Author’s Reply
To the Editor: We read with interest the comments of Drs. Lember and Rätsep on their study of diabetes mellitus clinical practice guidelines in Estonia. Although their results contradicted our findings in the United States, we suggest that the differences are largely accounted for by methodological differences in the study design.
Diabetes mellitus (DM) is a common disorder that affects all ages and is associated with significant morbidity if not properly treated. In addition, it is highly likely that most family doctors have daily or weekly exposure to patients with DM. Access and motivation to read DM guidelines may have been affected by these facts. In our study, the focus was upon 4 age-specific guidelines, with the realization that family doctors would vary in the number of patients they treated with each clinical problem. Motivation to read the guideline might be affected by the number of patients that are seen with that clinical condition as well as the potential impact of mistreating the condition.
Similarly, our study asked the doctor to respond to a very general question, “Have you changed your practice as a result of reading the guideline?” Although more respondents in our study stated that they had changed practice, the social desirability and general tone of the question makes it easy to respond affirmatively.
It is our opinion that the differences in study results highlight the limitations of both study designs—reliance upon self-report as opposed to observation of actual practice. As is the case in many areas of science, financial restrictions adversely affect research designs.
The low rates of overall compliance in both of these articles are not surprising. A number of studies have shown poor correlation between either continuing medical education (CME) or clinical practice guideline (CPG) dissemination and actual changes in patterns of care in clinical settings.1, 2 Our study found that approximately one third of the family physicians we studied were very aware of the guidelines, one third were somewhat aware of the guidelines, and one third were unaware of them.3 We would extend this to a general rule, probably applying to most specialties—the “law of thirds.” In medicine, as in most areas of life, there is a range of knowledge among the practitioners. The family physicians who are in the upper third probably have the most interest and/or ability in reading, learning, and keeping current, and probably the most flexibility in adapting their practice patterns to new changes. They will learn aggressively whether or not there are CME requirements or new CPGs showing up in their mailbox. The middle third will maintain some level of currency, but are probably more focused on clinical practice. Their currency will remain moderate regardless of CME or CPG requirements. Finally, the lowest third will do their best to keep up while maintaining a busy practice. They will meet their CME requirements, but this will not produce significant changes in their practice.
What is a guideline? It is essentially a well-researched and authoritative review article that has the blessing of the body promoting it. As such, it is meant not only to be a good review article, but also to set a standard of care. Hopefully, it is has more practical point-of-care usefulness in its organization and presentation than a good review article, but this is not always the case. As with any review article, how well physicians use it will ultimately reflect their pre-existing scholastic qualities—i.e. which “third” they reside in. Indeed, in our study, the rates of self-reported change in practice correlated closely with which “third” reported the change; we suspect that had Lember and Rätsep segregated their group by level of familiarity with the diabetes guideline, as opposed to availability, a similar trend would have been found.
There has been a great deal of research about how to improve physicians’ compliance with CPGs, but we believe that gentle reminders at the point of care, possibly linked to financial performance incentives, are the most likely to work.4 It may be that a gradual shift to electronic medical records will make this more practical in the future.