The following is taken from the First Plenary Session, “Balancing Affordability and Value: The Universal Challenge in Health Care Delivery,” presented at ISPOR 11th Annual International Meeting, May 2006, Philadelphia, PA, USA.

One approach to framing the issue of balancing affordability and value with respect to health care is to define the dimensions of interest and place them in priority order. While the initial framing of the question appears to focus on optimizing health care allocations and investments, the real purpose is to maximize health. While implicitly acknowledged, we have not explicitly agreed that this is our national priority. Quality health care is a secondary priority - as it is a critical element in enhancing the health of the population. The third priority is value - getting the best "bang for the buck"; the fourth priority is affordability - not spending "too much" (however that is defined).

While we have not, as a society, normatively defined what our targets are for each of these four dimensions, we do know that our targets need to move. The United States is well below international norms (as determined by the WHO) with respect to measures of population health for countries with similar socio-economic standards of living. Indeed, recent data from Marmot and colleagues (JAMA 2006;295:2037-45) suggest that the richest third of Americans are not as healthy as the poorest third of the British, even though access to health care is arguably much better for rich Americans; thus health care is but one of many determinants of population health . We also know - based upon the work of McGlynn and colleagues (NEJM 2003;348:2635-45) - that half the time we fail to deliver evidence-based care. Moreover, as quality measurement has become widely adopted over the last 15 years, we can expect that our ability to assess health care quality will only improve in the future and therefore our expectations regarding health care quality will increase. Today we largely focus on process measures; in the future - with the availability of inter-operable electronic health information systems - we will be able to assess true outcomes.

But today, without the ability to measure outcomes and given our limited understanding of the "health care production function," i.e., the incremental impact on population health of improved health care, it is difficult to know whether we are improving or enhancing value when we make additional investments in health care. Affordability is even more difficult to assess given U.S. cultural resistance to setting limits on access to care for insured individuals. And what about the uninsured? How can we achieve a goal of optimizing U.S. population health when more than 40 million Americans do not have coverage for essential preventive, diagnostic and therapeutic services? At the end of the day, what can we really afford? Is there a limit?

The next question to be asked is "What are health care payers and providers currently advocating and/or implementing and how does this match up with the four priorities?" A non-exhaustive list of current strategies that have been the focus of much attention would include: disease management, pay-for-performance, health care information technology, and new benefit designs (tiered co-pays, co-insurance, and consumer-directed plan designs). The first three strategies are designed to enhance health care quality. The last focuses on improving value and affordability. While some have argued that disease management and HIT may also improve affordability and value, data are limited and I suspect that the return on investment will be difficult to demonstrate across broad populations. Indeed, some have argued that disease management - originally promoted as a quality improvement strategy - was appropriated as a cost-containment tool. Current evidence suggests that disease management may decrease costs for certain populations (e.g., high risk patients) with specific conditions (e.g., congestive heart failure) but it is unlikely to be a panacea for rising health care costs.

Pay-for-performance can provide incentives for improving health care quality but issues remain - Where does the money come from? Who will pay more for higher quality? We do know that what you pay for gets done; therefore P4P should focus on important issues where there is a clear linkage to health outcomes, though whether that can best be accomplished through P4P or suitable payment for those services is open for discussion.

Newer benefit designs were largely created to address the issue of "moral hazard" - i.e., overutilization occurs when services are free or too inexpensive relative to their value. Without addressing the merits of this point of view, we do know that financial incentives are blunt instruments. Thirty years ago the Rand Health Insurance Experiment taught us that elasticity of demand operates in health care as everywhere else - if you raise the price of health care services, the utilization of both essential and nonessential services will decrease. A number of recent studies show that lower adherence and higher medication discontinuation accompany patients' increased out-of-pocket costs- who are also "consumers." Such behavior can lead to worse health outcomes and can lead others to delay seeking necessary medical care. Thus the overall impact of increased cost-sharing with patients is unclear as short-term decreases in "drug spend" may be offset by increases in future medical costs.

What I have addressed thus far can be summarized as follows: Our ultimate priority is to improve population health. While improving healthcare can contribute to improvements in health, it is not the only determinant. Therefore we are in a difficult situation in trying to assess the value of investments in health care quality. The four major developments in recent years that have impacted health care delivery have yet to prove that they improve health care quality, value, or affordability.

What I want to turn to now is another dimension of value - increasing health care quality through scientific advances in therapeutic interventions. Advances in medical technology have contributed greatly to the longer lifespan we now enjoy and the enhanced quality of life enjoyed by aging Americans. While the Congressional Budget Office and others have calculated that the benefits of new treatments have more than outweighed their costs when considered from a societal viewpoint, increasing costs are the subject of much consternation by patients and payers. Should health care costs continue to increase, driven in no small part by innovative diagnostic and therapeutics, payers (including consumers) may exhibit a decreased "willingness to pay." While the increased use of generic medications will provide some headroom for innovation - and there many "modern generics" that will be entering the marketplace in the next several years - payers are increasingly using formal health technology assessments to "raise the bar" for obtaining favorable coverage of new drugs. They are asking the following questions: "Does a new therapy really provide advantages over current therapy?" "Does it provide a good value?" These are reasonable questions and represent a shift from the latter 20th century when the questions were: "Does a new therapy provide significant benefit relative to any potential harms?" "Are there patients who would benefit from its availability?"

This provides a new challenge to pharmaceutical innovation. Investment in pharmaceutical R&D is sensitive to the prospects for commercialization which is in turn increasingly influenced by payer coverage decisions. Indeed given the long lead time between discovery and marketing, to make informed investment decisions manufacturers need to understand where the "goal posts" will be 10 years down the road. What is needed is for "rules of the road" to be developed that are acceptable to payers, providers, patients, and pharmaceutical companies. We at Merck & Co., Inc. embrace the transparent and appropriate use of evidence-based decision making as one of the "rules of the road."

To understand what I mean by this, it is important to distinguish between and understand the dynamic relationship of evidence-based review/synthesis and evidence-based decision making (as discussed in the editorial I co-wrote with Steve Teutsch (Med Decis Making 2005;Sept-Oct.: 487-9). [Figure 1]. An evidencebased review and synthesis is a critical review of the information regarding the benefits, harms if any, and costs associated with a therapy. This is an evolving and special discipline that answers the questions "What do we know? and "How certain are we about what we know?" It integrates basic, clinical, and economic information in a structured and a priori fashion guided by key questions and an underlying model of disease. It may aggregate data into evidence tables and/or include meta-analyses. Its focus is on the scientific evidence and may employ modeling to assess the economic impact of therapeutic choices with cost-effectiveness analysis. Evidence-based decision making takes these inputs on effectiveness, safety, and economic impact and interpolates it - in a transparent fashion - with other considerations including values/preferences, equity, acceptability, and budget constraints. When optimally performed, with adequate stakeholder involvement in a deliberative and transparent fashion, the basis for coverage recommendations are clearly understood and the fairness of decisions are more likely to be acceptable to key stakeholders, including those who may be disadvantaged by a particular decision. This is what Norman Daniels has called "accountability for reasonableness."

The use of formal evidence-based reviews and decision making remains controversial across the pharmaceutical industry, in part because of the manner in which they have been conducted. Some view this as a method to justify cost-cutting decisions. Greater acceptance would follow from insulation of individuals conducting the evidence reviews/syntheses from those making decisions; this would minimize bias and conflicts of interest. Transparency is also critical to acceptance. Reviewers and decision makers must inform stakeholders about what drives their deliberations and considerations; scientific evidence must be distinguished from social science evidence (e.g.,. economic models and resource constraints) and colloquial evidence (e.g., values, precedent, professional opinion,). Performed in this way, I believe that decision making will be improved and that a path forward can be found that balances the goals of affordability and value. It's in all of our best interests to make this work.