|
Competition And New Technology
Most real medical spending growth is accounted for by beneficial but costly new technology. This paper argues that a missing dimension of our concepts of competition among health plans is a focus on their policy toward new technology. In principle, plans could choose to move rapidly or slowly, inclusively or selectively, toward adopting new technologies, broadly defined. These policies would affect the rates of growth of their premiums, and consumers could choose among plans depending on both the technology policy and premium growth. Legal impediments, physician resistance, and membership turnover are all possible obstacles. Further thought should be given to making technology policy an explicit dimension of plan competition.
The U.S. medical care sector shows two faces. One expresses rising spending, at real rates well in excess of the growth rate of real income (and almost everything else); the other shows rising quality and improved treatment and prevention of illness, largely attributable to the invention, development, and marketing of new technology. To a considerable extent, these two faces reflect the same reality: consumers’ willingness to pay more for better outcomes and suppliers’ (whether drug firm, imaging center, or specialty practice) willingness to invent and then supply new services in return for these higher payments. This pattern of spending rising and outrunning income is not unique to medical care; it also characterizes the rest of the services sector (of which medical care is an appreciable but by no means dominant part). At this level of abstraction, there is nothing especially ominous or troubling about the fact that not all categories of consumer spending grow at exactly the same rate and at the same rate as income. With regard to health care, consumers in general might prefer to devote larger shares of their new wealth to enhancing the quantity and quality of their lives than to buying more and better automobiles, home furnishings, or accountant services.1 However, medical care surely garners more than its share of criticism because medical markets do not have the same competitive transparency that others have. That is, in most other cases, it is easy to believe that people spending more on most things are getting more value for more money. If, for example, people choose to spend their growing incomes for high-price gourmet coffee drunk in comfortable chairs, we don’t worry. But medical care, we think, is different. Although there is evidence that in the aggregate and on average, there is value for money from improved medical care, it is surely not precisely and transparently true for each person or product.2 Certain subsidies to public and private insurance might account for some of this additional spending, but they seem too small and too constant to explain the spending growth we observe; they also seem unlikely to be abolished anytime soon. In my judgment, it is reduction of the anxiety associated with this vagueness about value, rather than guaranteed improvements in quality or efficiency, that would be the biggest dividend of a transparently competitive health care market, even given the other defects in the system we have. What such a market might look like, and what might have prevented its emergence up to this point, are the two questions I address in this paper. Because of my focus on competitive markets, I do not discuss single-payer systems or other highly regulated options. Other countries have opted for a regulatory route; although there is a great deal of opinion about what works best, there is little agreed-upon fact, other than that both regulation and markets are less than perfect. We do know that recent rates of growth in health spending (not the levels) have been similar across developed countries. Instead of prolonging an inconclusive discussion, I stick to discussing competitive markets. The importance of technology. The growth rate of spending on new medical care technology is so large that any effective effort to slow it appreciably would soon swamp any alternative program to improve efficiency, say, by reducing geographic variation in spending or spending on overuse of tried-and-true medical services.3 Taking out the effect of economywide inflation (at 2 percent), more than half of the 6.8 percent per capita real health spending growth in 2002 is accounted for explicitly by technology (here called "utilization" and defined to include the "quality and mix of services"). Moreover, some of the second-largest component, medical price increases, almost surely includes some changes in the nature of the product.4 Yet almost none of the debate about competitive solutions focuses on new technology. Take, for example, the debate over introducing more marketlike competitive arrangements into Medicare. Advocates claim that through consumer choice, with payment rates set by competitive markets or bidding, sizable sums will be saved through greater efficiency of production. Critics fixate on possible harms to high-risk patients and "problematic" evidence that competition might not save money.5 All ignore the new-technology monster in the closet, and yet it is the technology-fueled and apparently inexorable growth of Medicare spending, not its possible current-period inefficiency, that makes Medicare as we know it infeasible in the relatively near future and will lead this program to consume much of the federal budget and a large share of gross domestic product (GDP).6 Beyond bland assurances that consumer-directed health care or government control will wring out enough waste, fraud, and abuse to pay for this new technology, what coping mechanisms should have been discussed? I address this question. What do we want that we can have? For Medicare or private coverage for the nonpoor nonelderly, what are the reasonable objectives or criteria for new technology? What we want is easy (for economists) to define, hard to persuade ordinary citizens to accept, and almost impossible to measure: We should want all of the new technology added every year that is worth more than its cost and no technology added that is worth less than its cost, no matter how much good it does. Clearly, any technology that improves outcomes and lowers cost will fill this bill. But so does cost-increasing but quality-enhancing new technology, up to a point. The point, of course, is where each person’s marginal benefit, to the patient and to any others in society concerned about the patient, equals the full marginal cost of developing and supplying that technology to that person. This straightforward view is not politically correct. In polite policy discussions and speeches, either quality per se is defined as always good, or quality is defined as always cost-reducing. The evidence on spending growth overwhelmingly disproves the latter, and any adult reasoning disproves the former. But it is impossible to find a policymaker, even a retired one, who is willing to advocate for limitations on effective medical technology that will surely mean that some good will be left undone. Since the general public believes that high and growing spending for medical care is generally a result of waste and fraud (or at least injustice), the absence of anyone willing to try to disabuse them of this belief (and its corollary that they should vote for the person who can eliminate waste and fraud) inhibits progress.7 Among the "reasons for higher costs," only 7 percent of respondents in a recent poll gave "the use of expensive high-tech medical equipment and expensive new drugs" as the most important reason; high drug company profits, malpractice lawsuits, and greed and waste were all ranked more highly, in the double digits.8
The idea that competition could reduce spending has some empirical and conceptual support. Before the advent of managed care (back when health spending was 7 percent of GDP), there was great concern about rising spending that was attributable to new technology. Economists responded to these concerns with a model and a solution. The model was one of "quality-only" or "medical arms race" competition. At least in hospitals, the only heavily insured part of health care at the time, the then-prevailing Blue Cross or commercial plans virtually all required completely open panels. Any licensed hospital was eligible for payment, which was usually enough to cover the entire bill. Economists then observed that there was no incentive for hospitals to compete on the basis of price; they competed for volume by offering the latest technology. The most popular variant had hospitals competing for the loyalty of physicians, who automatically brought patients with them; to attract physicians, hospitals had to offer the latest devices and programs. Whether patients were really so passive and whether technology was really so useless and wasteful as in this story was never established, but evidence was developed to show that after the repeal of the free-choice-of-providers rules, spending growth did slow differentially in those market areas that had the two crucial ingredients: high managed care penetration and high competition among hospitals.9 The cross-sectional finding of lower spending growth in more competitive markets after repeal seems quite robust. No research established exactly which technology was restricted in the high-competition, low-growth areas that was unrestricted in the areas with either provider monopoly or no managed care. Some of the most recent evidence points not to the slower diffusion of technology (or the end of the arms race) but rather to greater price discounting.10 With the benefit of hindsight, we can see that this genuine accomplishment of the managed care revolution contained the seeds for the public’s disenchantment. Discounting could not, and did not, go on forever. Some hospitals fought back against underpayment with the loss of millions of dollars and bankruptcy or near-bankruptcy. As managed care swept the field, there were no remaining fee-for-service plans (except for Medicare) to use as benchmarks for discounts. But I believe that the biggest reason for this relatively failed revolution was that it did not address the new technology that was then and is now at the heart of most spending growth, and probably is also responsible for some part of the increase in skilled hospital worker wages. So now we are back to the drawing board.
Health care markets could work to achieve an ideal rate of addition of newly developed, higher-quality, but more costly goods and services in two ways. The most obvious is the conventional model in which competitive firms invest in, develop, and sell new products. Firms that invent or discover some change in product characteristics that consumers value more highly than those in existing products can usually charge a premium price for the new product—higher than the cost of developing and making the product. These profits reward innovative efforts with "first-mover" advantages. Although other sellers might eventually be able to copy the latest fashion, the temporary profits assure some positive level of innovation, while the competitive supply assures that these products are eventually furnished at prices close to cost. In medical care, new types of surgery, new forms of wellness programs, and new ideas for disease management would fit this model. The other model is less general and less competitive but still familiar. For those novel ideas that can be embodied in a product or process, the U.S. government and most other governments permit a patent to be issued. A patent, to be blunt, grants a temporary monopoly to the applicant in return for providing a detailed public description of the innovation. This means that talk of free competitive markets in patented products is a logical contradiction. The police power of government is used to protect this monopoly, both domestically and internationally. The rationale for patents is that without them, there would be "insufficient" investment in research and development (R&D). Pharmaceutical products, medical devices, and information technology are patentable medical goods and services. In this case, the seller presumably sets the profit-maximizing monopoly price, and then consumers decide if the new product is worth its higher price. Consumers could wait until the patent expires instead of paying the monopoly price, but in the case of treatments for current illnesses, many will not. Once the patent expires, the public knowledge can be copied by competitive or generic firms. Determinants of new technology development. In deciding beforehand whether to invest in research to discover a new product, firms estimate what the monopoly profits would be if development were successful, multiply that profit by the probability of successful discovery, and compare these expected revenues with the cost of inventing and producing the monopoly quantity of the product; a rate of return higher than the (risk-adjusted) market rate of interest will trigger investment. Any product discovered and marketed under this system would be efficient, but not all efficient products would be developed and marketed. New products of high value for small or heterogeneous markets, in particular, would be ignored. That is, if anything, the volume of new products and therefore the growth in medical spending would be too slow.11 Determinants of new technologies’ market value. A key determinant of what technology it will pay firms to invent and develop is the price that consumers are willing to pay, and the qualities they are willing to pay for, at the quantities they are willing to buy. Most do not buy technology at full out-of-pocket price; instead, they have insurance, which reduces the out-of-pocket price, possibly to zero. So the usual signals of value will not be present. However, insurers can control access in a blunt way: They can refuse to cover a new technology at all, or limit its use explicitly, or put financial incentives in place for patients or doctors that apply specifically to a particular technology. Such tools, while not perfect, can impose broad limits. Plans can thus adopt different policies toward new technologies, and consumers who have a choice among plans can select them based on differences in coverage (broadly defined to include not only reimbursement but also rules, limits, and incentives) of new technology, the implied differences in the growth of premiums, and the value that the consumer places on one relative to the other. As long as consumers face premium differentials that reflect cost, they can in principle choose the optimal plan to limit (or not) the use of new technology. Some plans might permit all new technologies to be used without limit; others might limit them. Because of the tax exclusion, higher-income consumers will find expensive technology embodied in insurance coverage to be underpriced relative to true cost and would enroll in such plans to an excessive degree. This is the shining vision (or what qualifies for it) of how markets might work to choose the right rate of spending growth. Of course, to get things perfectly right, the terms and other conditions of patents also have to be exactly right (with no legal concessions for extra-long periods of monopoly power), unions have to be absent from the labor market, health professionals cannot conspire to divide up the turf or limit the number of others who will join them, other markets generally need to be competitive, and the tax exclusion needs to be capped.
Such a transparent market for technological options is not a reasonable description of today’s U.S. health care system. People cannot generally choose knowledgeably among a variety of plans characterized by explicitly different policies toward new technology. Instead, plans in most markets do not say much about their technology-rationing policies (beyond the "all medically necessary care" cop-out), and, as far as we know, health plans are fairly similar in terms of coverage of new technology. For reasons discussed later, they would probably think it market suicide to admit that they offer anything but the latest. Differences do occur in the breadth of preferred provider organization (PPO) networks and, to a lesser extent, the extent of cost sharing, which might or might not have indirect effects on use of new technology. No plan, so far as I know, explicitly emphasizes a "go-slow" approach to new technology in its marketing materials. Some large plans support technology assessment programs, but these programs, if they have any effect at all, are targeted at keeping out technologies that are useless or less effective and more costly than existing treatments. They do not profess to limit technology that improves outcomes more than any other alternatives but at higher cost, although high cost per unit of improvement can be the signal to revisit the data on effectiveness and require a higher standard of proof. There is no U.S. analogue to the National Institute for Clinical Effectiveness (NICE) in the United Kingdom, which in principle rejects or limits effective technologies on grounds of high cost but does not always do so very effectively. I should not overstate the point here. Certain health care delivery systems (especially academic ones) do have reputations of being earlier adopters of new technologies, so that the choice of such a network can be a proxy choice of technology. When high-tech networks are not priced as loss leaders, they cost more than networks of community hospitals and physicians. Still, the predominant impression is that plans generally avoid explicit discussions of technology policy and implicitly cluster together to pursue similar policies. Here are some of them. Adherence to "safe harbors" against liability. The most common explanation for the absence of differentiation among health plans is that the legal system, emphasizing standards of practice in the community as the safe harbor against liability, would seriously inhibit a plan that intended to do less than the latest to save money. Put slightly differently, it may be difficult for plans to specifically contract for a technology policy in the complex and nuanced way that they should. Clark Havighurst has made this argument most strongly.12 But Havighurst believes that "it is not clear that contracts could not be drafted to take many issues out of the never-never land of medical necessity.... Contracts could specify in general language how generous a plan intends to be in cases not otherwise specifically provided for, then a well-designed procedure using medical experts could give effect to that general commitment, thus ensuring reasonably consistent coverage decisions that reflect subscribers’ collective willingness to pay for marginally beneficial care."13 Little differentiation across patients. Another explanation is that health care providers, especially physicians, inhibit differentiation across patients because they regard it as unethical or inconvenient.14 Of course, closed and exclusive panels could avoid such constraints, but closed-panel plans have not been able to grow in the United States outside of some coastal regions. Member turnover. Still another limit on plans’ ability and incentive to differentiate new technology comes from turnover of plan members. Many of the more effective new technologies are long-run preventive in nature: For example, cholesterol-lowering drugs taken today improve future cardiovascular health, smoking-cessation programs reduce future chronic conditions, and even disease management programs usually do not generate their full benefits immediately. Although consumers have variable preferences for how they value the future relative to the present, an insurer’s attempt to offer a "high prevention" plan might be discouraged by its belief that it will not be able to capture all future benefits because of turnover. Here again, the inhibition may not be so serious. What plans will lose is any future cost savings associated with prevention; the consumer keeps the improved future health. In addition, there is evidence that offering aggressive, inconvenient preventive care that is not very effective can be a tool for selection of members with strong tastes for maintaining their health in other ways. Adverse selection. Adverse selection also might influence plans’ choice of technology, although the model of managed competition imagines that risk-adjustment mechanisms can make this problem less severe, as can employer adjustments to premium differentials and individual insurance underwriting.15 The most plausible assumption, if there were to be adverse selection, is that the higher risks would be more likely to choose plans that add technology more rapidly, so that any resulting adverse selection would drive out those "inflationary" plans compared with ones with slow premium growth that are reluctant to add expensive but beneficial technology. However, this kind of market behavior is exactly the opposite of what conventional wisdom imagines actually happens in U.S. private health insurance; instead, we supposedly get a lot (perhaps too much) of high tech. Therefore, I conclude that although adverse selection could be a reason for the existence of a smaller variety of plans than would occur in its absence, it does not seem to be a plausible explanation for what is happening with technology and spending growth. Given the recent evidence that adverse selection may not be of great importance in U.S. unregulated private health insurance markets in any case, selection alone would not seem to be a good candidate explanation.16 "Clustering" plans’ products. Finally, there are economic models of product differentiation that predict that firms will tend to cluster their products too "close together," relative to the ideal. Just as all flights on different airlines to Chicago tend to leave at the same time and all network and cable channels tend to present reality shows, so health plans may tend to pick the characteristics that appeal to those in the middle of the distribution of preferences (in the case of insurance, those of the tax-subsidized upper middle class) and to ignore the extremes. Beyond this, the economic question is whether a single insurer would in some sense be able to charge a monopoly price for coverage that rations technology. Providers may push technology (if monopoly profits are higher with high technology than with low technology). But insurers with market power will tend to resist technology. A possible solution. A plausible impediment to constraining new technology might arise from the common employer practice of offering multiple plan options at premium differentials that are not equal to the cost differential, and offering options sold by a single insurance firm rather than from multiple competing plans. This could be an inhibition if there were some plan type that was especially effective in rationing new technology but that was not available to many insurers. Prepaid group practice plans of the Kaiser type are not available through many insurers, but there is no evidence yet that they are better at limiting the growth of technology.17
So far I have discussed the abstract idea that in a perfect market, where people have different preferences about how much of their future income they would want to spend on insurance coverage for new technology, there would be many plans differentiated by their technology adoption strategies. The apparent failure of such a market to emerge thus has theoretical inefficiency costs. But there is another consequence that could have much higher costs not only for direct consumers but also for society. Insurers’ inability to offer plans with lower-than-average rates of added new technology could be contributing to uninsurance. Specifically, even if the household with the average income were willing to pay a higher premium for the new technology covered by the typical health plan, lower-income families might not be.18 The question is, why aren’t there lower-price health insurance plans with premiums that grow less rapidly? The short answer is that there are, but they often appear to be no better than no insurance at all. Insurance Web sites display insurance plans with low premiums, but those low-price plans always have one characteristic in common: high deductibles or coinsurance, often combined with a limited provider network. Premiums for such plans are much lower than those for the average plan, and the dollar amount of additional premium growth is also much less—even if the percentage rate of growth is the same. Many of these plans have now been rechristened health savings account (HSA) plans, and their lower premiums are frequently touted. But it is easy to see what the problem is: These plans cost less, but they also cover less of a person’s probable medical bills. They are not guaranteed to get the total medical bills down, at least not enough to offset the risk of higher out-of-pocket costs. Moreover, they concentrate their greatest protection on the very large expenses associated with rare but costly treatments—treatments so costly that even moderately well-off families might expect to receive charity care if they were to use them and to avoid using them if no such care was expected.
The news is not all bad. For one thing, as shown in Exhibit 1
If the majority of insurers continue to offer costly, undifferentiated plans, spending will continue to grow. At this point I cite the looming alternative: government takeover with spending limits and unvarnished exercise of monopsony power, the "brute rationing" that is so feared. The other horror story is a spreading of uninsurance or plans with too-meager coverage. Even if we cannot get a perfect system, are there things that can be done—by public policy or by honest entrepreneurs—that are better than these alternatives? Popular remedies with lower chances of success. The most popular current panacea for what ails the health care system is information technology (IT). Certainly, IT improves some outcomes and lowers some costs. And certainly, better and more timely information should help with any management task, including making difficult decisions about new technology. To the extent that limitation requires custom tailoring of coverage approvals in real time, for example, or coordination across multiple providers, quicker and better information will help. Likewise, easy, cheap, and convenient access to information could help patients understand and accept limitations more willingly. Still, it seems likely that in its current state, IT is aimed more at current clinical and payment practices and not specifically at new technology. Perhaps this could be changed. The other popular remedy is (somehow) to improve quality of care as defined by the Institute of Medicine (IOM), which is thought to be able to reduce costs.19 This assumes that overuse is a bigger problem than underuse or that increasing some underused care today will save costs tomorrow. I am skeptical about the direction of net change, but my main point is, as before, that this effort does not seem to be targeted at the addition of new technology. Typically, interventions are prioritized based on their level of spending (not their rate of growth), and most of the anecdotes of cost savings (which are all we have) apply to decades-old procedures, some of which physicians still fail to use properly. Here again, one could conceive of a "new quality improvement movement," but it has yet to be proclaimed. Again, perhaps it should be. Some promising strategies. The improvements about which I am more optimistic fall into two categories: improvements in the information needed to manage new technology; and improvements in methods to bring about consumers’, patients’, and politicians’ acceptance of limits on new technology. (1) Each health plan will need to make choices about how it will manage new technology. The principle of comparing benefits and costs is embodied in cost-effectiveness analysis, whose conceptual framework is by now reasonably well developed. Evaluation of a new technology as a whole, or its provision to sets of patients with different illnesses and other characteristics, requires only good data on additional costs and on health outcomes. Then a health plan could target a consistent strategy, and it might be able to explain a consistent strategy to consumers. For example, if it measured outcomes by the common quality-adjusted life year (QALY), a consistent strategy would be to set a benchmark value of dollars per QALY and then adopt all new technologies with costs below that level and none above. Setting a lower threshold would yield a lower rate of growth in spending; plans could therefore vary based on what level they chose for their threshold. Such a rule might qualify as the "well-defined procedure" suggested by Havighurst and might be a good way to describe the strategy to consumers.20 An alternative would be a bottom-up strategy in which the plan set a target level for spending growth and then used cost-effectiveness analysis to choose the set of new technologies whose cost fit within the limit and which maximized the number of new QALYs delivered. At the limit, a technology would be in the package only if its value of dollars per QALY were lower than that of all excluded technologies. I believe that this way of defining "basic coverage" is better than the more common alternative of calling health professionals together to negotiate what they think is needed, which tends to include the services they provide. (2) A policy of delaying new technology until the bugs are worked out and the prices fall might make sense on its face and could have helped with Vioxx (although not with Lipitor). Likewise, a "no new heroic measures" health plan might align some consumers’ preferences for end-of-life care (where much of the most costly new technology is used) with incentives for providers. There may be ways to add appeals, counseling, partial coverage for lower-value technology, and point-of-service options; creativity in crafting language will be needed. Unpacking the aggregate. Another improvement would unpack the U.S. government’s aggregate measures of new technology from direct measures of spending on new technology. Spending on brand-new products in the government’s statistical tables is now presented as part of a statistical residual—what is left of annual medical spending growth after price and demographic changes are factored out. By definition, no one knows exactly what is in the residual. We do know that it includes some wholly new products, but we also know anecdotally that it includes changes in medical practice that involve greater use of existing products. The growth in the use of cholesterol-lowering drugs because of recommendations for more aggressive treatment is probably the most prominent example. Probably, too, some of the residual is greater use of things that do little good. Recommending research is the obvious thing to do here. The heterogeneity of technical change also means that health plans that tried explicitly to announce their policies would find it complex to do so; they could not reasonably pick one or few technologies and expect to communicate information about them to the general public. The cost-effectiveness measures I discussed earlier are one way to aggregate information about aggregate behavior in a more meaningful way. It is likely that consumers (as opposed to investors and analysts) will not be especially concerned about whether a plan improves their health with just-invented products or with just-invented ways to make better use of old ones. But it surely is true that the residual, imprecise character of our measures of "technological change" is one of the reasons for the lack of transparency in our understanding of medical spending growth and the consequent lack of confidence in value.
Admit that access needs to be limited. The most useful thing that policy-makers could do is to stop posturing about finding the unattainable "higher quality at lower cost" and frankly admit that beneficial new technology will need to be limited in some fashion. IT, evidence-based medicine, and pay-for-performance all appear likely to be fine things, but they cannot deal with the inexorable addition of technology for more than a few years. Efforts to do this, whether in the form of the Oregon experiment or former Colorado governor Richard Lamm’s crusade against the myth of optional immortality, have not been successful. Someone has to try again. An optimist must admit that just as with old technology, some new technology may be of little or no value. But there is no evidence that enough new technology is truly marginal to make a serious dent in the rate of spending growth. Create safe harbors. Most obviously, some type of legal safe harbor has to be created for insurers that implement well-designed plans for limiting technology, and the "community or standard practice" and "medical necessity" concepts need to be jettisoned. The information needed to set up and operate the well-designed plan is a public good. Government has already taken some steps in planning to collect more evidence on the effectiveness of medical care for different patient subgroups; data on overall cost effects have as much claim for public support. Design plans to prioritize and limit spending growth. Either the government or foundations might sponsor efforts to design basic coverage plans that can prioritize and limit new spending growth and to conduct demonstrations of how they might work. This might even be done initially on a hypothetical basis, determining what new technology would not be provided and estimating the cost savings and the effects on health outcomes. Consider differential rationing. Finally, and most seriously, we need to develop a discourse on differential rationing. The challenge is enormous. Not only must we be able to explain to average citizens why they cannot get all of the care that will do some good, but we need to explain why the limits on new technology should be tighter for some people than for others. Probably permitting tighter limits for those at the same income level who value other types of consumption relative to health care will not be so objectionable. What will be objectionable is advocating less care for lower-income people. Logic says that without subsidies, there should be multi-tier medicine, but it is a hard logic to accept. That this policy stems from a maldistribution of income that society seems to tolerate may not be much consolation, but forthright consideration of this matter might prompt a willingness by morally embarrassed, better-off taxpayers to provide more support for low-income health care and for income redistribution in general. I should not end on a gloomy note. Almost the whole reason for the "problem" of health spending growth is the long-run success in discovering ways to improve the quantity and quality of life. If we returned to the medicine of the 1960s, defunded the National Institutes of Health, or abolished patent protection for drugs, we could control explosive health spending growth. But we probably should not do these things. Having to face trade-offs between better things is preferable to no trade-offs at all. But dealing in a forthright way with the future path of this effort is surely important, and rejuvenated markets with relevant health plan choices could help a lot.
Mark Pauly (pauly{at}wharton.upenn.edu) is the Bendheim Professor in the Department of Health Care Systems at the Wharton School, University of Pennsylvania, in Philadelphia. An earlier version of this paper was presented at "Health Care Market Competition: How Well Can It Work?" at Lansdowne, Virginia, 29 April 2005. The meeting was cosponsored by Health Affairs, the Kaiser Permanente Institute for Health Policy, and the Center for Studying Health System Change.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
How It Was Supposed...
The Vision: How Competitive...
