Evidence-based medicine (EBM) is based on the principle that providers and patients should adopt those medical treatments that scientific clinical trials find to be safe and efficacious. It advocates that clinical practice should be grounded in scientific evidence and continually improved by the results of robust research. However, the articles in this FORUM demonstrate that the process of adopting and diffusing health care innovations is far more complex than EBM proponents may have assumed. While EBM is desirable, it does not sufficiently explain the pattern of adoption and diffusion of many health care innovations. There is a significant and unexplained asymmetry between the innovations supported by research and the innovations adopted by clinical practice. Why are some health care innovations, though apparently lacking strong scientific support, widely adopted, while others with solid scientific evidence remain underused?

The rationale for EBM is theoretical argument. Researchers believe in the scientific method. Sound premises and rigorous testing, the principle of falsifiability, and the replication of studies: these are the things that convey legitimacy to researchers. But one person's sound argument may leave another unconvinced, especially when each side brings its respective beliefs and assumptions to the table. Given the range of backgrounds of the various players in health care innovation adoption, we can easily see how their beliefs and assumptions might diverge commensurately. How then can proponents of EBM make their innovation arguments more broadly convincing?

What makes information convincing and, therefore, utilized is a rhetorical question. Rhetoric is the use of persuasion to influence the thought and conduct of one's hearers (Random House unabridged dictionary). Aristotelian rhetoric includes a moral component lacking in the rigorously logical science of dialectics. To Aristotle, the art of persuasion comprises three elements: (1) logos-the message, especially its internal consistency (i.e., the clarity of the argument, the logic of its reasons, and the effectiveness of its supporting evidence), (2) pathos-the power to stir the emotions, beliefs, values, knowledge, and imagination of the audience so as to elicit not only sympathy, but empathy as well, and (3) ethos-the credibility, legitimacy, and authority that a speaker both brings into and develops over the course of the argument.1 Logos, pathos, and ethos, the foundations of the rhetorical triangle, are highly interrelated and together shape the persuasiveness of any communication. Using the rhetorical triangle as a framework for interpreting the case studies in this FORUM on health care innovation adoption and diffusion, we can identify the strengths and weaknesses in the rhetorical elements of EBM.

Perhaps the most obvious lesson we can take away from the rhetorical triangle is that scientific evidence is but one component of persuasion. Health care innovation adoption is not only a function of the argument, but also the credibility of the proponents and the values, experience, and interests of potential adopters. Innovations are more likely to be adopted when a convincing argument (logos) is presented by credible research proponents (ethos) who stir the interests, needs, and emotions of adopting practitioners (pathos). In an ideal world where EBM researchers and practitioners are persuaded by similar sources of evidence, influence, and credibility, innovation adoption is not a problem. But, as the articles in this FORUM demonstrate, the adoption of health care innovations in the real world is seldom so unproblematic. In more realistic and complex situations innovation adoption is often dependent on communications and negotiations among divergent stakeholders (researchers, developers, and users) who are persuaded by differing forms of evidence, are influenced by different benefits and costs associated with this evidence, and place credibility in the testimonies of different authorities. Moreover, as Denis et al. found, the forms of evidence, influence, and credibility for an innovation may develop unevenly by leading or lagging the other over time. As a result, the process of innovation adoption and diffusion was reported to vary from no discernable pattern (Dopson et al.) to one that reflected the political dynamics of a social movement involving ongoing conflict, contestation, and struggle (Maguire).

The rhetorical triangle is useful for disentangling some these complex dynamics. It provides a framework to reinterpret the case study findings in terms of the following propositions on the relative influences of logos, pathos, and ethos on innovation adoption and diffusion.

Effects of logos (strength and consistency of logical argument). When the logical argument is clear, when evidence is strong, and when reasoning is consistent, there will follow greater understanding and agreement among stakeholders, and the potential influence of logos on the adoption process will be greatest. This proposition has been the guiding assumption of the EBM movement. Conversely, as the case authors imply, the greater the uncertainty of scientific evidence, the greater the potential for different interpretations of the evidence, and the more influence pathos (particularly the values and interests of user groups) and ethos (the credibility and legitimacy of spokespersons) will have on the adoption process.

Effects of pathos (appeal to values, beliefs, and experience). The greater the divergence in values, beliefs, and experience between innovation stakeholders (researchers, developers and user groups), the more the innovation proponent must address these concerns of divergent stakeholders in the argument. To the extent that they feel their own values, beliefs, or experiences are underrepresented or misunderstood, stakeholder groups will turn to alternative forms of evidence (scientific, experiential, craft knowledge) and rely on the credibility and legitimacy of different authorities. As Dopson et al. point out, in this pluralistic situation there is no such thing as "the evidence." Instead, there are several competing bodies of evidence that are subject to multiple interpretations by different stakeholders. Different groups have different hierarchies on what constitutes credible evidence, and they tend to rely as much on trusted colleagues and trusted mentors as on scientific evidence for guidance.

Knowledge and data tend to be "viscous," and do not readily flow across group boundaries. Dopson et al. observed very few formal, regular forums in which groups meet to share perspectives and debate evidence. Yet, Lemieux-Charles et al. found that intensive communications between diverse groups over an extended period of time were necessary to build interorganizational networks for EBM change. They conclude that networks with an existing infrastructure and/or resources for network development were more successful in building knowledge that impacted the diffusion of the CCS innovation. In general, Denis et al. propose that the more the benefits and risks of an innovation map onto the interests, values, and power of adopting stakeholders, the easier it is to create a coalition for adoption and the faster the adoption process.

Effects of ethos (credibility). Relying on the testimony of credible authorities or advisors provides a pragmatic way to compensate for ambiguous evidence. The lower the credibility of innovation proponents, the more partisan interest groups will mobilize to challenge and replace not only the legitimacy of the evidence, but also the rules of evidence for an innovation. The interesting case study by Maguire shows that AIDS activists successfully transformed both the adopting system and the evidence-producing system. Community-based vs. "academic" medicine are two different and important forms of evidence for AIDS treatments. Academic researchers sought scientific evidence to generate universal knowledge on whether or not a treatment works. Clinicians and activists adopted a more pragmatic approach to determine whether a treatment worked in the messy real world of heterogeneous patients, poor compliance, and confounding effects of other treatments.

As Maguire states, the AIDS case supports a political model of adoption, involving ongoing conflict, contestation and struggle. It focuses on institutional change in the rules within which innovations diffuse. Promoters of claims and counter-claims competed for supporters and allies not only in contests of evidence ("fact-making" according to Maguire), but also in institutional "rule-making" contests over the legitimate use of evidence. Maguire points out that such struggles and their outcomes are intimately connected because the production of evidence and creation of institutional rules for an innovation occurs simultaneously. Knowledge and power are inextricably linked. Politics concerns the creation of legitimacy for certain ideas, values, and demands. Activists exploited latent conflict and critiques already present but dormant in the existing institutional status quo. Insightfully, Maguire notes that the legitimacy of evidence claims and credibility of institutional rules "appear to be subject to increasing returns." This is because they are simultaneously stake and weapon in struggles; those who emerge as winners acquire more legitimacy and credibility and are then more likely to win subsequent battles. Having one's claims accepted better positions actors to have their subsequent claims accepted because they become accepted as "legitimate," "authoritative," or "possessing expertise" (Maguire).

Multiple triangles and nested triangles. The rhetorical triangle (or indeed, any rhetorical model) begins with an assignment of roles: speaker and audience, writer and reader, persuader and target of persuasion. The rhetorical situation of health care innovation adoption and diffusion is not such a well-delineated forum. Each side plays both roles. Researchers, providers, patients, and advocates, each by turn will attempt to persuade the others. Each side will listen and react, formulating and reformulating its own argument as it learns from and is persuaded by the rhetoric of the others.

Complicating matters further is the nested nature of these triangles. Before researchers take their place as persuaders, invoking their own ethos and evoking the pathos of their audience, they are first themselves persuaded by the very same elements. The scientific method calls for observation (pathos), hypothesis (logos), and repeated testing of predictions (ethos). As Dopson et al. found, professional networks play a crucial role in persuasion. Physicians talk to other physicians, providing one another with a strong pathos element by virtue of shared experience and beliefs. Patients, too, can be vocal in their arguments, as is evidenced by the numerous health care rights initiatives promoted by advocacy groups. Thus, individuals persuade and inform institutions.

This multiple and nested nature of the rhetorical relationships in health care innovation diffusion form a complex web of interactions. Individuals inform and persuade the institutions to which they belong and the researchers, physicians, and patients seek to inform and persuade each other. It is important to note that even this represents a simplification of the situation, as individuals belong to and so create multiple institutions. A physician, for example, informs and learns from not just her health care provider system, but also the medical field in which she practices and the professional organizations of which she is a member. Similarly, the researcher belongs to the governmental regulatory institution, as well as the scientific community, insofar as his studies inform the F.D.A.

"EBM as logos"versus"EBM as complete triangle." Through the evidence presented in the articles of this special FORUM as well as the insights provided by the rhetorical triangle framework, we can see that EBM, as it is currently articulated, is insufficient as a vehicle for moving health care innovation from research into practice. To be successful at this mission, proponents of EBM must, as Aristotle exhorts, have the ability to reason logically, understand human character and goodness in their various forms so as to be a creditable witness, and understand human emotions so as to better appreciate the beliefs and experiences of others.

REFERENCE