Ms. Rosenbaum, Chair of the Department of Health Policy and Health Services at the George Washington University School of Public Health and Health Services, provided the committee with additional insights into “what kinds of issues were on the minds of lawmakers” as they drafted the “required elements for consideration” portion of Section 1302.2 These required elements for consideration, she said, were not included in the committee versions of what became the final statute. Instead, these provisions were added when Senate Majority Leader Harry Reid assembled the final Senate bill that was ultimately signed into law. On the House side, members and staffers expressed concern about having protections for people who have greater than normal health care needs. Ms. Rosenbaum noted that the required elements are not part of the benefit entitlement itself. Rather, the inclusion of the required elements was a directive to the Secretary of the U.S. Department of Health and Human Services (HHS) to articulate standards for benefit design and coverage determinations to modify “what has been historically at the discretion of insurers.” Specifically, this section of the law addresses having balance among categories of care and not having coverage decisions, reimbursement rates, incentive programs, or benefits that discriminate based on age, disability, or expected length of life. This discretion relates not only to individual medical necessity decisions, but also to coverage design at the macro level.
Ms. Rosenbaum noted it is very unusual for federal law to “deal with the content of health insurance” from a regulatory viewpoint. The Patient Protection and Affordable Care Act (ACA) echoes Medicaid’s heretofore unique nondiscrimination rule.3 Even the Americans with Disabilities Act of 1990 (ADA)4 “does not deal with the content of health insurance,” an issue addressed in a leading federal case whose decision was allowed to stand by the U.S. Supreme Court.5 Furthermore, the Supreme Court has expressly held that Section 504 of the Rehabilitation Act,6
1 Financial support for development of analyses in Ms. Rosenbaum’s statement comes from The Commonwealth Fund in conjunction with development of an expanded article on the subject published subsequent to this presentation (Rosenbaum et al., 2011).
3 Code of Federal Regulations, Title 42 Section 440.230(c) barring nondiscrimination in the provision of required services on the basis of condition.
4 U.S. Code, Title 42 Sections 12101 et. seq.
5Doe v. Mutual of Omaha Insurance Company, 179 F.3d 557 (1999) cert. den. 528 U.S. 1106 (2000).
6 The Rehabilitation Act of 1973, Public Law 93-112 § 504 (September 26, 1973).
A key question for HHS is to consider how to comment on the use of limits—which are prevalent in benefit plan administration—given the ACA’s intent to reform insurance industry practices, she said. Health insurance coverage entails legal and financial risk. For this reason, insurers logically seek to structure their products to provide risk exposure protection against the covered population. These risk avoidance techniques, however, go well beyond strategies for assuring that insurance pays only for medically necessary care. Ms. Rosenbaum cited an example of a once prevalent, but now precluded intoxication exclusionary clause from when many believed alcoholism was a behavioral choice not a medical condition: if the injury or the illness was related to intoxication, the plan would not provide coverage on the grounds of this “morals clause” (Rosenbaum et al., 2004). She urged the committee to keep this example in mind when developing recommendations for the Secretary. She clarified that it is necessary to put some limits on what is a covered treatment, but she believes the ACA is trying to convey that insurers cannot come up with treatment distinctions by labeling underlying conditions in ways that push specific individuals outside of the scope of the contract.
Under the ACA, the Secretary has the primary responsibility for setting up a federal framework for essential health benefits (EHB) coverage, but Ms. Rosenbaum opined that the Secretary may decide on a framework that delegates a “fair amount” of authority back to plans. For example, the Secretary could conceivably instruct insurers and plan administrators to utilize the terms and definitions in their most popular group health products as long as such terms and definitions do not discriminate on the basis of disability by, for example, defining a benefit as covered only in cases in which restoration or recovery are possible. This approach to tempering delegated authority is consistent with other laws, such as the Mental Health Parity Act,9 which gives insurers and health plans discretion over whether to cover mental illness and substance use disorder benefits but prohibits the use of discriminatory quantitative and non-quantitative coverage limits. “It is wrong,” she said, “to conclude that the essential health benefit statute leaves the status quo in play. The essential benefit statute is an enormous, profound departure from the way in which we’ve conceptualized the discretion of insurers, but that is not to say that [the Secretary] could not specify a fair and nondiscriminatory framework and process to avert discrimination in benefit design and making coverage determinations.”
The clearest precedent for the prohibitions against discrimination in the EHB statute can be found in the Medicaid statute. Since the Medicaid statute’s original enactment,10 its “reasonableness” provision11 has been understood by both the agency12 and the courts13 as barring arbitrary limits in required services based solely on an individual’s condition, diagnosis, or type of illness. Furthermore, at least two recent court decisions suggest that at least some courts will reject coverage denials under Medicare where the basis of the denial is the arbitrary exclusion of otherwise covered services based on absence of “recovery” potential.14
The required elements for consideration provision of Section 1302, Ms. Rosenbaum argued, is designed to address the issue of insurer discretion to discriminate against certain types of conditions in the context of benefit design and coverage determinations. In some cases, excluded conditions may be quite specific, while in others, a proxy of “recovery” or “restore” is commonly used to differentiate chronic conditions for which there may be
7 U.S. Code, Title 29 Section 794.
8Alexander v. Choate, 469 U.S. 287 (1985).
9 The Mental Health Parity Act of 1996, Public Law 104-204, 104th Cong., 2d sess. (September 26, 1996).
10 The precursor to the federal non-discrimination rule can be found in the Handbook of Public Administration, Supplement D, issued in 1966 by HEW.
11 U.S. Code, Title 42 Section 1396a(a)(17).
12 U.S. Code of Federal Regulations Title 42 Section 440.230(c).
13 See, for example, Pinnecke v. Preiser, 623 F. 2d 546 (8th Cir. 1980).
14 See, for example, Papciak v. Sibelius— F.Supp. 2d —, 2010 WL 3885605 (W.D. Pa.). and Anderson v. Sibelius, F. Supp. 2d, 2010 WL 4273238 (D.Vt.).
no “recovery,” from those that are acute and time-limited and for which recovery is possible. Disability and age both raise this issue, since age can affect the potential for recovery. The presence of developmental disabilities can also trigger “recovery” discrimination. The question should be whether treatment aids functioning and serves to maintain health or avert deterioration in health, not whether recovery can be expected.
For age: The statute bars discrimination, but not the use of patient characteristics when such characteristics rest on a reasonable clinical and scientific evidentiary base. A decision cannot be made “on the basis of age,” but a decision based on clinical factors (age is a recognized clinical factor when, for example, age is used to determine when to immunize a child against certain diseases) would not be a decision based on age.
For disability: Limits that require recovery or restoration inherently discriminate against individuals for whom the expected impact is the attainment of, improvement in, or maintenance of function, or the aversion of functional loss. Exclusionary clauses that limit coverage only to situations in which recovery can be expected create insurmountable problems for individuals who would benefit from medical care but for whom recovery is not possible. Similarly, exclusions that depend on whether the need for treatment also has been noted in an employment plan, individualized education plan, or some other document addressing the work-related, social, developmental, or educational needs of a patient could be considered discriminatory. The central question is whether the treatment is medical in nature and whether the individual can be expected to medically benefit from the treatment. The fact that the health benefit will have spillover effects in educational, employment, or social contexts should be irrelevant to the coverage determination.
For expected length of life: In the case of hospice treatment, for example, where length of life is accepted as a core element of the intervention, length of life might be a proper consideration. On the other hand, the ACA withdraws expected length of life as a criterion where reasonable clinical and other relevant evidence shows an individual’s ability to benefit from a treatment. The concept of “to benefit” encompasses attainment and maintenance of health as well as avoidance of deterioration.
Ms. Rosenbaum made a distinction between what she called “across-the-board limits” and condition-specific limits. Across-the-board limits, she said, apply to everyone in the plan, such as the exclusion of infertility treatments. Such an exclusion does not single out people with disabilities. Conversely, an embedded guideline might state “we cover infertility treatment, but not in situations in which the woman has one of four different conditions.” Such flat exclusion without evidentiary consideration of whether the underlying condition is one that impairs the ability to benefit from fertility treatment would be discriminatory. Hospitalization limits of 30 days per spell of illness or 60 physical therapy treatments may limit medically necessary care, but they do so without regard to the underlying condition. Such limits, Ms. Rosenbaum said, fall with particular severity on the sickest members of the coverage groups and are undesirable for many reasons (the most desirable result is, of course, to have enough scientific and clinical evidence to be able to make coverage design and administration decisions solely on the basis of the evidence). But courts have ruled that across-the-board limits on scope are not discriminatory against persons with disabilities under federal civil rights laws, nor, presumably, would across-the-board limits “discriminate” against individuals because of age or expected length of life.
Preventing Discriminatory Definitions
Ms. Rosenbaum suggested the Secretary consider cautioning insurers about using definitions, whether broad definitions for medical necessity or service specific ones, that inherently discriminate on the basis of a condition. A medical necessity standard limited to medical conditions, for instance, could be interpreted as not addressing coverage of treatments for developmental disabilities because they are often not considered medical conditions. There are, however, medical treatments, including speech therapy and physical therapy, appropriate for children with developmental disabilities. “Just because the condition is developmental does not,” Ms. Rosenbaum said, “make benefits aimed at alleviating or ameliorating the condition any less medical in nature.” The question is whether coverage can be expected to produce a beneficial impact, not whether that impact is “restorative.” She believes if an insurer were to exclude medical treatments based on an underlying condition and without regard to evidence from the patient’s record and, when available, research evidence, it would be applying a discriminatory definition. Similarly, if an insurer were to define speech therapy as “a therapy that is needed to allow somebody
to recover speech,” that definition would be discriminatory as it would withhold therapy from someone who never attained speech. Nor should it be at the discretion of the insurer to decide if therapy for a multiple sclerosis (MS) patient, for example, will be covered to avert a loss of functioning as opposed to recover prior functioning. As a general rule, she said, limitations that discriminate on the basis of an underlying condition should be avoided.
When committee member Ms. Ginsburg asked Ms. Rosenbaum to comment on how one would define “what is medical,” Ms. Rosenbaum clarified that despite being “an insurance lawyer” for her entire career, she has “no idea what a medical condition is.” It is more helpful, Ms. Rosenbaum said, to “look at the intervention. If the intervention is by a licensed clinical professional, it is regulated under a state scope of practice act governing the health professions.” Then, she said, it should be “recognized by insurers as a form of treatment when furnished by a licensed professional acting within the scope of practice.”
Committee member Dr. Selby asked whether the committee should be “cautious” about ensuring policies do not discriminate against conditions such as rare diseases for which there is less likely to be convincing, rigorous evidence. Ms. Rosenbaum reiterated that first, one needs to determine “what it means to discriminate” and suggested that her opinion is that discrimination is any “arbitrary distinction” and that may or may not relate to a rare condition. It will be decades, she said, before an evidence base will be available for rational distinctions based on the condition. Until then, she said, “we would really like to move away from condition-based distinctions in coverage … to focus on medical management across the board, not just in mental health, but generally, value-based coverage, incentives, yet not draw distinctions based on conditions, or age, or expected length of life.”
Committee member Dr. Sabin asked if it would be discriminatory for an insurer, for example, to acknowledge that children with autism spectrum disorders have severe speech delays, but regard those delays as an “educational problem.” In response, Ms. Rosenbaum cited a case in which an employer-sponsored plan administrator decided that certain types of treatments for a particular child with developmental disabilities were educational in nature and were therefore excluded from reimbursement.15 All of the treatments in question were medical treatments furnished by medical professionals in a clinical setting. They were, in Ms. Rosenbaum’s opinion, “unquestionably covered medical treatments, except that the child’s condition was labeled behavioral and the treatment was labeled educational, and that was the end” of coverage for the child.
Her response prompted Dr. Sabin to argue that some medical treatments, including physical therapy, can be used to enhance, rather than treat, a condition. Ms. Rosenbaum responded that in the absence of expressed exclusions in plan documents, courts have generally resisted efforts to define medically necessary as: (a) limited by the kind of condition or (b) limited by the notion of recovery.16 Medical necessity, she said, should cover the treatment regardless of whether it will aid in recovery or whether, instead, it is necessary to develop, maintain, or avert loss of a function. She noted the inherent weakness in a definition of medical necessity that is limited to “medical conditions.” The phrase “is a mechanism for eliminating certain kinds of conditions that some people do not consider medical.”
Committee member Dr. McGlynn noted that when Dr. Mark McClellan was the administrator of the Centers for Medicare & Medicaid Services (CMS) from 2004 through 2006, he attempted to require, that when a treatment lacked evidence, that such evidence had to be generated as a condition of coverage. Is this practice, Dr. McGlynn asked, “inherently discriminatory?” Ms. Rosenbaum stated that Dr. McClellan’s attempt “was brilliant.” Currently, the disability community faces a dearth of evidence and consequently experiences denials. As suggested by Dr. Gary Ulicny in Chapter 7, she said the rehabilitation treatment community itself is attempting to gather the evidence needed to justify interventions. The population with disabilities and others with heightened health care needs “welcome a greater focus on evidence-based practice and reporting.” A review of the case law on coverage suggests that individuals most seriously burdened by health conditions are the most affected by denials, Ms. Rosenbaum said. Their high representation in judicial decisions underscores the enormous costs they face when an insurer or plan administrator denies coverage and the corresponding importance of braving a lengthy and difficult appeals process.
15Mondry v. American Family Mutual Insurance Company, 557 F. 3d 781 (2009), 130 S. Ct. 200 (2009).
16 See, for example, Bedrick v. Travelers Insurance Company, 93 F. 3d 149 (4th Cir., 1996) and McGraw v. Prudential Insurance Company of America, 137 F. 3d 1253 (10th Cir., 1998).
Ms. Rosenbaum advised the committee against allowing insurers to embed treatment guidelines in contracts of coverage and plan documents. These normative, undisclosed, and fixed limits on treatment do not allow for any deviation based on underlying condition. As a result, she said, they can be considered essentially definitional. She suggested that guidelines are often not based on evidence; they are simply conclusory statements by an actuarial firm. Furthermore, even if clinically sound, a guideline may be used out of context as a coverage limitation as opposed to an informal guide, where one would weigh other clinical evidence.
When committee member Ms. Monahan asked for more details about how insurers should use evidence-based guidelines regarding disabilities, Ms. Rosenbaum acknowledged that “guidelines may be used completely out of context” to limit coverage. The purpose of these guidelines, she said, is to suggest that certain underlying conditions may inhibit or impair the success of treatment. Guidelines can help inform decisions, but they should not be embedded in plan documents as an actual limitation on coverage. Committee member Dr. Santa followed up on this exchange by asking Ms. Rosenbaum whether she would agree that if “the evidence is more robust, you are immunizing yourself from discrimination, but to the degree the evidence is not robust, then you’re susceptible to discrimination.” Ms. Rosenbaum concurred, also noting that flexibility can help mitigate against arbitrary distinctions that are often made when the evidence is weak. Flexibility would allow determinations to consider, she said, whether “there is a reasonable basis” for the treatment to be covered, even in the absence of good evidence. Medicaid, for instance, cannot draw arbitrary distinctions solely on the basis of a condition.17 HHS should also worry about the evidence base to justify inclusions and denials, and often, the evidence-free zone is showing up on both sides.
Ms. Rosenbaum concluded by suggesting the committee be “less worried about the specific terminology and more worried about what the consideration section tries to do.” Section 1302(b)(4) of the ACA provides “required elements for consideration” and tries to “bring some balance to content limits on coverage.” The ACA is trying to prevent discriminatory types of limitations and exclusions (i.e., that are not solely determined on the basis of age, expected length of life, or disabling conditions) in favor of limits that are reasonable.
Rosenbaum, S., H. Van Dyck, M. Bartoshesky, and J. Teitelbaum. 2004. SAMHSA policy brief. Analysis of state laws permitting intoxication exclusions in insurance contracts and their judicial enforcement. http://www.gwumc.edu/sphhs/departments/healthpolicy/dhp_publications/pub_uploads/dhpPublication_3626D84B-5056-9D20-3DE5C10098AB28B8.pdf (accessed May 4, 2011).
Rosenbaum, S., J. Teitelbaum, and K. Hayes. 2011. The essential health benefits provisions of the Affordable Care Act: Implications for people with disabilities. New York, NY: The Commonwealth Fund.
17 U.S. Code of Federal Regulations, Title 42 Section 440.230(c).