Cost-Effectiveness Analysis in the Courts
Recent Trends and Future Prospects
Peter D. Jacobson, University of Michigan School of Public Health
Matthew L. Kanna, J.D./M.H.S.A. Candidate, University of Michigana
This article was originally published in the Journal of Health Politics, Policy and Law. All rights reserved. This material may be saved for personal use only, but may not be otherwise reproduced, stored, or transmitted by any medium, print or electronic, without the explicit permission of the copyright holder. Any alteration to or republication of this material is expressly prohibited.
It is a violation of copyright law to reproduce any copyrighted information from this publication without first obtaining separate permission directly from the copyright holder who may charge fees for the use of such materials. It is the responsibility of the user to contact and obtain the needed copyright permissions prior to reproducing materials in any form.
Permission requests should be directed to:
Duke University Press
Durham, NC 27708
Fax: (919) 688-3524
Organization of the Article
Cost-Effectiveness Analysis (CEA)
Using CEA in Clinical decisions
Barriers to Using CEA in Medical Decisions
CEA and Cost-Containment in the Courts
The Standard of Care
CEA and Cost-Benefit Analysis (CBA) in General Negligence Cases
CEA and Cost-Containment in Medical Liability Cases
CEA in Cases Challenging Government Regulation
The Standard of Care in the Managed Care Era
This article provides an initial look at how managed care organizations (MCOs) might incorporate cost-effectiveness analysis (CEA) into their decision-making process and how the courts might respond. Because so few medical liability cases directly involve CEA, we must look at other areas of the law to assess potential MCO liability for applying CEA. In general negligence cases, courts rely on a risk-benefit test to determine customary practice. Likewise, in product liability cases, courts use a risk-utility calculus to determine liability for product design defects. And in challenges to government regulation, courts examine how agencies use CEA to set regulatory policy. The results have been mixed. In product liability cases, CEA has led to some punitive damage awards against automobile manufacturers. But courts have integrated it in negligence cases without generating juror antipathy, and generally defer to agency expertise in how to incorporate CEA. The article discusses the implications of these cases for MCO use of CEA and outlines various options for setting the standard of care in the managed care era.
Cost control is a primary objective of the managed care environment.1 It is no longer possible to provide health care without regard to cost or patient demand. The question is not whether there will be cost containment, but how to structure and oversee its implementation. Neither the courts—to whom patients frequently turn when the general need for cost containment turns into the specific need to deny treatment—nor other policy makers can successfully resolve managed care disputes by ignoring or wishing away the fundamental fact of scarcity. At issue is how courts will respond to managed care's cost containment initiatives in developing liability principles for the managed care era.
A potentially significant cost containment approach is the use of cost-effectiveness analysis (CEA) in making clinical and payment decisions. This article provides an initial look at how managed care organizations (MCOs) might incorporate CEA into their decision-making process and how the courts might respond. We examine how courts have viewed CEA when used by private parties to set industry custom and by governmental agencies in promulgating regulations. We also consider what these judicial trends portend for future legal doctrine in health care. At present, there is a lot of skirmishing over cost containment. MCOs seem reluctant to publicize the role of cost constraints in clinical and payment decisions, and the courts are slowly figuring out how to construe the reality of scarce resources in applying liability standards to MCOs. It probably won't be long before the skirmishing breaks out into direct conflict. In anticipation of the conflict, this essay suggests ways in which courts might incorporate CEA into the standard of care. We conclude that CEA should be treated as one piece of evidence to be considered by a jury rather than being used to determine the standard of care.
The next section discusses the concept of CEA, how it might be used in clinical decisions, and barriers to its use. Then we look at how CEA has been interpreted in the courts. After briefly describing how courts establish liability (the standard of care), we discuss how courts have used CEA in a range of case types, including nonmedical negligence litigation, medical malpractice cases, environmental law, and product liability cases. We then examine how CEA might arise in future health care cases and how the courts might respond. We conclude with a discussion of what the standard of care should be in the managed care era.
Cost-effectiveness is a widely used but imprecise term that means different things to various users. Like the term medical necessity, there is no common conceptual understanding of what it means or how it should be used (Gold et al. 1996). According to A. M. Garber et al. (1996: 26), "Cost-effectiveness analysis is a method designed to assess the comparative impacts of expenditures on different health interventions."2 CEA assesses the advantages and disadvantages of alternative interventions to examine the inevitable trade-offs in resource allocation (Berger and Teutsch 2000). Even though the definition is seemingly straightforward, CEA is deceptively difficult to apply.
In assessing alternatives, CEA uses a ratio where the denominator is the gain in health (such as adverse reactions avoided) and the numerator is the incremental cost of obtaining the benefits. The denominator may be expressed in years of lives saved or undesirable outcomes averted. The primary advantage of CEA is the ability to compare two interventions aimed at the same outcome. But a disadvantage is that the denominator may omit important aspects of quality of life, satisfaction, different preferences, values, etc. (see, e.g., Goold 1996; Gold et al. 1995). Another problem is that by definition, cost-effectiveness analysis presupposes knowledge regarding the overall effectiveness of a given clinical intervention, which is often not known. This problem plagues many areas of clinical decision making, helping to explain why few clinical decisions are grounded explicitly or primarily on the results of a CEA.
CEA differs from cost-benefit analysis (CBA) and cost-utility analysis (CUA) in how the benefit is expressed.3 In CEA, the common measure is one of nonmonetary effectiveness, while CBA is expressed in dollar terms and CUA is expressed in quality adjusted life years (Titlow et al. 2000). Because of discomfort with ascribing dollar values to health outcomes, CEA is more often used by health services researchers.
As an example, consider the debate over whether to provide yearly mammograms to women under fifty. According to one synthesis of the literature, yearly mammogram screening is not as cost-effective for women under fifty as it is for women between fifty and sixty-nine ($105,000 per cost of year gained for the former relative to less than $50,000 for the latter) (Salzmann, Karlikowske, and Phillips 1997; see also Eddy 1997; Hirth et al. 2000). Suppose that based on this information, a plan decides not to offer mammograms to women under fifty and that a forty-five-year-old then sues the plan for failure to detect her breast cancer. She alleges that the failure to screen resulted in a more advanced disease. The plan will argue that the money saved by not investing in the less cost-effective screening can better be allocated to more productive uses. The patient will argue that the use of cost-effectiveness ignores unmeasured benefits and that there is no widely accepted threshold for determining when a particular intervention should be provided (Hirth et al. 2000). Although other evidence will be presented, such as professional consensus statements, how the jury reacts to the plan's CEA will likely determine the decision.
So far, there is little direct evidence that CEA is an integral aspect of clinical or payment decision making among managed care plans. For example, M. R. Gold et al. (1995) examined the state of the art in cost control mechanisms as of 1994 but do not mention CEA in their report. In a recent and thorough review of managed care practices, Gold (1999) makes no mention of CEA as an integral element in MCO cost containment practices. It seems as though CEA is more talked about than implemented (Prosser et al. 2000; Jacobson et al. 1997). Nonetheless, MCOs may be using CEA but implicitly and below the radar screen. For example, cost considerations could be incorporated implicitly into the development of clinical practice guidelines or in rendering utilization review decisions. Because MCOs are cost-conscious organizations, these considerations must enter into their decisions.
If CEA were to operate as well as its proponents expect, CEA would become a basis for clinical decision making so that the most cost-effective clinical alternative would be chosen. At this point, however, how MCOs might actually use CEA in making clinical or payment decisions is an evolving process. While the literature does not specify how that process is working right now, there are several identifiable areas where CEA can be expected to emerge.
Benefit Packages. CEA could be used to determine the set of benefits to be offered to managed care subscribers. Plans might contract with employers to use CEA in making benefit determinations for subscribers. For example, the contract might state that based on CEA, yearly mammograms will only be provided for women over fifty.
Medical Necessity Determinations. Despite the continuing controversy over how to define and use the concept of medical necessity, it remains a commonly used concept for making clinical and insurance payment decisions. David M. Eddy has perhaps been the foremost proponent of including cost-effectiveness as an element of clinical decisions. Working with the National Institute for Health Care Management, Eddy (1996) developed a proposed approach that would require health plans to cover interventions based on the following criteria: (1) used for a medical condition; (2) sufficient evidence to draw conclusions about the effects on health outcomes (including the quality of life); (3) evidence that the intervention will produce the intended effects; (4) beneficial effects outweigh the harmful effects; and (5) the intervention is the most cost- effective method available to address the medical condition (NIHCM 1994). One way to incorporate CEA in clinical decisions is through the use of clinical practice guidelines based on cost-effectiveness criteria. Kaiser Permanente did so in developing guidelines for using contrast agents (liquid dyes used to improve the clarity of Xrays) without any resulting litigation (though not all physicians were exactly thrilled with the idea) (Eddy 1996; Jacobson and Rosenquist 1996).
It is not clear how widely CEA is incorporated into medical necessity decisions. One study found that some medical necessity definitions incorporate CEA but that respondents were divided as to its efficacy (Jacobson et al. 1997). The study found that administrators generally favored its use, while physicians (including plan medical directors) viewed it as an intrusion into their clinical autonomy. In contrast, a more recent study in California also found that only a few medical necessity contractual definitions included cost-effectiveness, but that private plan and medical group medical directors mentioned cost-effectiveness as a prominent criterion for clinical decision making (Singer et al. 1999). If confirmed by other studies, the Singer et al. results suggest that there is a gap between what plans actually do and what is reported publicly or in the literature. Another important finding from the California study is that many respondents would approve an equally effective but more costly treatment "for fear of litigation or backlash if it were to be discovered that they had considered cost" (ibid: 57).4
Utilization Management. After a physician determines that a particular intervention is medically necessary, the MCO might then review the decision as part of its utilization management (UM) regime. At this point, the MCO could use a range of factors, including CEA, to determine whether the recommended intervention should be provided. Nothing in the literature shows that CEA is now an explicit component of the UM process, although it seems clear that cost considerations play an important role. Many of the studies cited above indicate that the incentives inherent in managed care are factored into the decision, including the use of clinical practice guidelines and preauthorization for high-cost medical interventions.
Pharmacoeconomics. A fourth area in which CEA could play a prominent role is in the emerging field of pharmacoeconomics. Because pharmaceutical costs are rising at such a rapid rate, many health care providers are turning to pharmacoeconomics to help them determine which pharmaceuticals to cover in the drug formulary. By definition, this process involves CEA. Most of these analyses are conducted by pharmacy benefit managers (PBMs) not by MCOs (Titlow et al. 2000; Rosoff 1998).
Indeed, a recent survey found that MCOs rely on information provided by PBMs, including CEA, literature reviews, drug utilization reviews, and recommendations for disease management programs (Titlow et al. 2000). The study also found that cost was only the third most often cited factor determining pharmacy coverage decisions, behind the drug safety profile and FDA approval. Costs were an important factor in shaping the overall formulary but not in deciding whether to cover individual pharmaceuticals. The authors noted that survey respondents "might have been reluctant to openly acknowledge the dominance of cost," but other data supported their findings (ibid.: 245; see, e.g., Jacobson and Rosenquist 1996). Nevertheless, this raises the question of which entity would bear the potential liability if an MCO is sued for excluding a pharmaceutical based on a PBM's CEA.
In theory, MCOs would appear to be perfect proponents and users of CEA. After all, the guiding concept of managed care is to provide high-quality care at a lower cost than in the fee-for-service system. Yet it appears that CEA is not widely used by MCOs, at least not explicitly or as reflected in the published literature. One of the few exceptions to this, as noted above, is the use of CEA in Kaiser Permanente's decision to use lower cost contrast agents for low-risk patients (Eddy 1992).
In a recent article, L. A. Prosser et al. (2000) suggest that serious barriers limit MCOs' use of CEA. They argue first that MCOs are not aware of the usefulness of CEA or, if they are aware, may still question its applicability. Clinical effectiveness, as opposed to cost-effectiveness, appears to be driving the decision-making process. Surprisingly, Prosser et al. argue that MCOs actually have limited incentives to use CEA. Right now, plans are able to achieve cost control through the other mechanisms, such as capitation, noted above that shift the cost control burden to physicians. That way, an MCO may have less risk of liability and be subject to less public anger.
Not surprisingly, Prosser et al. (ibid.) note the negative public perceptions (i.e., public relations problems) that could emanate from explicit use of CEA. Noting that Oregon had to retreat from using CEA in setting its priority list, the public may not be ready to accept an explicit CEA clinical approach.5 Many MCOs are already reeling under the onslaught of the managed care backlash and may be reluctant to push much further. Physicians have also reacted negatively to including CEA in clinical practice guidelines as an intrusion into physician autonomy. Interestingly enough, Prosser et al. do not include litigation as a potential barrier. Given the recent jury verdict against General Motors for using CEA in deciding where to locate the gas tank (discussed below), it would be surprising if MCO administrators were not at least somewhat concerned about liability.
Currently, there are both methodological and process difficulties in implementing CEA. The methodological problems include data timeliness, bias in the data sources, the lack of comparative studies, disputes over measuring costs and effectiveness, concerns about the subjectivity inherent in defining the denominator, and disagreement about the benchmark cost-effective threshold to be used (see, e.g., Hirth et al. 2000). CEA may undervalue health benefits and there may be limited evidence of clinical effectiveness for a wide variety of interventions. There is also no standard CEA methodology that can be pulled off the shelf to use, making it time-consuming to develop. Two editorials commenting on the Prosser et al. (2000) article are also cautious about using CEA based largely on the methodological problems involved (Livak, Long, and Schwartz 2000; Langley 2000).
The process barriers are equally daunting. For one thing, the quality of the analysis is an important consideration and often in doubt (Rennie and Luft 2000). For another, it is imperative that the process be transparent to gain public legitimacy. As Rennie and Luft (ibid.: 2159) argue: "The key requirement for any cost-effectiveness analysis is that the assumptions, models, and possible biases are well described, transparent, and fully supported by evidence, the strength of which is made easily available to any critical reader." Without this transparency, the essential aspects of patient understanding and informed consent cannot be fulfilled. Without patient understanding and informed consent, the process lacks public legitimacy.
To date, there is little health care litigation interpreting the use of CEA. Eventually, MCOs will begin more explicit use of CEA in one or more of the areas just described, so it seems inevitable that courts will directly confront the issue. If challenged, MCOs might justify the denial of care to an individual patient by arguing that it would not be in the best interests of the patient population to provide benefits that are not cost-effective. Plaintiffs, on the other hand, may want to introduce CEAs to discredit the MCO by showing that the denial of care was not based on sound methodologies, undervalued care for the individual patient relative to the patient population, or was contrary to stated plan criteria.
These efforts will raise several questions. For trial courts, to what extent will this be a question of admissibility versus the weight of the evidence? (Admissibility is a judicial determination that evidence is relevant and can be introduced into trial. Once introduced into evidence, the jury's responsibility is to weigh CEA against all the other evidence.) If the former, what criteria will be used to guide admissibility? If the latter, how can judges balance the probative value of the CEA with the potential prejudice against the defendant seen in automobile safety litigation? If admissible, should CEA be used to define the standard of care, or be treated as simply another piece of evidence for the jury to weigh?
One way to address these questions is to examine how courts have responded to CEA in other areas of the law and to other managed care cost containment innovations. In this section, we consider general liability decisions, medical liability cases, product liability cases, and cases challenging governmental regulation (particularly environmental issues). For those not familiar with the underlying legal standards, we first provide a brief overview of liability principles.
Establishing Liability. At its simplest, tort law (torts are civil wrongs such as negligence) establishes standards of behavior that individuals and businesses are expected to meet in avoiding unreasonable risk of harm to third persons. Each state's court system establishes its own body of common law negligence principles, although all states use the same general framework. This means that legal doctrine will vary across states, so that what may be negligent in one state will not be negligent in another.
To win a negligence case, the plaintiff must prove the following four elements by a preponderance of the evidence: (1) a duty of due care; (2) breach of that duty; (3) the conduct caused the injuries; and (4) the injury produced actual damages. To determine whether a defendant has breached his duty of due care, courts often look to custom in the industry, and, in the absence of custom, to reasonableness to set the standard of care. How would a reasonable person have acted under the circumstances? The reasonable person standard allows the jury to make an informed decision about whether the defendant's activity met the community's standard of due care or created an unreasonable risk of harm.
One of the hallmarks of the common law is its ability to adapt incrementally to changing social and economic circumstances. By deferring to industry custom, courts give the marketplace considerable flexibility to determine how and when to introduce the latest technology or safety advances. Not every conceivable safety precaution must be taken-only those that are justified by the costs of injury prevention. Thus the utility or social value of the conduct must be weighed against the risks (United States v. Carroll Towing Co., 159 F.2d. 169 [2d Cir. 1947]; Restatement [Third] of Torts: General Principles, Article 4 ). In negligence cases, the courts are free to overrule industry custom and impose more stringent standards of care if the industry is slow to adopt technologies or systems that would avoid injury (see, e.g., The T. J. Hooper, 60 F.2d 737 [2d Cir. 1932], cert. denied, 287 U.S. 662 ).
Establishing Medical Liability. In medical liability cases, the standard of care is exclusively set by the medical profession itself based on what is customary and usual practice, as established through physician testimony and medical treatises. A typical statement of the law is that each physician must "exercise that degree of skill ordinarily employed, under similar circumstances, by the members of [the] profession" (Lauro v. The Travelers Insurance Co., 261 So.2d 261 [La. 1972]). In effect, this means that the same level of care must be provided to all patients, regardless of resource constraints. The primary reason why medical liability diverged from general negligence is deference to professionalism; courts did not feel capable of second-guessing customary medical practice. Instead, courts have consistently held that nonphysicians do not have sufficient training to establish customary and reasonable medical practices (Prosser 1978; Peters 2000).6
Most courts presume that a physician's failure to adhere to customary practice constitutes negligence. If there is more than one recognized course of treatment, most courts allow some flexibility in what is regarded as customary (known as the respectable minority rule). In relatively rare instances, courts will allow a plaintiff to challenge the adequacy of customary medical practice, resulting in a higher standard of care than that determined appropriate by the profession.
Custom and cost containment are not inherently in tension. The point of cost containment is to reduce the amount of inappropriate health care that led to the explosion in health care costs under the fee-for-service system. At a minimum, cost containment aims to change what has been the custom. The assumption is that what has been customary is in fact "too much" care, unjustified either from the standpoint of societal allocation of resources or, more often, from the standpoint of medically necessary care for the individual patient. In this sense, managed care is designed to influence how the standard of care will be set by physicians (by implementing, for example, treatment protocols). One of the common accusations against managed care is that it aims to depress the required standard of care too far, below an optimal level. Those who advocate a move away from the professionally determined standard of care are instead concerned that levels of care have been set too high and argue that resource intensity does not necessarily equal quality of care (Hall 1997; Havighurst 1995; Epstein 1997; Morreim 1997).
The use of CBA and CEA in litigation grew from legal standards governing general negligence. Well into the twentieth century, there was little mention of any standardized formula for calculating the proper amount of precautions necessary to avoid claims of negligence. The reasonable or "prudent man" standard dominated both in academic discussions and in the courtroom, although there were pockets of what could be considered CBA scattered throughout the case law (Green 1997; see, e.g., The T. J. Hooper, 60 F.2d 737 [2d Cir. 1932]).
In the most famous case in which a cost-benefit approach was explicitly adopted, United States v. Carroll Towing (159 F.2d 169 [2d Cir. 1947]), the defendant owned a tug and was moving a line of unmanned barges out to sea when one broke loose and collided with a nearby tanker. The tanker's propeller punctured the hull of the barge, which then began to take on water. Eventually the barge sank, along with its cargo. Faced with the absence of precedent determining when barge owners were liable for not maintaining watch to ensure that their vessels did not break away from their berths, Judge Learned Hand (159 F.2d 169, 173 [2d Cir. 1947]) reasoned:
Since there are occasions when every vessel will break from her moorings, and since, if she does, she becomes a menace to those about her; the owner's duty, as in other similar situations, to provide against resulting injuries is a function of three variables: (1) the probability that she will break away; (2) the gravity of the resulting injury, if she does; (3) the burden of adequate precautions. Possibly it serves to bring this notion into relief to state it in algebraic terms: if the probability be called P; the injury, L; and the burden, B; liability depends upon whether B is less than L multiplied by P: i.e., whether B is less than PL.
In short, negligence occurs when the burden (cost) of investing in accident prevention is less than the expected liability (P 3 L). Likewise, if the cost is greater than the expected liability, the defendant would not be negligent. Conceptually this formula makes sense, and its similarity to modern cost-benefit analysis formulas is readily apparent. Additionally, it is an easy-to-grasp (if not always easy to apply) guideline that allows for a great amount of flexibility. Of course, it suffers from the same problems that plague all cost-benefit and cost-effectiveness analyses, but that has not hindered its adoption by the courts.
The Hand formula and its derivations now dominate general negligence law. Just eighteen years after the decision in Carroll Towing, Section 291 of the Restatement (Second) of Torts (1965)7 provided a risk-benefit test for unreasonable conduct and negligence, reflecting the standard put forth by Judge Hand. And the current draft (1999) of the Restatement (Third) of Torts clearly expects negligence cases to be decided based on a risk-benefit test (General Principles, Article 4, Comment d). The Carroll Towing analysis has also become "a staple of the law and economic scholarship addressing tort law" (Green 1997).
But even so, the common jury charge to this day involves some variant of the "reasonable man under like circumstances" standard and not a firm admonition to weigh costs or risks against benefits. Thus what remains very murky is whether the Hand formula is actually used to establish the standard of care, is just implicit in the reasonable man jury instruction, or is used by appellate courts in reviewing case outcomes.