I think I am like many practicing physicians in my “love-hate” relationship with clinical practice guidelines. On the one hand, it is often helpful to look up a set of evidence-based recommendations on a particular clinical issue, and I feel particularly fortunate that the American College of Cardiology and the American Heart Association have collaborated to produce high quality guidelines on a wide-range of subjects relevant to my practice. On the other hand, I am well aware of the shortcomings of practice guidelines, including the limitations of the underlying evidence base, the challenge of synthesizing the available evidence into guidelines, and the often limited applicability of recommendations to clinical practice.
Even these well-known problems with practice guidelines don’t capture the broader issue of “guideline overload.” There are now so many guidelines, that the “old” problem of “keeping up with the literature” has been matched by the contemporary problem of “keeping up with the guidelines” and some areas of clinical practice have many competing guidelines with inconsistent recommendations. Just for yucks, I searched for “hypertension” on the National Guidelines Clearinghouse site and got a list of 548 relevant guidelines, all of which had met the strict criteria for inclusion on the site. Thanks a lot.
A related aspect of guideline overload is the challenge of caring for patients with multiple conditions, each of which may be the subject of recognized guidelines. A recent paper in the Annals of Internal Medicine (Ann Intern Med. 2013;159(3):161-168. doi:10.7326/0003-4819-159-3-201308060-00005) presented an interesting response, at least with regard to guidelines for preventive care. Researchers from NYU School of Medicine built a mathematical model to help clinicians prioritize preventive interventions (e.g. quitting smoking vs. losing weight) for patients with multiple co-morbid conditions. They limited the “inputs” to the US Preventive Services Task Force recommendations, and were clear that their effort was intended only as a “proof of concept,” but it seemed to me that the approach has real merit.
They used two different imaginary patients, and were able to show that their profiles of the most effective preventive measures were different. They conclude: “models of personalized preventive care may help clinicians prioritize … recommendations at the patient level. Future work may help determine whether model-based personalization is feasible at the point of care and is associated with improved health outcomes.”
In other words, if this kind of information can be presented to clinicians as they care for patients (say, by imbedding it in an EMR), and if it can be shown to make a real positive difference in how patients do over time (the real test of utility of everything we do), then this is a real advance. I agree!
What do you think?