Levels of Evidence Supporting Guidelines

Okay, so evidence-based medicine, using treatment guidelines that supposedly are backed by superior evidence of outcomes, is the way we are going to improve quality and control costs.  And all those guidelines really can be trusted, right, no reason to think we would ever change them?  Of course that is crap, and we should all recognize its crap.  Gotta be more skeptical, people.  And a study in the Journal of the American Medical Association shows why.   (JAMA Article)   The authors looked specifically at guidelines related to cardiovascular care.  In a study from 2009 only 11% of cardiovascular treatment guidelines were supported by the highest level of evidence.  And even that makes me queasy because the guideline compilers decide what level of evidence they think supports the recommendations.  The authors looked to see if the current, as of February 2019, guidelines had any better evidentiary support.  26 American and 25 European guidelines were included.  Only 8.5% of recommendations were classified (again, by the people who compile the guidelines) as having the highest level of support.  And just to show you how complex and confusing guidelines are, each guideline had a median number of 121 recommendations!  For the 25 European guidelines, 14% of recommendations had the highest level of evidence.  There is no significant trend in the last few years indicating that more recent recommendations or guidelines have higher levels of evidence to support them.  What the authors didn’t look at, but would be even more instructive is the number of recommendations that got modified or even eliminated.  Several times in recent years, guidelines for widespread chronic diseases have had very significant changes, suggesting that prior guidelines were actually recommending care that was not in fact optimal.  A health-related example we are all familiar with is how frequently dietary recommendations change; in extreme cases saying something that we were told not to include in our diet is now actually fine and vice versa.

Gives you every reason to be skeptical about corresponding efforts to reward or penalize doctors for following or not following guidelines.  What gives policymakers any confidence that they know what the right care should be.  Scientists in general need to stop pretending that they know more than they do; they need to be a lot more skeptical about what they believe they know, and acknowledge that even in the hard sciences new evidence continually is developed that indicates that our prior understanding about how things work is wrong.  And when people’s health is at issue, there is even more reason to be very cautious, and to respect the role of individual clinician judgment about what may be best for a patient.