After 10 to 15 years of trying to improve the quality of medical care by developing and disseminating literally tons of clinical guidelines, it is time to admit that we have been doing it wrong. The best evidence for my claim is summarized in the latest report on quality from the Institute of Medicine.¹ Eleven years after publication of the first of its three reports on guidelines,² this report describes not just a gap between what we know and what we do, but a chasm. The report concludes that, if we are to bridge that chasm, a revolution is needed in the way we organize and provide medical care.

The problem is not that we need different physicians or need to get care to people who do not come to see us, it is that we physicians are not consistent and comprehensive in our care for those patients we do see. We just do not provide preventive and chronic care very well, managing everything as acute problems (although there are gaps here, too). Part of the problem is that we are too busy, we are not capable of remembering everything, and we concentrate on responding to the issues that our patients bring to us.³ More fundamentally, however, it is because we have not set up organized support systems in our offices. We appear to believe that we have to do everything ourselves, and we do not understand how to implement guidelines.

A recent article in Family Practice Management highlights the problem.⁴ It describes a group practice in Louisiana that decided to compare their care with the National Institutes of Health (NIH) “General Practice Guidelines for the Diagnosis and Management of Asthma.” The practice's chart reviews revealed that their treatment was focused on managing acute exacerbations of asthma, with little prevention, maintenance, or patient education being done. With that knowledge, they designed and conducted two educational sessions on the key recommendations in the NIH guideline and “challenged the physicians and staff to adopt them into regular practice.” A repeat audit showed no positive changes. Then they implemented a simple reminder system and found improvements in all four remindertargeted behaviors. The authors concluded that the extensive randomized trial literature was correct about the failure of continuing education to affect behavior, as well as about the efficacy of reminder systems.

Unfortunately, their report illustrates the problem of quality improvement and guideline implementation in a much larger way than they described. First, it demonstrates that we have not only failed to implement evidence-based guideline care, we have largely failed to even become aware of the extensive evidence base on various approaches to changing clinical behavior. Like the first efforts of these physicians, I still see most well-intentioned efforts to improve practice being based on education, exhortation, feedback of data, and the use of opinion leaders and so-called “academic detailing.” There are so many good randomized trials demonstrating the ineffectiveness or marginal benefits from these approaches that there have actually been 47 high-quality systematic reviews of a much larger number of scientific trials evaluating implementation strategies.⁵ There is, indeed, a place for the strategies mentioned above, but only as a foundation for the truly effective approaches of task delegation and office systems.

Secondly, even if there were more widespread understanding of the literature on this topic, the literature itself fails us.⁵ For example, while there is indeed excellent evidence supporting the use of reminders, scientific trials of reminders are usually relatively short-term and do not test the likelihood that, over time, reminders lose much of their effectiveness. Much worse, though, is the failure of the literature to emphasize the systems nature of what does work, or the organizational leadership and change management that is needed to put these systems into place. Only recently are these lessons from other lines of work starting to appear in the medical literature.^6,7

Because the medical groups in our region have been forced to be organized in order to survive, and because many groups have worked collaboratively through the Institute for Clinical Systems Improvement (ICSI) to develop and implement guidelines, they have had an unusually good opportunity to learn these lessons.^8,9 When ICSI began, there was considerable belief that if local physicians became actively engaged in the development of a guideline, they would implement it. After developing 50 primary care guidelines, practices of all sizes learned this belief was untrue. Implementation required work to set up systems, and focusing primarily on changing physician behavior was futile.

Fortunately, we had the resources to conduct small studies of guideline implementation efforts. One of the best studied a systematic effort by one medical group to implement a guideline for management of simple cystitis episodes in women.¹⁰ The implementation approach used was to combine physician and nurse educational efforts with referral of most phone calls to nurses. The nurses used guideline-driven protocols to collect data and to provide three-day antibiotic treatment without cultures to patients who fit the protocol. After implementation of the protocol, these clinics demonstrated a dramatic improvement in adherence to the guideline. However, when the cases managed by nurses were compared with those of similar complexity managed by physicians in these same clinics, it was clear that all of the improvement came from the nurse protocol cases. Physician behavior had not changed at all.

Recently, we conducted in-depth interviews of the people from these medical groups (12 physicians and staff) who had the most successful experience with leading guideline implementation, so-called “insightful implementers.”¹¹ Their recommendations emphasized organizational changes and systems and the use of multiple strategies, not single ones. Only two of the top 22 aspects of change management that they recommended related even partly to individual clinicians or to characteristics of the guideline itself.

All of this might suggest that only large clinics could implement guideline-driven systems. Nothing could be further from the truth. The one clinic that consistently did the best job—the real star of the ICSI collaborative—was a four-physician practice in a small town. These physicians understood the value of systems and strongly supported the efforts of their laboratory technician, who was very good at setting up the processes that supported desired guideline actions. My experience in a two-physician practice was similar, and the smoking cessation system that we used helped more than 200 of our regular patients to quit smoking in a four-year period.¹² Our system became the model for the American Academy of Family Physicians' Stop Smoking Kit.

In fact, a solo physician who understands systems principles and who is willing to delegate supporting roles in care to other office staff can actually implement guidelines better than anyone else. Such a physician can establish registries of patients with common chronic conditions, have their charts labeled, and have nursing staff identify and remind or address guideline-recommended care needs for these patients during any of their visits. Staff can also mail or call needed reminders to patients who do not come in to the office.

If your practice group is interested in implementing evidence-based guideline care and in showing the world that there does not need to be a “quality chasm,” you can do it. All you have to do is discard outmoded ideas about how to change your practice and build your own office systems. You too can be an “insightful implementer.”