This volume presents an enormous amount of information that will take students of consumer-driven health care a very long time to read and digest. It will be tempting for both advocates and opponents of the movement for greater consumer control to browse through the work and pick out and trumpet those nuggets of information that suit their predispositions.
This would be unfortunate because the information that runs counter to our biases is the most important information to understand. Good policy can be developed only when we listen closely to honest criticism and respond accordingly—as difficult as that may be.
Still, the work presented here requires some context. Consumerism in health care is in its infancy. We do not yet know what the optimal approach is and we are in a period of experimentation and trial and error. Like most other new ideas, the initial models will need to be revised and improved. Prototype designs are almost never without flaw.
One of the marvels of any market-based system is the ability to make those corrections and revisions quickly as more information becomes available.
Too many health policy analysts take a governmental program approach to design questions—the model must be irrefutably effective before it is ever implemented. Once a program is “the law of the land” it is nearly impossible to change. Witness the protracted debate over adding prescription drug coverage to Medicare.
Fortunately, consumer-driven health care (CDHC) was born in the market and will be revised in the market. To the extent there has been governmental involvement (such as the IRS guidance on Health Reimbursement Arrangements), it has been extraordinarily flexible and permissive.
Vendors and employers are free to refine their products in accordance with changing conditions and growing knowledge. In that context, identifying problems is seen not as an attack on cherished ideas, but as a welcome opportunity to improve the product offerings. Criticism is valued as product feedback. A company that wants to succeed in the market is eager to hear what the problems may be.
Market approaches have some other advantages over a governmental orientation, as well. Government programs are essentially political. They are aimed at pleasing 50 percent+1 of the population. Opinion surveys are conducted to see how close a new idea is to achieving that goal.
Few companies in the private market think in those terms. If a new product or a new company feels it can reasonably attract even just 10 percent of a market, it views the prospects as very promising. Hertz is not the only success in the rental car business. Avis and National and Budget and Alamo and many others manage to succeed without being Number One.
Readers of the papers in this volume will likely conclude that the experience at Humana was not very favorable, the experience of the Definity-covered University of Minnesota was more favorable, and the large, unnamed Definity-covered employer was very favorable. What does that mean? Clearly different locations and different designs lead to different results. If CDHC were a government program, this might be worrisome—have we chosen “the right” model? But because CDHC is a market-oriented approach, it is not discouraging at all. Definity is doing something right and will build on it. Humana may revise its approach or drop the program altogether. It does not matter in the slightest. Humana is not disadvantaged because Definity is succeeding. And Humana's problems do not detract at all from Definity's success.
Certainly there are things to be learned in both cases, and market-oriented companies will study these experiences closely. But no company—including Humana—is stuck with a problematic design. Humana's product did not allow rollovers and the funds in the “allowance” could be spent only on in-network provides and for covered services. These features remove the most promising elements of consumer-driven health designs—consumer choice and the opportunity to save money for future needs. It is simple enough for Humana to incorporate those features in its next round of offerings.
Market-oriented companies also know that early adopters are different than the rest of the market. The people who are the first to sign up for a new product or service tend to be risk-takers. They accept risking the unknown for the privilege of trying something new. They also tend to be younger and better educated than the rest of the market. They volunteer to be “test cases” and product developers rely on them to refine their offerings. People oriented toward government programs may view this as a selection problem, but innovators expect this to occur in the first couple of years of new-product roll out. If the product is successful at this stage, word gets out and the new idea attracts a wider market segment.
The research presented here does not address the “early adopter” phenomenon very effectively. We are told that the enrollees in the Humana program tended to be actuaries and financial service personnel. These individuals are presumably better educated than most Humana employees, and they certainly know their way around a benefits program better than the average person. It is interesting, for instance, that the studies report no end-of-year rush to consume unspent dollars in the allowance, even though Humana included a use-it-or-lose-it provision characteristic of flexible spending accounts (FSAs). This contrasts with the Countrywide Financial experience that did have an FSA-type year-end rush, even though those employees were able to roll over unspent balances. It is possible that the self-selected Humana employees understood the dynamics of forfeited balances and did a better job of spending their money through the course of the year, while less-savvy Countrywide employees stuck to their FSA-induced spending habits.
Most of the studies report income disparities between CD-selectors and nonselectors. It will be interesting to see if this difference continues over the years, but it is also possible that income is a proxy for education. This should certainly be the case at the University of Minnesota where educational attainment should correlate closely with income. If it is true that early adopters tend to be more highly educated, we would need to control for differences in education before concluding there is an income effect unique to CD health.
We also think of early adopters as being younger, but that does not seem to be the case here. If anything, CD-selectors appear to be somewhat older than nonselectors (though age is another underreported variable in these studies). Is it possible that early adopters for electronic gadgets are different from those for health insurance programs? Perhaps younger people pay so little attention to their health care needs that a choice of benefits plan is of little interest to them.
Since we cannot yet distinguish between the behavior of early adopters and a more mature market for consumer-driven health, the research presented here is of limited (but not unimportant) value. Most of this work looks at baseline information in 2001, first enrollment in 2002, and renewals in 2003. That means there is only a single year's worth of data. Given that the IRS did not issue guidance until June 26, 2002, the products were very tentative and in some cases did not incorporate the more attractive features of the approved health reimbursement arrangement (HRA) model. It was not at all clear at the start of 2002 that the IRS would allow year-to-year rollover and buildup of unspent balances.
Even more importantly, none of the pioneer models anticipated that postemployment access to the funds would be allowed. The prospect of saving money for future needs even after leaving one's current employer could very well skew enrollment decisions from what this research presents. Lower-income workers in particular might find that prospect more attractive.
The body of research presented, then, is looking at a moment-of-time of an extremely fluid and dynamic environment. Much of the experience studied predates the IRS guidance. And, while Humana and Definity are both very serious and credible players, they are not the only vendors, nor the only models available. Destiny Health, for instance, takes a radically different approach to the market and to product design. It distinguishes between “discretionary” and “nondiscretionary” spending and applies the cash account only to the former. It also requires portability for account balances. Its market targets fully insured smaller companies, rather than the larger self-funded employers studied in this research. It would be worthwhile knowing the experience of this different design and different market segment.
It is impossible to know ahead of time if the Destiny model is superior to the Definity model or the Humana model (or the models from Aetna, HealthMarket, Lumenos, or dozens of other variations). Clearly, behind each design are a number of credible and serious people who believe their approach is superior to all others. It will not be academia that answers the question of which approach is best, but the market.
Another example of the limitation of the research is the role of consumer support. Critics have complained that good comparison data does not yet exist, so it is difficult for individuals to become smart shoppers in the health care marketplace. That is unquestionably true—at this point in time. The research in this volume touches on what information services and customer support were available during the study period, and it all seems pretty rudimentary.
But 10 years ago the Internet was rudimentary, too. One thing we have learned beyond doubt is that information systems explode once the right incentives are in place. It is probable that the support services available to companies buying consumer-directed plans in the winter of 2004 have evolved considerably from what were available in 2002. We cannot begin to imagine what health systems information may look like 10 years from now. All we really know is that for the first time in history, individual consumers of health care services have a reason to demand having reliable and accessible information and the tools to make use of this information.
The context of this information revolution is important. For at least two decades policymakers have bemoaned the lack of quality incentives, patient education programs, transactional efficiency, price competition, and so on. We have created massive government agencies, behavioral modification programs, public service announcements, efficiency initiatives, and health education efforts. We have Institute of Medicine studies, and those from the Agency for Health Care Research and Quality, and the Leap Frog Group. We have worksite wellness programs, quality assurance requirements, and certificates of need. All an endless string of well-intentioned badgering at a system that is largely indifferent. And still we have epidemic-sized obesity problems, diabetes, smoking, HIV infections, physicians writing prescriptions in illegible handwriting, and massively inefficient hospitals. For all of the effort invested, nothing we have done has been very effective.
Consumer-directed health care supposes a new formulation—one driven by consumers with cash-in-hand, demanding to know for themselves who is the best urologist in town, what are my treatment alternatives, why is this hospital billing so much for a Tylenol, why can't I read this prescription, where is the nurse when I need one, how do I get the most value for the money I'm spending?
Information systems to support this movement will grow exponentially. But the information is only ammunition. It is not an end to itself. The real revolution will come when health care consumers use that information to reward higher quality and punish the mediocre, to demand efficiency in the use of their health care dollars, to educate themselves about their treatment alternatives and become invested in the decisions they have made, and to learn that their own behaviors are what drives their need for health care services.
Nothing we have tried in the past has accomplished this transformation. If we keep repeating the same old patterns, we will keep getting the same old results. Consumer-driven health care gives us an opportunity to change the pattern.
But it will take a little patience to get there. Already the environment has changed dramatically from what was in place in early 2002 when these programs started. As mentioned, the IRS put its imprimatur on HRAs in June of 2002. Legacy companies like Aetna, Cigna, United Healthcare, and even the Guardian have entered the market, bringing legitimacy and marketing clout to the movement. Costs for traditional coverage continues to rise and employers everywhere are increasing cost sharing with employees. The IRS now allows FSAs and other cash accounts to pay for over-the-counter drugs and weight-loss programs, and to use debit cards. Even more recently (January 1, 2004) Health Savings Accounts have been made available to all 250 million nonelderly Americans.
The coming twelve months could see additional changes, such as refundable tax credits for people who do not get coverage on the job, FSA rollovers, and possibly some form of association health plan or joint purchasing for individuals and small employers.
By all means, let the research continue. Let's dig deep into the experience we've had and learn as much from that as we can. But let's also understand that the process of research necessarily means looking backward into what has already happened. We are white-water rafting here and the river changes by the minute. The experience of two years ago is important, but it is already out-of-date.