The Enormity of the Task: SUPPORT and Changing Practice

Abstract

To interpret the SUPPORT study, it is important to place it in the context of design of interventions aimed at changing behavior. This specialized field draws on theories and empirical research from several spheres, including studies in social psychology, adult learning, professional education, knowledge and research utilization, communications research, social marketing, and the adoption and diffusion of innovations. Models of behavior change have been developed that provide a theoretical foundation for shaping health promotion and disease prevention strategies. For the most part, these theories have been applied to the design of health promotion campaigns that aim to encourage individuals to adopt healthy behaviors or refrain from ones that put them at increased risk of illness or injury. Recently there has been growing interest in learning how to encourage physicians to adopt new behaviors. This interest arises in part from a desire to ensure that medical practice reflects state-of-the-art scientific knowledge, in part from a desire to curtail unnecessary, often expensive procedures and tests, and in part from the recognition that traditional continuing medical education has had very little impact on physicians' clinical practice.[1] Whether directed to physicians or others, there is a standard method for designing interventions that includes the selection of an appropriate target audience, program goals, and messages as well as methods for delivery, implementation, and evaluation. Well-designed interventions are informed by empirical research on the current beliefs, attitudes, and practices of the target audience and on the barriers and incentives that are likely to impede or encourage adoption of the new behavior. Ideally, there is an articulated theory of behavior change driving decisions about the content and structure of the intervention and an explicit understanding on the part of the design team about the values upon which the intervention is premised. In short, sound behavior change interventions are theory driven, explicit about value assumptions, and empirically based. Although the SUPPORT study does not address the theory of change that guided the design of its intervention, it does indicate substantial attention to the values that shaped the program. These included a commitment to basing decisions about the use of life-sustaining medical treatments on patients' personal values and goals as well as a desire to improve communication among nurses, physicians, patients, and families. Furthermore, the intervention appears to have been based on considerable empirical research that allowed the researchers to develop well-grounded hypotheses about the problems with current decisionmaking patterns and likely ways to improve the process. The strategies designed to topple these barriers seem reasonable and not unlike other interventions that have produced a positive effect. For example, A. J. Dietrich and colleagues found that an enhanced role for nurses and simple shifts in office protocol, such as the use of flowsheets that cued behavior, were successful in improving primary care physicians' cancer screening and detection practices.[2] Yet in the SUPPORT study similar strategies had no effect on five key measures related to terminal care decisionmaking. What are some plausible ways to interpret these results, and what should those of us who are committed to improving terminal care in the United States do now? My answer is threefold. First, we should learn more about the content of the SUPPORT intervention and how it was implemented. It is quite possible that minor modifications in program content or in the way in which it was delivered and implemented could have made a major difference in program outcome. Until we learn more about the content of the SUPPORT intervention and the process by which it was implemented, we should be wary of drawing overly broad or negative conclusions about the power of an intervention, either the SUPPORT intervention itself or other interventions, to effect improvements in terminal care. …