July 14th, 2014

Poor Retention of Guideline Recommendations

In an article published in the May 28th edition of JAMA, a group of physicians addressed the question of the retention of clinical practice guideline recommendations over time. The investigators assessed the durability of Class I recommendations to provide a perspective on how often actions that are deemed mandatory become less enthusiastically endorsed as more evidence emerges.

Of 619 Class I recommendations in 11 index guidelines published between 1998 and 2007, 20% of the recommendations did not appear in the subsequent guideline and 9.2% were downgraded or reversed. Even among the recommendations that were supported by multiple trials, 5.7% were not retained in subsequent versions of guidelines and 3.8% were downgraded or reversed. The Class I recommendations at the highest risk for being omitted, downgraded, or reversed were those based on consensus opinion, followed by those based on a single randomized trial or nonrandomized trials.

I would have thought that the strongest recommendations in a guideline would be very unlikely to change over time. In this disturbing study, a substantial proportion of Class I recommendations were not durable over even a 10-year period, showing us that uncertainty surrounds even the strongest guideline recommendations. I wonder whether the writing committees would have been able to predict which of the Class I recommendations were most likely to be omitted, downgraded, or reversed based on future studies. I’d bet they could not have guessed correctly.


Why do you think such a high percentage of guideline recommendations are omitted, downgraded or reversed? Does this information affect your opinion of the importance of existing guidelines?

2 Responses to “Poor Retention of Guideline Recommendations”

  1. H Robert Silverstein, MD says:

    “Sic transit gloria mundi”: thus passes the glory of the world–should not be forgotten. More HUMILITY in the strongest of terms should be required of the guideline writers: they are de facto judge, jury, and executioner for anyone who would deviate from those gospellines. There is much too much pomp in the declaration of the makeup and delivery of these guidleines. Not enough position is give to those who disagree & not enough safety is suggested to support those who disagree or do not completely agree. HRS, MD, FACC

  2. Matthew Carr, MD says:

    I am not sure we should ever give class one status to any guideline that is based solely on expert opinion!
    Evidence based medicine requires better evidence. I am an expert in a few things but am able to resist the temptation to turn expert opinion into dogma unless backed up by randomized studies.
    We had a saying in medical school. “This years test questions will be the same as last year, but the answers will be different”. Want a good laugh? Look at the history of sulphonylureas for diabetes and all the studies that have been done.