INTRODUCTION

The diffusion and adoption of innovations are old problems that have a large classic literature associated with them.1 The topic has always been of particular interest to the health care field. Classic diffusion studies have generally looked at predictors of innovation and adoption rates or developed models of the processes in terms of contagion or phases.2 This special FORUM, based on a symposium developed for the 2001 Academy of Management meeting in Washington, D.C., takes a new look at the topic, stimulated by three issues that have come to the fore in recent years:

1. The evidence-based medicine/management movement: The ambition of this influential movement is to introduce rationality into the adoption and diffusion of innovations.3 The promotion of evidence-based medicine draws attention to questions of how scientific knowledge influences professional and organizational practices. It also raises critical issues concerning the nature of evidence and the applicability of scientific knowledge in different contexts.

2. The nature of innovations: The boundaries of innovations are not as easy to define as often supposed. This is particularly the case for many of today's health care innovations that have both clinical and organizational content that may take a variety of shapes. Christensen et al.4 claim that many recent health care innovations are also more "disruptive" than those of the past, with potential to question the entire health care industry landscape. In addition, as several writers have suggested, there is a great need to look at new practices in a less pro-innovation value-loaded manner.5

3. The emergence of new conceptual perspectives: Traditional theories of innovation saw the definition of innovations, evidence, and organizations as fixed and unproblematic. Newer perspectives recognize the fluidity of boundaries among social contexts, objects, and knowledge. For example, recent work on information technology has drawn on ideas from structuration theory6 and actor-network theory.7,8 These perspectives are also relevant to health care innovations and organizations. They imply a need for longitudinal and contextualist research methodologies to generate an understanding of the processes that shape practices, evidence, and/or organizational arrangements over time.

This FORUM presents four different empirical studies that take into account these issues. All use qualitative case study methodologies and examine in different ways the role of "evidence" in the processes of adoption and diffusion of health care innovations.

The first study by Dopson, Fitzgerald, Ferlie, Gabbay, and Locock pools data from seven recent projects on the diffusion of health care innovations by two different research teams in the U.K. The authors draw on an impressive database of 49 cases and a total of over 1,400 interviews, taking up the challenge of developing cumulative insights across multiple qualitative cases and teams. Their article focuses specifically on the nature and role of "evidence" as used in practice, drawing attention to a number of issues that need to be considered in any attempt to improve clinical practice. For example, they observe that the creation of evidence ("fact-making" according to Maguire-see below) appears to be a social as well as a scientific process, that different professional groups need different kinds of evidence, that this hampers communication across professional boundaries, and that context is a poorly understood mediator of adoption patterns.

The second article by Lemieux-Charles, McGuire, and Blidner is based on an in-depth study of a major initiative for the improvement of stroke care in Toronto that involved multiple organizations and professional groups and complex organizational and clinical issues. The study followed the patterns of diffusion and adoption in four pilot sites. As well as identifying many of the same issues concerning the nature and role of evidence noted by Fitzgerald et al., this article is particularly interesting for its observations of the role of interorganizational and interprofessional network structures (committees, working groups) in diffusing the stroke strategy and in acting as a focal point for both problem-solving and the discussion, appropriation, and confrontation of various forms of evidence. Nonaka's9 theory of knowledge creation and transfer is found to be particularly helpful in understanding the diffusion process. The article further reveals the complexities of implementing an integrated program that has three components of varying degrees of ambiguity where different strengths and types of evidence are associated with each component.

The idea that many health care technologies may have both hard (well-defined) and soft (ambiguous) elements that affect the adoption and diffusion process is pursued in the following paper by Denis, Hebert, Langley, Lozeau, and Trottier, based on case studies of four innovations. These authors focus specifically on the social context for innovation, arguing that the way in which the distribution of benefits and risks maps onto the interests, values, and power distribution of the adopting system (organization, interorganizational group, set of professionals) is critical to understanding the way in which innovations diffuse. The softer aspects of the innovation provide opportunities for the negotiation of several different pathways. Denis et al. see "scientific evidence" as one of several potentially legitimating tools that contribute to this essentially political process.

Finally, in the last article, Maguire takes a look at innovation in the field of HIV/AIDS treatment. Medical innovations are usually assumed to follow a logical sequence in which an innovation is first developed, then tested to produce evidence of its efficacy, and finally implemented in practice. In the case of HIV/AIDS, Maguire's analysis shows how the traditional logical sequence broke down and was turned on its head. The way in which AIDS activists took charge of both the evidence-producing process and the treatment approval process has provided a new and powerful model for how scientists, professionals, and client/patient groups might interact to produce solutions in domains where informed and concerned lay people are willing to invest time and effort in ensuring that their views are heard. Maguire's analysis of this remarkable case places in sharp relief phenomena that underlie to some extent the observations of our other symposium participants: "the nature and role of evidence are not given, stable and exogenous to the innovation and adoption processes of analytical concern. They emerge, rather, out of ongoing processes of discursive struggle in both the scientific and political arenas" (Maguire). The other articles suggest that the "discursive struggle" extends beyond these arenas to everyday professional practice contexts.

Andrew Van de Ven and Margaret Schomaker round out the FORUM with a commentary that both integrates and reaches beyond the ideas in the individual articles. They draw on Aristotelian rhetoric to reflect on the strengths and weaknesses in the rhetorical elements of the evidence-based decision-making model, concluding that as currently formulated, the model's persuasiveness tends to be limited by its insistence on "logos" (the logical pole of the rhetorical triangle) at the expense of "pathos" (appeal to values and experience) and "ethos" (credibility). Turning the argument back on ourselves, we hope that the evidence presented in these articles may strike a chord among readers, and in some way broaden our thinking about routes to improving health care practice.

REFERENCES