Original Article: Why Are We So Slow to Adopt Some Evidence-Based Practices?

Issue Date: December 15, 2018

See additional reader comments at: https://www.aafp.org/afp/2018/1215/p709.html

To the Editor: I read this editorial with great interest. Although I believe that many of the factors that the authors presented have contributed to the lack of adoption of evidence-based processes, there is one more that I would like to add. For those who have been in medicine long enough to see evidence-based processes be proved wrong with follow-up studies, it is hard to jump on board a change that one's experience also has shown to be false. The good news is that medicine continues to grow and change, and physicians scientifically evaluate and incorporate new data.

In my family medicine consulting practice, physicians refer patients to me for colposcopies and loop electrosurgical excision procedures. I had never seen so many cervical intraepithelial neoplasia 3 biopsy results until guidelines began recommending that Papanicolaou (Pap) tests be performed every three to five years. As physicians, we sometimes forget that many research studies are conducted under optimal conditions, but in practice there seem to be more false-negative Pap test results.

In the past, physicians have prescribed fish oil supplements for cardiovascular disease; however, studies subsequently found that these supplements are ineffective for primary or secondary prevention of cardiovascular events.¹ Many other examples could be listed.

Evidence-based practices need to incorporate contradictions in studies, change as medicine progresses, address concerns about changing for change's sake (despite being well-meaning), evaluate real-life vs. study processes and results, protect our patients because less is not always more, and realize that the latest is not always the best.

Thank you for the editorial in American Family Physician. It obviously got the attention of this reader.

In Reply: Thank you for your thoughtful comments. We agree that there are factors beyond those we discuss in the editorial that contribute to practice change. We feel that there are two kinds of evidence: disease oriented and patient oriented. Many studies primarily report disease-oriented outcomes, such as improvements in a biomarker, a physiologic measure, or other surrogate outcome. These often mislead us: examples include hormone therapy,¹ vitamin E,² omega-3 oils,³ and tight control of type 2 diabetes mellitus.⁴ More reliable guidance comes from well-designed, unbiased studies that report patient-oriented outcomes such as improvements in morbidity, mortality, and quality of life. These studies are also less likely to be reversed than evidence from disease-oriented studies.

Regarding the impact of recommendations to change the interval for cervical cancer screening with cytology alone from one to three years, we are not sure that your experience of an increased prevalence of cervical intraepithelial neoplasia (CIN) 3 is typical. A 2017 study reported the incidence of CIN1, CIN2, and CIN3 in females in New Mexico from 2007 to 2014 and found significant decreases in the incidence of each finding.⁵ A possible explanation for the perception of more CIN3 is that the denominator has changed, with fewer normal Pap tests being performed in the setting of screening every three years. That said, it is important that physicians be vigilant and that we adhere to longer screening intervals, and not let three years (or five years if screening for high-risk human papillomavirus with or without cytology in women older than 30 years⁶) become five or seven years. Our goal should always be the right amount of care, for the right patients, at the right level of intensity. Both too much and too little care can be harmful.

Editor's Note: Dr. Ebell is Deputy Editor for Evidence-Based Medicine in AFP, and Dr. Shaughnessy is an Assistant Medical Editor for AFP.