DISCUSSION AND REVIEW PAPER

The Promise of Accountable Care Organizations: “The Code,” Reimbursement, and an Ethical No-Win Situation for Behavior Analysts

Abraham Graber1 & Matthew O’Brien2

Published online: 25 January 2018 # Association for Behavior Analysis International 2018

Abstract Clinical ethics, with its emphasis on the actions of clinicians, risks overlooking the ways in which broader health-care structures influence the behavior of health-care providers. Analysis of a factual case study demonstrates that status quo reimbursement practices may place behavior analysts in a position where, no matter how they act, they risk acting unethically. By contrast, the reimbursement model set by accountable care organizations (ACOs), part of the Patient Protection and Affordable Care Act (also known as Obamacare), may offer a solution. However, making good on the promise of ACOs will require more resources than any individual behavior analyst possesses. In order to encourage institutional structures that facilitate ethical practice, behavior analysts’ professional organizations should engage in contemporary political discussions about the state of American health care.

Keywords Ethics . Reimbursement . Accountable care organizations . The professional and ethical compliance code for behavior analysts . Affordable care act

With its narrow emphasis on the behavior of clinicians, the literature on ethics in the clinic too often overlooks the ethical import of health-care institutions (Hafferty & Franks, 1994). From the perspective of behavioral psychology, overlooking the environmental contributions to ethically problematic be- havior is particularly worrisome (Skinner, 1971). In this article, we zoom out from the clinic and highlight the ethical signifi- cance of institutional factors. In particular, status quo reim- bursement practices may place behavioral analysts in a posi- tionwhere, nomatter how they act, they risk acting unethically. The reimbursement model set by accountable care organiza- tions (ACOs), part of the Patient Protection and Affordable Care Act (also known as Obamacare), may offer a solution. Our case analysis serves to illustrate a broader point: In order to encourage institutional structures that facilitate ethical practice, behavior analysts should engage in contemporary political dis- cussions about the shape of American health care.

The Case Study

This article grounds its ethical analysis of reimbursement practices in the concrete details of a factual case study. After presenting the case study, an analysis of the status quo reim- bursement practices illuminates how these practices put be- havior analysts in an ethically untenable position.

The case study takes place within the context of Clinic 1, a tertiary care intensive outpatient behavior clinic at a major teaching hospital with a focus on assessment and treatment of severe and challenging behavior. Clients who visit this clin- ic receive behavior–analytic services for 3 h/day for 10 con- secutive business days, totaling an entirety of 30 h of care. Clients referred to this clinic have generally already received care from less intensive top tier behavioral outpatient tertiary care providers but have failed to respond to treatment or are in need of more comprehensive assessment.

The client, Jane Doe, was 9 years 4 months old at the time of the appointment. At the age of 3 months, Jane Doe was in a car accident that caused a skull fracture, intracranial bleeding, and brain injury. Jane Doe had mild cerebral palsy, intellectual disability, and a history of epilepsy (whichwas resolved by the time of the appointment).

Jane’s problem behavior began approximately four years prior to the appointment. At the time of the appointment, she displayed three topographies of self-injurious behavior: head

* Abraham Graber [email protected]

1 Department of Philosophy and Classics, University of Texas at San Antonio, 1 UTSA Circle, San Antonio, TX 78249, USA

2 Center for Disabilities and Development, University of Iowa Children’s Hospital, 100 Hawkins Dr., Iowa City, IA 52252, USA

Behavior Analysis in Practice (2019) 12:247–254 https://doi.org/10.1007/s40617-018-0209-0

banging (forward, backward, side), head hitting (contacting head with wrist, fist, or hand), and eye poking (poking or rubbing eyes with fingers). Each of these behaviors resulted in tissue damage.

Jane Doe received 30 h of assessment and treatment at Clinic 1, broken up into ten 3-h sessions. Clinic 1 billed each session as a psychotherapy session lasting longer than 53 min (Current Procedural Terminology [CPT] Code 90837). This code reimburses a flat rate for any care provided at or above 53 min. Thus, for every 3-h session, Clinic 1 was reimbursed the same as if it had only provided 53 min of therapy. Clinic 1 provided a total of 1,800 min of care and was reimbursed as if 530 min of care had been provided. Consequently, a total of 1,270 min of care could be regarded as nonreimbursable time. Thus, Clinic 1 was reimbursed for only 29.44% of the care provided—a full 70.56% of care was not reimbursed.

Furthermore, because Clinic 1 specializes in the treatment of individuals with severe problem behavior, appointments at Clinic 1 involve up to four clinicians. However, the psycho- therapy code under which Clinic 1 bills was initially designed for individual psychotherapy involving a single clinician, a pat ient , and/or a family member (APA Pract ice Organization, 2013). Not only is Clinic 1 reimbursed for a mere 29.44% of the care provided, but the CPT code Clinic 1 uses is also designed to cover a single clinician even though Clinic 1 requires up to four clinicians.

It is important to note that Jane Doe is not an unusual case for Clinic 1. Clinic 1 faces a similar billing challenge for all the patients it sees. As an intensive tertiary care outpatient clin- ic—and the only one of its type in the state—Clinic 1 sees the most challenging cases in the state. Concomitant with both the severity of the challenging behavior displayed by patients at Clinic 1 and the lack of success of previous behavioral inter- ventions, Clinic 1 provides 30 h of assessment and treatment for all of its patients. As a general rule, Clinic 1 can only bill using the CPT psychotherapy code for 53+ min, leaving the majority of the care provided by Clinic 1 unreimbursed.

Applied Behavior Analysis International (ABAI) has long been aware of the kind of reimbursement problems faced by Clinic 1. In response to this type of concern, ABAI has worked to develop adaptive behavior assessment codes (ABACs). ABACs were designed with behavioral interven- tion in mind and allow clinicians to bill for comparatively lengthy periods of time (Shain, n.d.). Interventions that qualify as “adaptive behavior treatment by protocol” can be billed using one code for the first 30 min (i.e., 0364T) and have a distinct code that can be used to bill for additional 30-min increments (i.e., 0365T; Shain, n.d.). ABACs are currently classified as Category III CPT codes (Thompson, n.d.). Category III CPT codes are temporary; after a 5-year period, they will either become Category I CPT codes or they will be eliminated (Thompson, n.d.). Furthermore, whereas Category III CPT codes do not have standardized nationwide

reimbursement rates, Category I CPTcodes do. Designed spe- cifically for behavior analysts, ABACs hold significant prom- ise for resolving the financial difficulties faced by Clinic 1.

Despite efforts on the part of the hospital administration, insurers continue to be unwilling to reimburse any care pro- vided by Clinic 1 using ABACs. Even in states where insurers allow providers to bill using ABACs, an autism spectrum diagnosis is frequently required (see, e.g., ValueOptions, n.d.), and there are often age restrictions that rule out a large number of patients (see, e.g., Arkansas BlueCross BlueShield, 2015; BlueCross BlueShield of Louisiana, 2017). Even if Clinic 1 could bill using ABACs, these codes would likely only cover a small percentage of the patients the clinic treats.

To some extent, the difficulties Clinic 1 faces with billing using ABACs may be resolved when ABACs are upgraded to Category I CPTs. Unfortunately, as documented in a recent e- mail to ABAI members, ABACs have recently encountered a further hurdle. Medically unlikely edits (MUEs) indicate “the maximum number of units of service that a provider would report under most circumstances for a single beneficiary on a single date of service” (J. W. Minton, T. Thompson, G. Green, J. Carr, & M. Wasmer, personal communication, April 26, 2017). ABAI wrote: “Last year, we learned that CMS had established overly restrictive MUEs for the Category III CPT codes for adaptive behavior services” (J. W. Minton, T. Thompson, G. Green, J. Carr, & M. Wasmer, personal com- munication, April 26, 2017). Although ABAI has had some success rolling back these overly restrictive MUEs, these re- strictions remain in place for “three codes1 [that] are critical to the effective delivery of ABA services” (J. W. Minton, T. Thompson, G. Green, J. Carr, & M. Wasmer, personal com- munication, April 26, 2017). Although ABACs hold promise for Clinic 1, significant obstacles must still be overcome be- fore this promise can be realized.

Ethical Analysis of Current Billing Practices

Is This an Ethical Dilemma?

Themajority of this section will be dedicated to demonstrating that Clinic 1 is in an ethical no-win situation; no matter how Clinic 1 proceeds, it risks acting in a way that is ethically impermissible. There are two distinct ways to conceptualize this kind of ethical no-win situation. An ethical no-win situa- tion may constitute an ethical dilemma. An ethical dilemma

1 The three codes that would regularly be used in Clinic 1 are 0365T, adaptive behavior treatment by protocol administered by a techni-

cian; 0369T, adaptive behavior treatment with protocol modification adminis- tered by a physician or other qualified health care professional (QHCP); and 0370T, family adaptive behavior treatment guidance administered by a physi- cian or QHCP. (J. W. Minton, T. Thompson, G. Green, J. Carr, & M. Wasmer, personal communication, April 26, 2017)

248 Behav Analysis Practice (2019) 12:247–254

exists when “no matter what … [an agent] does, she will do something wrong” (McConnell, 2014). There is, however, significant disagreement at the level of ethical theory regard- ing whether moral dilemmas can exist (McConnell, 2014).

From an alternative point of view, in an ethical no-win situation, although there are notable ethical reasons for avoiding all of the available options, there is nonetheless an ethically permissible course of action. Understood this way, although in an ethical no-win situation there is still an ethically best course of action (i.e., a permissible action), the ethically best option is best only because it is being compared to even worse choices. Although there may be an ethically best choice in an ethical no-win situation, ethical no-win situations are nonetheless problematic because, although there may be an ethically best choice, there are no ethically good choices.

Does the case of Jane Doe constitute an ethical dilemma? Answering this question goes beyond the scope of this inqui- ry. There are notable moral reasons for avoiding any of the available options in the case of Jane Doe. Consequently, the case of Jane Doe constitutes an ethical no-win situation. Nonetheless, there may (or may not) be an ethically best decision.

If there may be an ethically best option, why is the empha- sis not on determining which option is ethically best? As noted at the outset, ethical examinations of clinical encounters too often limit themselves to considering the proximate clinical environment. In so doing, ethicists often fail to consider broader questions about the ethical implications of institution- al structure (cf. Hafferty & Franks, 1994). Although it may be important to offer ethical guidance to clinicians regarding con- crete clinical encounters, it is equally or more important to draw attention to institutional structures that give rise to ethi- cal problems in the clinical context.

Ethical Analysis of Options for Clinic 1

Option 1: Close the clinic Clinic 1 is reimbursed for less than 30% of the care it provides. Clinic 1’s ongoing efforts to find alternative billing options, bolstered by support from hospital administration, have been unsuccessful. In this section, we will offer an ethical analysis of the options available to Clinic 1. Where applicable, we will further apply the Professional and Ethical Compliance Code for Behavior Analysts (PECC) and the ethical guidelines of the Association of Professional Behavior Analysts (APBA EG) to the case study.

The heart of the ethical challenge lies in the financial diffi- culty faced by Clinic 1. Any clinic that is only reimbursed for 30% of the care it provides will have difficulty covering its costs. In response to financial pressures, Clinic 1 could close. Therapists at Clinic 1 could then reallocate their time to clinics where therapists are reimbursed for a greater percentage of the care provided.

Although closing Clinic 1 would resolve relevant financial questions, it would also deprive patients of needed services. Clinic 1 is the only clinic of its type in the state. Were Clinic 1 to close, families would have to travel hundreds of extra miles to receive recommended services. For many of the families served by Clinic 1, such travel is an economic impossibility. Consequently, Clinic 1’s closing is tantamount to making re- quired behavioral interventions inaccessible to families in need. There are thus notable ethical considerations militating against the closing of Clinic 1.

A number of provisions in the PECC and APBA EG are relevant. Both codes hold that “the behavior analyst’s behav- ior conforms to the legal and moral codes of the social and professional community of which the behavior analyst is a member” (Association of Professional Behavior Analysts, 2010; cf. Behavior Analyst Certification Board, 2017). Were Clinic 1 to close, many profoundly vulnerable individuals— all with severe challenging behavior and most with significant disabilities—would no longer be able to receive the care they need. This is an ethically problematic result; as behavior ana- lysts are bound by the ethical norms of the society in which they practice, therapists at Clinic 1 must consider the impact closing the clinic would have on families needing the services that only Clinic 1 provides.

The PECC requires that “behavior analysts act in the best interests of the client and supervisee to avoid interruption or disruption of service” (Behavior Analyst Certification Board, 2017, p. 10), and the APBA EG holds that, “behavior analysts provide for orderly and appropriate resolution of responsibil- ity for client care … with paramount consideration given to the welfare of the client” (Association of Professional Behavior Analysts, 2010). Neither prescription is immediately relevant regarding the decision to close Clinic 1. Only indi- viduals to whom “behavior analysts provide services” count as clients (Behavior Analyst Certification Board, 2017, p. 6; cf. Association of Professional Behavior Analysts, 2010). Consequently, neither clause establishes an ethical obligation on the part of therapists at Clinic 1 to individuals who need, but have not yet received, services fromClinic 1. Nonetheless, both clauses capture the ethical importance of providing con- sistent and ongoing care to individuals in need. Although closing Clinic 1 would not violate the letter of either clause, by effectively denying care to individuals in need, closing Clinic 1 would violate the spirit of both.

Option 2: Make Clients Pay Out of Pocket Although there are notable ethical reasons not to close Clinic 1, less drastic mea- sures are available. Rather than closing entirely, Clinic 1 could choose to prioritize patients who can afford to pay any ex- penses not covered by insurance. For example, if a patient’s insurance will only cover 29.44% of the care provided, Clinic 1 could choose to primarily see patients who could pay the remaining 70.56% out of pocket. Clinic 1 could continue to

Behav Analysis Practice (2019) 12:247–254 249

see patients who, like Jane Doe, cannot afford to pay out of pocket, although Clinic 1 would see patients like this with significantly less frequency than is the clinic’s current practice.

As part of a state-governed university hospital, Clinic 1 cannot choose to approach reimbursement in this manner. The clinic adheres to the hospital’s policy that no individual be denied services based on the ability to pay. Although Clinic 1 cannot choose to prioritize clients who can pay out of pock- et, it is nonetheless worth identifying the ethical concerns associated with this reimbursement model, as it is a model available to many behavioral clinics.

There are significant ethical concerns associated with pre- ferring patients who can afford to pay out of pocket. Jane Doe has no control over the family into which she was born, nor does she have control over her family’s economic situation. From an ethical perspective, the economic status of Jane Doe’s family is irrelevant to the extent to which Jane Doe deserves therapy. Similarly, from a clinical perspective, the economic status of Jane Doe’s family is irrelevant to the extent to which Jane Doe needs therapy. Nonetheless, if Clinic 1 prioritizes clients who can afford to pay for services out of pocket, a client’s economic situation will determine his or her access to services. Put bluntly, if Clinic 1 opts to primarily see patients who can afford to pay out of pocket, economically disadvantaged patients like Jane Doe will not receive therapy. Although this would be a troubling result regarding any clinic, given that Clinic 1 is the only clinic of its type in the state, it is uniquely troubling. Clients too economically disadvantaged to receive care at Clinic 1 are also likely to be the same clients who cannot afford to travel hundreds of miles to receive ser- vices elsewhere.

Clauses in both the PECC and the APBA EG suggest that prioritizing patients who can pay out of pocket is ethically impermissible: “In their work-related activities, behavior ana- lysts do not engage in discrimination based on … socioeco- nomic status” (Behavior Analyst Certification Board, 2017, p. 5; cf. Association of Professional Behavior Analysts, 2010). As written, this clause is ambiguous and can be interpreted in two ways. In one reading, “discrimination” is being used in its technical sense, where discrimination is “any difference in responding in the presence of different stimuli” (Catania, 2007, p. 387). If this is the intended reading, showing a pref- erence for clients who can afford to pay out of pocket is clearly forbidden by the PECC and the APBA EG.

In another reading, “discrimination” is being used in its colloquial sense, which Merriam-Webster defines as “the un- just or prejudicial treatment of different categories of people or things” (“Discrimination,” 2018). From this interpretation, it is likely impermissible to prioritize patients who can afford to pay out of pocket. So long as one believes that Jane Doe’s socioeconomic status is irrelevant to the extent that she de- serves treatment, it would be unjust to deny her access to

treatment on the grounds that she cannot afford to pay out of pocket.

Option 3: Provide Less CareClinic 1 could resolve its financial difficulties by opting to provide less care than is deemed nec- essary. For example, because Clinic 1 is not reimbursed for services provided beyond 53 min/day, Clinic 1 could limit care per client to 53 min/day.

Patients are referred to Clinic 1 only after other behavior– analytic tertiary care providers have determined that intensive outpatient care is required. Patients referred to Clinic 1 need more intensive therapy than can be provided in 53-min ap- pointments. Thus, limiting care to 53 min/day would consti- tute knowingly providing care that is unlikely to be effective. Were Clinic 1 to limit therapy to 53 min/day, therapists at Clinic 1 would be charging for care that comes with the risk of countertherapeutic outcomes (Cooper, Heron, & Heward, 2007) while knowing that the therapy is unlikely to be suc- cessful. Such behavior is ethically impermissible; it is doubly harmful to clients, requiring clients to expend resources and exposing them to the risk of countertherapeutic outcomes, and has minimal promise of benefit.

Limiting care to 53 min/day would violate both the PECC and the APBA EG. Both ethics codes hold that “clients have a right to effective treatment” (Behavior Analyst Certification Board, 2017, p. 8). Clients referred to Clinic 1 require therapy that is more intensive than is allowed by 53-min sessions. Were Clinic 1 to only provide care in 53-min intervals, the clinic would fail to respect a client’s right to effective treatment.

Beyond Clinic 1 Clinic 1 is in an ethical no-win situation. Reimbursed for 29.44% of the care it provides, Clinic 1 is financially unsustainable. Were the clinic to close, some of the most vulnerable individuals in the state would be unable to receive the care they need. Were the clinic to only see patients who can pay out of pocket, it would limit clients’ access to needed care on grounds that are both ethically and clinically irrelevant. Finally, if the clinic were to limit the care it provides to patients, therapists at Clinic 1 would knowingly be providing ineffective care. Status quo reimbursement prac- tices put Clinic 1 in an ethical no-win situation whereby there are notable moral reasons to avoid acting on any available option.

In some ways, Clinic 1 is unique. As a tertiary care intensive outpatient clinic, Clinic 1 serves a client popula- tion with challenging behavior that is both particularly se- vere and particularly difficult to change. To a large extent, the ethical challenges faced by Clinic 1 are the result of the intensive care needed by Clinic 1’s clients. Therefore, it may be unclear how the ethical challenges faced by Clinic 1 are relevant to behavior analysts who work in other clinical settings.

250 Behav Analysis Practice (2019) 12:247–254

Despite initial appearances, the financial and ethical pres- sures faced by Clinic 1 are relevant to behavior analysts who work in less intensive clinical settings. This is the case for at least two reasons. First, although Clinic 1 exclusively sees clients who require intensive outpatient therapy for severe and challenging behavior, these clients generally reach Clinic 1 through a series of referrals. Each referral comes from a clinic that is not equipped to provide therapy for the client in question. Although most behavior clinics do not specialize in clients with severe and challenging behaviors, they are none- theless likely to encounter such clients. In order to discharge their ethical obligation to protect the well-being of clients with severe and challenging behavior, clinics that do not specialize in severe and challenging behavior need to be able to refer these patients to clinics that specialize in working with this type of client. For this referral process to work, places like Clinic 1 must exist. Consequently, although most behavior analysts do not work in clinics that specialize in severe chal- lenging behavior, they nonetheless have an ethical stake in such clinics existing.

Moreover, the ethical analysis of Clinic 1’s billing structure is relevant to the average behavior analyst in a more immedi- ate way. Although we are unaware of data to support our anecdotal account, many of our friends and colleagues report being frustrated with preset upper limit constraints on the amount of time they can dedicate to an individual client. Althoughmany clients can receive the therapy they needwith- in the time constraints, many others need more time with the therapist than is allowed. Such time constraints are the inevi- table result of a reimbursement structure that will not reim- burse therapists for care provided in excess of some prespecified time limit (e.g., 53 min). Current reimbursement structures push behavior analysts to limit time with clients and to pack their schedules with as many clients as possible. Yet it is doubtful that such practices are in the best interests of cli- ents. Although the ethical implications of current reimburse- ment structures are most stark when considered in the context of Clinic 1, related ethical concerns are likely ubiquitous in the practice of behavior analysis.

ACOs, Quality Benchmarks, and Jane Doe

ACOs are a new type of health-care institution introduced by the Patient Protection and Affordable Care Act. ACOs “are groups of doctors, hospitals, and other healthcare providers, who come together voluntarily to give coordinated high qual- ity care to their Medicare patients” (Centers for Medicare and Medicaid Services [CMS], 2014). ACOs have been charged with achieving the “triple aim”: decreasing health-care costs, increasing health-care quality, and improving population health (Berwick, Nolan, & Whittington, 2008).

ACOs’ reimbursement structure sets them apart from other health-care institutions. ACOs are assigned a population of

Medicare patients based on where patients receive the major- ity of their primary care. Although ACOs are reimbursed un- der a standard fee-for-service (FFS) reimbursement model, ACOs also have the opportunity to share savings with CMS (CMS, 2014). Thus, in addition to receiving reimbursement through FFS, if an ACO can reduce expenditures on its assigned patient population, the ACO can receive up to 50% of the money it saved CMS (CMS, 2017, p. 7).

Offering providers an incentive to reduce costs is potential- ly problematic. The easiest way to cut costs is to deny care. Fortunately, a variety of mechanisms are in place to counter this concern. Of these, quality benchmarks are of particular interest. In order to qualify for shared savings, ACOs must meet 34 quality benchmarks (RTI International, 2016). These benchmarks include factors such as receiving high pa- tient satisfaction ratings, controlling high blood pressure, pro- viding fall risk assessments, and providing influenza vaccina- tions (RTI International, 2016).

There are, broadly speaking, two types of quality bench- marks: process benchmarks and outcome benchmarks (Rubin, Pronovost, & Diette, 2001). Process benchmarks determine if care providers have followed a set of prespecified steps. The benchmarks for fall risk assessment and influenza vaccination are both good examples of process benchmarks. Each checks to see if certain steps were followed in the provision of care (i.e., did patients receive a fall risk assessment and an influen- za vaccine?).

By contrast, outcome benchmarks eschew the focus on process. Outcome benchmarks measure the extent to which patients achieve predesignated measures of health. Good ex- amples are benchmarks that look at blood pressure or blood insulin levels (RTI International, 2016).

Quality benchmarks are particularly salient to our case study. By way of illustration, suppose the quality benchmarks for ACOs included the following process benchmark: Patients with self-injurious behavior receive recommended assessment and treatment. Developing quality benchmarks is an extensive process that requires calibrating the specificity and sensitivity of the measure while also balancing health-related consider- ations with the financial and data collection demands on an ACO. Proposing a plausible candidate for a process bench- mark related to behavior–analytic services is a project that falls well beyond the scope of this article. Therefore, although the aforementioned benchmark is too ambiguous to constitute a plausible candidate for an actual ACO quality benchmark, it can nonetheless usefully illustrate the promise of process benchmarks for behavior analysts in the clinical setting.

Jane Doe was referred to Clinic 1 because her previous tertiary care provider felt that she needed the level of intensive care provided by Clinic 1. Thus, the recommendation is that Jane Doe receive 1,800 min of assessment and treatment. If Clinic 1 is only being reimbursed for a fraction of the recom- mended 1,800 min of care, an ACO will not have the paper

Behav Analysis Practice (2019) 12:247–254 251

trail it needs to demonstrate that it has provided the recom- mended level of assessment and treatment. Consequently, the introduction of a process benchmark for the assessment and treatment of problem behavior would require ACOs to fully reimburse Clinic 1.

There are at least two distinct avenues an ACO could pur- sue in order to achieve this outcome. First, an ACO could attempt to renegotiate what codes are approved for reimburse- ment—for example, an ACO could fight to allow Clinic 1 to bill using ABACs, including billing using these codes for patients without autism spectrum disorder (ASD) and, de- pending on the ABAC, independent of a patient’s age. Because failing to meet outcome benchmarks threatens an ACO’s shared savings, an ACO would have a significant in- centive to throw its weight behind renegotiations. Alternatively, an ACO could choose to reimburse Clinic 1 on its own. Although this would be an expensive option, it would likely cost less than the ACO would