that regulatory change came about, andthen only as the result of a congressional mandate that theagencies remove barriers to electronic prescribing.

671 Fed. Reg. at 45113. Note that the agencies also created a Starklaw exception (42 C.F.R. § 411.357(v)) and AKS safe harbor (42 C.F.R.§ 1001.952(x)) for the provision of items and services that are necessaryand used solely to transmit and receive electronic prescription informationincluding hardware, software, Internet connectivity, and training andsupport services. The electronic prescribing exception and safe harborwere mandated by Congress in section 101 of the Medicare PrescriptionDrug, Improvement, and Modernization Act of 2003 (MMA). The electronichealth records exception and safe harbor were promulgated by CMS andthe OIG under the authority of sections 1877(b)(4) (Stark law) and 1128B(b)(3)(E) (AKS) of the Social Security Act.

7Interestingly, the congressionally mandated electronic prescribingexception and safe harbor were limited to those items and services thatwere necessary and “used solely” to transmit and receive electronicprescribing information.

8See U.S. General Accounting O�ce, “HHS's E�orts to PromoteHealth Information Technology and Legal Barriers to its Adoption,” GAO-04–991R (Aug. 13, 2004).


279

§ 7:13 Application of antitrust and fraud and abuselaws to hospital/physician collaborations—Selection, breadth, and removal of physicians

Selection and removal criteria for membership in aprovider network is relevant to the antitrust analysis of aclinically integrated network because it can shed light onwhether a network is likely to achieve substantial e�cien-cies. The use of strict membership participation criteria canbe a very useful tool for �rst limiting the provider panel toindividuals who meet certain minimum standards, and thenas an ultimate “enforcement mechanism” to ensure adher-ence to the venture's program. However, as with many otheraspects of clinically integrated networks, there may be varia-tions in approach, and some ventures may choose, particu-larly at the outset, to be less strict in their membershipcriteria in order to o�er a wider breadth of services or toimprove the quality of care across a wider spectrum ofproviders through the clinical program itself.1

The relative size of a venture's provider panel can be animportant element of the antitrust analysis under the rule ofreason because generally networks with a high share of therelevant market are more likely to have market power. Afull market power analysis requires de�ning the relevantproduct market (i.e. which types of physician specialtiesprovide services that are substitutes for each other) and rel-evant geographic market (i.e., from what geographic areacan substitute providers be found), and considering whatbarriers to entry into the market exist. Also, market powerconcerns will be reduced if a network is truly nonexclusive,that is, its members are free to contract directly with healthplans outside of the network. The federal antitrust agenciesin their Health Care Policy Statements include a “safetyzone” for nonexclusive �nancially integrated networks with

[Section 7:13]1Note that exclusion or expulsion of speci�c providers from a venture

could prompt an antitrust lawsuit from that provider. Such challenges,however, are similar to challenges to a denial of hospital privileges andare rarely successful because, among other things, the plainti� typicallyhas di�culty showing that competition has been actually harmed as op-posed to simply harming competing providers. See American Bar Assoc.,Antitrust Law Developments 122 (6th ed., 2007) (citing case law in thearea of antitrust challenges to denial of hospital privileges).


280

market shares of less than 30%, and for exclusive networkswith market shares of less than 20%.2 These are just safetyzones, however, and networks with substantially higher mar-ket shares may still not have market power. However, clearlythe extent to which a network has a high market share inone or more specialty will increase the potential for antitrustscrutiny.

From the fraud and abuse perspective, the selection ofphysicians to participate in a venture should be based on thenumber and specialty of physicians necessary to successfullyachieve the cost savings or the quality or patient safety goalsfor which the collaboration is formed—i.e., each willing par-ticipant, or at a minimum a critical mass of physicians if notall are willing, whose e�ort and contribution is required toachieve the goals of the clinically integrated venture. In thisregard, the antitrust consideration is similar in that a keyissue is whether the venture has the minimum scale neededto achieve its goals and whether adding more memberswould make it even more e�cient. As noted above, however,where the venture might have so many participants that itmay have market power, antitrust concerns arise and theincreased e�ciencies will need to be weighed against thegreater potential for anticompetitive e�ects.

In contrast, fraud and abuse regulators look most favor-ably on ventures that o�er the entire group of relevant physi-cians the opportunity to participate. The OIG would viewthe selection of any subgroup of relevant physicians assuspect, potentially indicating that the opportunity had beeno�ered only to those expected to generate large numbers ofreferrals (or as a reward for having done so). In this context,the OIG does not view the absence of referral requirements,or even explicit acknowledgement that referrals are not acondition of collaboration, as dispositive of whether the op-portunity is in fact tied to the referral of patients amongparticipants in a venture. Experience has taught the OIG tolook beyond the form and controlling documents of a trans-action to examine whether its putative structure re�ects thereality of its implementation.

That said, the OIG (in an AKS context) and CMS (in aStark law context) have each recognized that participation

2Health Care Policy Statements, Statement 8.


281

in certain hospital-related activities can, and in some casesmust, be conditioned on performing a minimum number ofprocedures, and thus admitting a certain number of patients,to the hospital. For example, a hospital that is recruiting aphysician to work at the hospital may require a practitionerto have and maintain sta� privileges, but may not conditionrecruitment payments on aggregate admissions by thephysician.3 Moreover, in the context of sta� privileges gener-ally, CMS recognizes that a “hospital may impose conditionsintended to ensure quality of patient care such as requiringthat a physician have [sic] performed a minimum number ofa particular type of care before performing the procedure atthe hospital.”4 In the shared savings programs that the OIGhas reviewed through the advisory opinion process, hospitals'limited program participation (and thus payment) to physi-cians already on the hospitals' medical sta� for at least oneyear at the time the program began. In the OIG's view, thatlimitation actually reduced the likelihood that the arrange-ments would be used to attract referring physicians or toincrease referrals from existing practices.

Like selection criteria, the bases for disciplining andeliminating nonproductive or low-quality physicians from acollaborative venture can be viewed somewhat di�erentlydepending on whether the analysis is under an antitrust orfraud and abuse context. As noted above, antitrust enforcershave found that clinical integration programs that utilizeenforcement mechanisms to maintain a provider group thatis actively engaged and integrated in the clinical protocolstends to demonstrate real indicia of integration.5 Venturesthat discipline or eliminate repeatedly poor performers dem-onstrate that the program is not a mere sham through whichcompetitors are coordinating their activities. The fraud andabuse laws, in contrast, view skeptically e�orts to separateout “non-productive” physicians—at least to the extent “non-productive” is interpreted as physicians who fail to have ahigh volume of patients. Clearly, physicians investors shouldnot be actively encouraged to make referrals to the joint

364 Fed. Reg. 63518, 63543 (Nov. 19, 1999).464 Fed. Reg. at 63543–44.5See MedSouth Opinion.


282

venture nor encouraged to divest their ownership interest ifthey fail to sustain an “acceptable” level of referrals.6

In summary, selection criteria and panel size can beimportant considerations for both the antitrust and fraudand abuse analyses. Under both assessments, it is importantthat whatever criteria are used be carefully and objectivelybased and are developed and applied with the goal of achiev-ing the legitimate aims of the venture. Under the antitrustlaws, selectivity with respect to membership and enforce-ment of selection criteria through expulsion of noncompliantmembers are important to establishing that the venture ismore than a vehicle for competitor collaborations. Under thefraud and abuse laws, a particular focus is on ensuring thatparticipants are not chosen or excluded based on theirpotential or history for making referrals. Given that concern,very broad participation by an existing medical sta� is actu-ally favored under the fraud and abuse analysis, while o�er-ing participation to individuals not currently on sta� isdisfavored. In contrast, under the antitrust laws, very broadparticipation so that a high share of the physicians in one ormore specialties in the geographic market could raise mar-ket power concerns. Note that with respect to this last issue,the focus is the share within a properly de�ned relevantmarket. Thus, a venture that included a very large share ofall of the physicians on a hospital's medical sta� is unlikelyto have market power if there are other nearby physicianswho are not on the medical sta� and who do not participatein the collaboration.

§ 7:14 Incentives

In order to achieve the level of commitment among physi-cians necessary to achieve real e�ciencies, many programs,in addition to providing feedback to members on their per-formance, implement �nancial incentives to reward provid-ers for meeting or exceeding performance benchmarks orpenalize those with poor performance. Indeed, some observ-ers believe that �nancial incentives are essential to a suc-

6See Gainsharing SAB.


283

cessful clinical integration program.1 However, �nancialincentives paid to physicians can raise some of the most seri-ous questions under the fraud and abuse laws. In this sec-tion, we �rst address the antitrust issued raised by �nancialincentives, and then turn to the fraud and abuse consider-ations.

§ 7:15 Incentives—Establishing appropriateincentives under antitrust law—Basis andamount of the �nancial incentives

Many clinically integrated programs choose to provide�nancial incentives to reward and penalize physicians fortheir performance. The �rst step in developing such an incen-tive program is to determine what will be measured, whatwill be the targeted benchmarks for evaluation, and whatadjustments will be made periodically to re�ect performanceimprovements or shortfalls. There are various approaches todetermining what will be measured. Some programs mayconcentrate on the process of care—to what extent do provid-ers follow certain protocols that have been linked to betteroutcomes or lower costs. Other programs may focus on actualoutcomes in the form of actual cost savings or clinicaloutcomes. Both approaches have bene�ts and drawbacks. Al-though outcome-based measures have the advantage of go-ing directly to the core goals of the program, it is often dif-�cult to tease out how much of the ultimate outcome shouldbe attributed to a provider's e�orts, and how much is due toother variables which are beyond a provider's control. Undereither approach, a physician's performance can be measuredagainst a variety of benchmarks including his own historicalachievements, those of others in the collaboration, or local ornational performance levels.

In an antitrust analysis of a clinical integration arrange-ment, enforcers will be interested in whether and how theventure enables the providers to achieve signi�cante�ciencies. Such e�ciencies may be re�ected in improvedscores relative to certain objective criteria, and the �nancialincentives typically will be structured around those criteria.The enforcers are likely to consider several questions: Are

[Section 7:14]1See Rosch Remarks.


284

the goals the venture is seeking to accomplish meaningful interms of seeking to achieve substantial quality improve-ments and/or cost savings? Is the venture able to measurereliably whether the providers are making progress in reach-ing these goals, and is such progress greater than they wouldlike have achieved without the venture? Are the rewardsand penalties in the program su�cient in size to adequatelymotivate the participating providers? Are the �nancial incen-tives structured in a way that rewards may be earned byproviders if they improve their performance, but those whodid not improve will not be rewarded or might even be penal-ized?

When third-party payors establish �nancial incentive ar-rangements such as P4P and similar programs to reward orpunish providers based on their performance, it is reason-able to presume that the goals, measurements, and structureof the program have been designed by the payor to ac-complish goals that it believes are important. Presumablythe payor would have little reason to develop the program ifthat was not the case. Of course, that does not mean thatthe program will succeed, but absent extraordinary circum-stances there would be little reason to assume that the e�ortwas not a genuine attempt to achieve meaningful results. Inassessing a clinically integrated arrangement, however, theantitrust enforcers may be more skeptical about the sub-stance of a program because it is developed not at arm'slength by a payor, but rather by the providers themselves.

One way a venture can address this issue is by trying towork with health plans and employers (essentially theventure's “customers”) to determine what kinds of improvedperformance they believe is most important, and whichwould they value most. In some situations, it may be dif-�cult for a provider venture to work with health plans insuch a way. In such cases, the venture should be able toexplain why it chose the goals, targets, benchmarks, andincentive structure that it did. For example, if the collabora-tion seeks to improve its performance on a set of clinicalmeasures that have been broadly viewed as important, andcan show that its approach is objectively sound and reliable,this should satisfy enforcers that there is a genuine e�ort toobtain signi�cant e�ciencies.


285

§ 7:16 Incentives—Establishing appropriateincentives under antitrust law—Individual vs.group performance

Financial incentives can be based on individual or groupperformance, or a combination of the two. Individual-basedincentives provide a reward (or assess a penalty) to aprovider based on his or own individual performance. Incontrast, group-based incentives depend on how well anentire group of providers perform. Whether a participatingphysician prefers an individual or group-based incentivemay depend on the level of his performance and his commit-ment to the program. High-scoring individuals, or those whoare very committed to improving their own scores, may favorindividual-based performance incentives because they maybelieve they have more control over their scores and aremore likely to be rewarded. On the other hand, group-basedperformance incentives encourage the kind of interdepen-dence that is the hallmark of an integrated joint venture. IfDr. Jones knows that his potential incentive payments isbased on how well Dr. Smith, Dr. Green, and his other col-leagues perform, he will do more to work with them toimprove their scores.

Closely related to the question of individual or group per-formance is where the source of the reward money comesfrom and to whom any penalty amounts are forfeited. Forexample, one alternative is for the group to withhold acertain amount, and to distribute the withholding so that es-sentially the poor performers are penalized by payingbonuses to the high performers. While this approach has theadvantage of simplicity, it may actually provide a disincen-tive for participants to work together to improve their collec-tive performance: Dr. Jones may believe he is more likely toearn a bonus if Dr. Smith does poorly. An alternative is tohave a shared risk pool, so that if the group as a whole doespoorly, it essentially forfeits money collectively to a thirdparty. This is the kind of arrangement that was more com-mon in the 1990s when payors entered into various risk ar-rangements with IPAs and other provider groups. Currently,fewer payors are interested in such arrangements, soprovider ventures need to be more creative if they wish toemulate a shared-risk situation. Some ventures have decidedto donate “forfeited” funds to charities. Others keep themoney in escrow for distribution at a future time. Some


286

health systems that include multiple hospitals have some ofthe forfeited funds from one PHO put into a common fund sothat they might be distributed to another PHO the followingyear.

From an antitrust perspective, rewards and penalties thatare based on shared-risk and group performance are favoredsince they provide the greatest incentive for the kind ofinterdependence and e�ciencies that a joint venture mightachieve compared to that of individual e�orts. Indeed, this isthe basis in the Health Care Policy Statements for recogniz-ing that entities that have �nancial risk-sharing throughcapitation or substantial withholds warrant rule of reasontreatment. But individual-based incentive programs also canbe a means by which a group can encourage desired perfor-mance and achieve signi�cant e�ciencies and thus are alsorelevant to the antitrust analysis.

§ 7:17 Incentives—Establishing appropriateincentives under fraud and abuse laws—Shared savings (“gainsharing”) advisoryopinions

At the time of this writing, there is a great deal ofuncertainty under fraud and abuse laws regarding whattypes of shared savings or incentive payment arrangements(�owing from hospitals to physicians) may be utilized as partof a clinically integrated model. This uncertainty stems, inlarge part, from a tortured regulatory history in this area.

In 2001, only a few years after the OIG released its SpecialAdvisory Bulletin concerning hospital-physician gainsharing,the OIG began issuing advisory opinions, exercising itsenforcement discretion and permitting certain arrangementsto go forward. In those opinions, the OIG concluded that theproposed arrangements incorporated su�cient safeguardssuch that, while the arrangements implicated the CMP lawand could potentially generate prohibited remunerationunder the AKS (if the requisite intent was present), the OIGwould not impose administrative sanctions because, asstructured, the arrangements posed a low risk of programabuse.

The �rst opinion involved a gainsharing arrangementthrough which a group of cardiovascular surgeons couldshare a percentage of the hospital's cost savings arising from


287

the group's implementation of a number of separate costsaving procedures. These procedures, designed to “curb theinappropriate use or waste of medical supplies,” involved: (1)not opening surgical packs until needed; (2) standardize usewith respect to certain high cost, clinical preference devices,and substituting less costly items to be used by the surgeons;and (3) limited use of an antihemorrhaging drug to onlyhigh-risk patients as recommended by independent, objec-tive clinical standards.1 “Properly structured, cost sharingarrangements can serve legitimate business and medicalpurposes,” the OIG wrote. However, the OIG stated that, thesame arrangements can “potentially in�uence physicianjudgment to the detriment of patient care” and said that it isconcerned about: “(i) stinting on patient care; (ii) ‘cherrypicking’ healthy patients and steering sicker (and morecostly) patients to hospitals that do not o�er such arrange-ments; (iii) payments in exchange for patient referrals; and(iv) unfair competition (a ‘race to the bottom’) amonghospitals o�ering cost savings programs to foster physicianloyalty and to attract more referrals.”2

Similar programs were proposed in 2005,3 in 2006,4 and in2007.5 Of the �rst 10 programs, six involved cardiac surgery,three involved cardiology, and one involved anesthesiologyservices provided during cardiac surgery. Unfortunately,these programs are all fairly similar,6 and so the opinionshave provided only a limited, and somewhat formulaic, ap-proach aligning hospital and physician incentives.

In general, the opinions consistently identify eight pri-mary features as contributing to its decision not to imposecivil monetary penalties:

[Section 7:17]1OIG Advisory Op. 01-1, at 3 (Jan. 18, 2001).2OIG Advisory Op. 01-1, at 6.3OIG Advisory Op. 05-01 (Jan. 28, 2005); OIG Advisory Op. 05-02

(Feb. 10, 2005); OIG Advisory Op. 05-03 (Feb. 10, 2005); OIG Advisory Op.05-04 (Feb 10, 2005); OIG Advisory Op. 05-05 (Feb. 18, 2005); OIGAdvisory Op. 05-06 (Feb. 18, 2005).

4OIG Advisory Op. 06-22 (Nov. 9, 2006).5OIG Advisory Op. 07-21 (Dec. 28, 2008).6This is likely due, in no small measure, to the fact that all the

programs were developed, and submitted for Advisory Opinion approval,by the same health care consulting company.


288

1. The arrangements are transparent, which allow for“public scrutiny and individual physician accountabilityfor any adverse e�ects of the arrangement,” includingdi�erence in treatment based on nonclinical indicators.

2. There is “credible medical support for the position thatimplementation of the recommendations [to reducecosts] will not adversely a�ect patient care” and the ar-rangements are “periodically reviewed . . . to con�rmthat the arrangement was not having an adverse impacton clinical care.”

3. The amount to be paid is “calculated based on allsurgeries regardless of the patients' insurance cover-age,” with a cap on payment for federal health careprocedures. Procedures subject to the arrangement arenot “disproportionately performed on federal healthcare” bene�ciaries. These cost savings are “calculatedfrom the hospital's actual out-of-pocket acquisitioncosts.”

4. The arrangements utilize “objective historical and clini-cal measures to establish” a �oor beyond which no sav-ings accrued to the group. Baseline measures are re-lated to “comparable hospitals' practices and patientpopulations.”

5. Inappropriate reductions in services are prevented inproduct standardization recommendations by “ensuringthat individual physicians still had available the sameselection of devices and supplies after implementationof the arrangement as before.”

6. Patients possibly a�ected by the arrangements aregiven written disclosures of physicians' involvement inthe arrangement with an opportunity to review the ar-rangement prior to admission (or, where preadmissionconsent is impracticable, prior to consenting to surgery).The OIG notes that disclosure is insu�cient standingalone, but this practice helps prevent abuses of patienttrust.

7. The �nancial incentives under the arrangement arereasonably limited in duration and amount.

8. Pro�ts from arrangements are distributed to individualphysicians on a “per capita basis” (thereby reducing theincentive for an individual physician to generatedisproportionate cost savings).


289

Similarly, the OIG found that su�cient safeguards existedto mitigate kickback risks.7 Among the safeguards identi�edby the OIG were that the arrangements contained provisionsthat reduced the likelihood that they were being used to at-tract referring physicians or to increase referrals from exist-ing physicians and were structured to eliminate the risk ofrewards for referrals to the group practices.

Finally, the OIG identi�es certain factors it believes would“heighten” the risk that payments will lead to inappropriatereductions of services. The risk would be heightened where:

1. There is no demonstrable direct connection between aphysician's activity and any reduction in the hospital'sout-of-pocket costs (and any corresponding “gainshar-ing” payment).

2. The performance that would give rise to the savings isnot identi�ed with speci�city.

3. There are not su�cient safeguards to protect againstthe risk that other unidenti�ed actions, such as prema-ture hospital discharges, might actually account for any“savings.”

4. The quality of care indicators are of questionable valid-ity and statistical signi�cance.

5. There is no independent veri�cation of cost savings,quality of care indicators, or other essential aspects ofthe arrangement.

More recently, the OIG has approved arrangements involv-ing an anesthesiology group,8 an arrangement between ahospital and orthopedic surgeons and neurosurgeons,9 aswell as arrangements that are three years10 and two years11

instead of the one year that had been in place for the previ-ous proposals.

7The OIG noted that the arrangements potentially met the AKS safeharbor for personal services and management, but that they ultimatelydid not �t because the payments under the arrangements to the grouppractices were calculated on a percentage basis, and as such were not setin advance, as required to meet the safe harbor.

8OIG Advisory Op. 07-22 (Dec. 28, 2007).9OIG Advisory Op. 08-09 (July 31, 2008).

10OIG Advisory Op. 08-15 (Oct. 6, 2008). For purposes of calculatingthe payment to each cardiology group, the actual costs incurred for theitems speci�ed in the recommendations when used by cardiologists in theparticular cardiology group during the speci�ed procedures (the “currentyear costs”) are subtracted from the costs for the same items when used


290

§ 7:18 Incentives—Establishing appropriateincentives under fraud and abuse laws—Incentive payment advisory opinion

The most signi�cant change for the OIG, however, is anadvisory opinion from October 2008 that addressed not againsharing/shared saving program but an incentive pay-ment proposal wherein a hospital proposed to share with aphysician-owned entity certain performance-based compen-sation available to the hospital under a quality and e�-ciency agreement with a private insurer.1 As in the advisoryopinions that address shared saving programs, the OIGdetermined that this incentive payment proposal couldviolate the CMP law as well as the AKS; however, the OIGfurther determined that it would not penalize the hospital inconnection with the arrangement because of safeguardsagainst fraud and abuse in the proposed arrangement.

Under the arrangement, a hospital participated in a P4Pprogram implemented by a private insurer. The programdictated that, in addition to the amount the private insurerwould pay the hospital for the care of patients in a givenyear, the insurer would also pay the hospital a percentagebased on the extent to which the hospital met certain qual-ity and e�ciency standards. These quality of care standardsrequired data reporting and the achievement of qualitytargets described in the Speci�cations Manual for NationalHospital Quality Measures published by the Joint Commis-sion; these are quality measures that have been derived froma joint e�ort of CMS and the Joint Commission.2 The incen-tive payment arrangement between the insurance companyand the hospital would require that the hospital track qual-ity measures and outcomes for all patients, not just enrolleesof the insurer's plans. The e�orts of the hospital and its sta�to achieve the quality care standards would thus a�ect

during comparable procedures in the respective base year (the “base yearcosts”).

11OIG Advisory Op. 08-21 (Nov. 25, 2008).

[Section 7:18]1OIG Advisory Op. 08-16 (Oct. 7, 2008).2See The Joint Commission, Speci�cations Manual for National

Hospital Quality Measures (2008), at http://www.jointcommission.org/PerformanceMeasurement/Current+NHOM+Manual.html.


291

Medicare and Medicaid bene�ciaries. As such, the arrange-ment could violate the CMP law and could violate the AKS ifit were used by the hospital to induce doctors for referrals.According to the requestor of the advisory opinion, thehospital could not achieve these quality targets without theassistance and cooperation of its medical sta�.

The physician entity in the arrangement would be com-posed of physician members of the hospital's active medicalsta�. Each physician who joined the entity would make anequal capital contribution which, in the aggregate, wouldprovide for working capital needed for the entity. The entitywould need at least 10 participating physicians before goingforward. Under a three-year quality enhancement profes-sional services agreement between the hospital and thephysician entity, the entity would require its members toundertake various tasks to ensure that the quality targetswere achieved, including developing policies and procedures,conducting peer review, and auditing medical records. Thehospital would pay the physician entity as much as 50% ofits incentive payments to the doctors for their e�orts in help-ing the hospital meet quality targets established by theinsurer. The entity would then distribute the payouts to themember doctors on a per-capita basis.

Although the CMP law would apply, the OIG identi�edthe safeguards in the proposed arrangement, including: (1)credible medical support for the likelihood the proposed ar-rangement would improve patient care and was unlikely tohave adverse e�ects on patient care; accordingly, the “physi-cians are to be compensated for speci�c actions which havebeen recognized as improving patient care”; (2) the lack ofan incentive (i.e., no reduction in compensation) for a physi-cian to apply a speci�c standard in medically inappropriatecircumstances; (3) a reasonable relationship between thequality targets to the practice and patient population of thehospital; (4) transparency, through noti�cation to patientsa�ected by any of the performance measures; and (5) moni-toring of the quality targets “to protect against inappropri-ate or limitations in patient care or services.”

The safeguards that helped mitigate the OIG's AKS-related concerns were: (1) the structure of the program (e.g.,limiting it to physicians who have been on active medicalsta� for at least a year and a cap on compensation) wouldnot likely attract referring physicians to the hospital or


292

increase referrals from doctors already on the hospital's sta�;(2) per-capita distributions which would reduce the risk thatthe arrangement would be used to reward individual physi-cians who refer patients to the hospital; and that participa-tion would not be limited to any particular group of high-referring physicians, but could include all physicians, subjectto the requirement that they have been on the active medi-cal sta� for at least a year; (3) transparency, which wouldensure that the purpose of the program would be to improvequality rather than to reward referrals; and (4) the oversightrole of the private insurer, whom the OIG viewed as havingno incentive to overcompensate either the hospital directlyor the physicians indirectly and having “every incentive” topay the portion of the bonus only as earned through theachievement of the quality targets. Acknowledging thatachieving the quality targets is an important part of theprogram, a �nal mitigating factor noted by the OIG is thatthe hospital had “certi�ed that it is not feasible to achievethem without the assistance of its medical sta�.”

§ 7:19 Incentives—Establishing appropriateincentives under fraud and abuse laws—Starklaw exception for hospital shared savings andincentive payments

In 2004, CMS published a discussion of the application ofthe Stark law to payments hospitals might make to createincentives for physicians to assist the hospital in achievingcertain quality and e�ciency goals. Speci�cally, in the pre-amble to the Stark II, Phase II, Interim Final Rules, CMSaddressed comments requesting that the Stark law's employ-ment exception be extended to permit hospitals to pay suchincentives to employed physicians. The comments arguedthat such incentive payments should not be construed as be-ing based on the volume or value of referrals for purposes ofthe Stark law. CMS replied:

There is no exception in the statute or in these regulationsthat would permit payments to physicians based on theirutilization of DHS, except as speci�cally permitted by the risk-sharing arrangements, prepaid plans, and personal service ar-rangements exceptions. None of those exceptions permit thosepayments other than in the context of services provided toenrollees of certain health plans. We believe that the Congressintended to limit these kinds of incentives consistent with the


293

civil monetary penalty provision at section 1128A(b)(1) of theAct that prohibits a hospital from paying physicians to reduceor limit care to hospital patients. Given that prohibition, wecannot say that payments based on lowering utilization pres-ent no risk of fraud or abuse.1

However, then in April 2008, CMS advised providers thatit was considering adding a speci�c Stark law exception forcertain gainsharing arrangements and sought comments onhow to circumscribe the exception to avoid abuse.2 CMS thenissued a proposed regulatory exception to the physician self-referral law that would allow hospitals to pay incentive pay-ments to physicians who participate in quality improvementand shared savings programs.3 CMS ultimately declined to�nalize its proposed exception to the Stark law that wouldhave expressly permitted either incentive payment programs,such as pay-for-performance arrangements, or that allowedphysicians to share in the cost savings attributed to their ef-forts (i.e., “shared savings programs,” such as gainsharing).This exception arguably was needed to ensure that suchprograms did not run afoul of the self-referral restrictions ofthe Stark law.

Rather than �nalize its proposal, however, CMS extendedthe comment period for 90 days and requested additionalinput from the industry on a number of topics relevant tosuch an exception. In doing so, CMS conceded that it has“had limited opportunity to review incentive payment andshared savings programs for compliance with the physicianself-referral law, and [it] lacked familiarity with the speci�csof measuring achievements and calculating payments undersuch programs.” Through the solicitation of comments, CMShopes to obtain the information it needs to �nalize a newStark law exception that will achieve transparency and ac-countability and ensure quality of care, while guardingagainst the risks of disguised payments for referrals,participants cherry-picking healthy patients and steeringsicker patients to other hospitals, and, for shared savings ar-rangements, physicians not using cheaper devices, tests, or

[Section 7:19]169 Fed. Reg. 16054, 16088 (Mar. 26, 2004).273 Fed. Reg. 23528, 23692 (April 30, 2008).373 Fed. Reg. at 38502.


294

treatments but which are of inferior quality and dischargingpatients quicker than clinically appropriate.

The best that can be said of recent fraud and abusepronouncements on gainsharing and incentive payments isthat CMS and the OIG recognize that the health caremarketplace is developing faster and in more directions thanthe current regulatory scheme can accommodate and thatsome of those developments are at risk of being unnecessar-ily constrained. The history of the OIG's treatment of theCMP law and AKS, through the advisory opinion process,and of CMS's evolving position on these types of clinicallyintegrated structures, reveals a growing consensus both onthe value of these programs and a recognition that the lawshould allow arrangements that, while establishing somesafeguards against sham arrangements, provide the �ex-ibility to deal with innovative ways of achieving clinicalintegration through hospital-physician models. That said,other than a few tightly crafted advisory opinions, fraud andabuse guidance o�ers only uncertain comfort to those whowould develop innovative collaborations. Again, the �ex-ibility of the antitrust rule of reason analysis—an analysisthat recognizes value in innovation itself—if properly ap-plied, has been shown to be better suited to accommodatingnew forms of clinical integration than the regulatory ap-proach under the fraud and abuse statutes.

§ 7:20 Incentives—Putting it all togetherAs discussed above, there are several aspects of �nancial

incentives for which the antitrust considerations will bealigned with those assessed under the fraud and abuse laws.For example, both favor (a) incentives that encourage clini-cal protocols or practice patterns based on credible medicalsupport that they will reduce costs without adversely a�ect-ing care or which might improve quality; (b) the use of ap-propriate benchmarks involving appropriate comparables; (c)the use of quality of care indicators that are valid andstatistically signi�cant; and (d) independent veri�cation ofcost savings, quality of care indicators and other essentialaspects of the arrangement.

A number of factors that fraud and abuse regulatorsconsider favorably in assessing incentives generally wouldbe viewed neutrally under the antitrust laws. These factorsinclude:


295

E payments calculated based on all performance regard-less of insurance coverage;

E limits on the duration and amount of incentives;E distribution of funds on a per-capita basis;E ensuring provider access to devices and supplies that

are available outside of the program; andE transparency of arrangements so that there can be pub-

lic scrutiny and individual accountability for any pos-sible adverse e�ects.

While the existence of these latter factors would not weighagainst an arrangement in an antitrust assessment, theycould have the e�ect of restricting the tools that a collabora-tion has available to it to most e�ectively achieve its goals.For example, under the antitrust laws, there generally wouldnot be strict limits set on the amount or duration of incen-tives if the collaboration thought it would best drive optimumprovider conduct. Similarly, a collaboration might determinethat sharply limiting access by providers to nonapprovedsuppliers would best achieve cost-savings, and barringextraordinary circumstances, it would be free to do so underthe antitrust laws. Although transparency is often desirable,some ventures may consider some of their arrangementsproprietary and not wish to share them with competitors.Perhaps the most serious limitation might be the apparentpreference under the CMP statute for the distribution offunds on a per-capita basis; that might be viewed assubstantially less e�ective at incentivizing desired behaviorthan would be the distribution of funds based on individualperformance or the amount of services provided by thephysician.

In summary, it does not appear that hospital/physiciancollaborations face inconsistent requirements under theantitrust laws and the fraud and abuse laws when theyconsider a �nancial incentive program. There are areas,however, where the fraud and abuse laws, at least asre�ected in the guidance available so far, may restrict therange and structure of some incentives, perhaps limitingtheir overall e�ectiveness.

§ 7:21 ConclusionThe question remains, what is the best strategy for


296

hospitals and physicians that wish to collaborate on improv-ing quality or reducing costs to assure compliance with boththe antitrust laws and the fraud and abuse statutes?

The starting point should be a clear understanding byeveryone involved in the collaboration about what theventure legitimately can—and cannot—seek to accomplish.For example, if members of a PHO believe that their goal isto simply develop some way in which they can collectivelynegotiate with health plans to “level the playing �eld” andachieve the reimbursement levels to which they believe theyare entitled, the collaboration is likely to create ongoingantitrust concerns—particularly if it is successful in achiev-ing its goal absent any quality improvement. The same ap-plies to a venture, for example, which is viewed by hospitaladministrators as a means of rewarding referring physiciansand increasing their admissions. In such circumstances, theventure is bound to raise serious fraud and abuse questions.

Once the legitimate goals are clear, the organizers of thecollaboration should then consider, given their particularneeds and circumstances, what are the best ways to structuretheir arrangement. This will include addressing such issuesas the criteria for participation, what speci�c kinds ofconduct should be encouraged (or discouraged), how tostructure the incentive payments, and how the necessaryinfrastructure will be funded. Although it may be helpful toexamine successful arrangements that others have developedelsewhere or to consider those that have been favorablyreviewed by regulators, these examples should not be viewedas the entire universe of acceptable approaches. Rather, ifthere are compelling reasons for fashioning a collaborationin a novel way that will best achieve legitimate goals, thosepotential arrangements should be fully considered.

We emphasize always keeping an eye on how best to ac-complish the collaboration's legitimate goals for severalreasons. Doing so will maximize the chance that the venturewill succeed from a practical standpoint because it will beable to provide higher-quality or less costly services that arevalued by patients and payors. However, it will also helpincrease the chances that the venture will survive legalscrutiny. As we have discussed above, both the antitrust andthe fraud and abuse enforcers have a threshold concern thata venture may be a “sham” designed simply as a cover toobtain market power or induce referrals. By documenting


297

the rationale for each of the steps it has taken and demon-strating that these were all carefully considered to achievereal results, an entity can help minimize the risk that enforc-ers will �nd it to be a sham. What is more, the venture likelywill run more smoothly if the collaboration is taking stepsthat seem to “make sense” to its own participants so as toachieve its own particular goals, as opposed to trying to fol-low an approach designed simply to pass regulatory revieweven though it may poorly �t the situation at hand.

Notwithstanding the above, the collaboration may face dif-�cult legal questions. For example, the venture may concludethat in order to maximize the commitment and focus of itsparticipants, and thereby increase e�ciencies, it wishes tolimit participation to only those physicians whose practice ispredominantly centered at the hospital. For the samereasons, it may wish to pay members not the same per-capita, but rather based on individual performance or howmany services each physician provides through the venture.Likewise, it may decide that performance will not be mea-sured across all payment sources, but only those who areproviding increased P4P bonuses.

These, and similar questions, can be challenging. However,as we have explained, fundamentally there should be waysto structure a core hospital/physician arrangement that canmeet both antitrust and fraud and abuse considerations.Most of the open questions relate to the extent to which theventure can take steps that are intended to increase e�cien-cies, but which lack some of the safeguards regarding fraudand abuse issues that the OIG or CMS have identi�ed. Onthese issues, the venture, guided by its counsel, will need tocarefully consider the importance of the proposed steps, thepotential for abuse, and whether there are alternative ap-proaches that could produce similar results. In some cases,the collaboration may wish to proceed further only afterobtaining feedback from the enforcers.

As noted above, both the OIG and CMS have been consid-ering an increasing number of physician/hospitalcollaborations. Health care reform may increase the push forsuch collaborations, and with it, the pressure for greaterguidance and �exibility from the enforcers on sucharrangements. If this occurs, we will be able to determinemore conclusively whether such e�orts truly hold thepromise for signi�cant advances in health care quality and


298

e�ciency.


299

hospital-physician collaborations: antitrust and …/media/hogan-lovells/pdf/... · 2017-01-27 ·...

Documents