Implementing Guidelines One Patient at a Time

First, a few disclosures: Not only do I regularly use clinical practice guidelines in my practice as a general pediatrician, but I am also a guideline author. By synthesizing complex information, guidelines can be great tools, especially as new tests and treatments become available. Furthermore, guidelines can reduce inappropriate variations in care, potentially improving health outcomes, reducing costs, and minimizing disparities [1].

My colleagues complain about “guideline fatigue” and “cookbook medicine”—complaints that are decades old. Anecdotally, it seems that these concerns have been exacerbated by worsening time pressures and by the implementation of electronic medical records. Clinicians quite reasonably believe that they understand the course of illness in their patients, as well as their patients' preferences for treatment. In addition, they often have vast experience with the particular condition being treated.

It is not hard to identify problems with guidelines. One of my biggest frustrations is the unfortunately common use of ambiguous and vague terms [2]. Expressions like “[a specific drug] may be considered” are not helpful. What does “consider” mean? A similar problem is, for example, the recommendation that testing be performed “in childhood.” This is not helpful. When in childhood? How frequently?

Another set of problems is that guidelines are often not transparent in how they were developed. How was evidence used and graded? Are there conflicts of interest? Are the recommendations really generalizable for the population for which the guidelines were written? In addition, guidelines may not consider real-world considerations such as who will pay for the recommended actions and how they will fit into the flow of practice. This can be especially troublesome for expensive treatments that are provided repeatedly (eg, palivizumab, to prevent respiratory syncytial virus infection in high-risk infants).

Despite these challenges, guidelines often become the quality metrics by which clinicians are judged, with target goals set by practice plans or other outside entities. I believe that the recommendations developed by the National Academy of Medicine for the development of high-quality guidelines [3] should help to improve guideline development (assuming that the guideline developers can follow guidelines) and to serve as a standard for assessing clinical practice guidelines.

When I am working on guidelines, I look to a framework published more than 15 years ago to understand why physicians do not always follow clinical practice guidelines [4]. As that article notes, guideline developers would do well to consider issues of health care provider knowledge, the degree to which there will be agreement with the guideline, motivations for adoption, costs, and resources. I also believe that guidelines should be accompanied by toolkits to facilitate implementation and that guidelines should undergo real-world testing prior to dissemination. Although it is rarely possible to conduct a randomized trial of guideline implementation, there are implementation study designs that allow testing and refinement of a guideline before it is released.

Traditionally, guidelines have been released to care providers like Moses carrying down the Ten Commandments. Guideline development can be one-sided, with no feedback from the end users. I think the solution is that guidelines should be implemented in the context of a learning health system [5]. We should take the opportunity to learn from each patient, to refine how care is provided, and to make guidelines better. Such data would allow us to better tailor the guideline to specific patient needs instead of dismissing the guideline altogether. I realize that this goal is in large part aspirational. However, most providers have electronic medical records that could help us to evaluate patient outcomes, at least for certain conditions. One of the most exciting developments in clinical practice guidelines is the use of Standardized Clinical Assessment and Management Plans (SCAMPs), which are flexible clinical practice guidelines that evolve over time using real-world clinical experience [6].

I am hopeful that the need to participate in quality improvement activities for maintenance of certification or the emergence of other reporting requirements could prompt clinicians to work together on efforts to strengthen guidelines and make sure that they are clinically relevant and effective. Looking into the near future, I believe that guidelines will increasingly be explicit in the incorporation of patient values and unique personal characteristics (eg, polymorphisms related to drug metabolism). These are some of the exciting goals of personalized medicine.

I do believe that guidelines make me a better pediatrician. I also believe that there is great opportunity to make guidelines better and that we would all do well to be critical, but not dismissive, in appraising new guidelines when they are released. I believe that it is important that we critically evaluate the care that we provide and use that evaluation to help improve the overall quality of care that patients receive. Finally, I believe that deviations from guideline-recommended care should be justified, and outcomes should be assessed to either refine the guideline or reinforce the need to follow the guideline as developed.

• ©2015 by the North Carolina Institute of Medicine and The Duke Endowment. All rights reserved.