Class 1 lesson: 20 percent of guideline recommendations revised

It may be time to re-evaluate the means and frequency of cardiovascular guideline updates, based on an analysis published online May 27 in JAMA that determined one in five Class 1 recommendations were later downgraded or omitted.

Mark D Neuman, MD, MSc, of the Perelman School of Medicine in Philadelphia, and colleagues examined Class 1 (“procedure/treatment should be performed/administered”) recommendations in 11 American Heart Association/American College of Cardiology guidelines and their status between revisions. The guidelines were published between 1998 and 2007 and revised between 2006 and 2013. Of 619 indexed recommendations reviewed, researchers found that 20 percent were either downgraded or omitted in subsequent revisions to the guideline.

“Our findings stress the need for frequent re-evaluation of practices and policies based on guideline recommendations, particularly in cases where such recommendations rely primarily on expert opinion or limited clinical evidence,” they wrote. However, the current process takes a median of six years and re-evaluates all recommendations, not just the ones that have new evidence or findings behind them.

“Finding ways of building on the evidence produced as part of the surveillance process, rather than starting each guideline update from scratch, is a promising area for reducing the timeframe to produce updates,” opined Paul G. Shekelle, MD, PhD, of VA West Los Angeles Medical Center, in his editorial response.

The research team reviewed the recommendations on whether they were retained, downgraded, reversed or omitted, but also on the rigor of the study. In 90.5 percent of cases, recommendations that were retained were supported by multiple trials. Researchers noted that the odds of a recommendation based on one trial or opinion being reversed, omitted or downgraded were higher than those that were supported by multiple trials. One in eight recommendations based on a single trial were downgraded, reversed or omitted and about a third of the recommendations based on opinion or standard of care were downgraded, reversed or omitted.

As guidelines affect practice and policy, these findings need considering.

Neuman et al stated, “Such information may aid clinicians and policy makers in quantifying the potential risks of measuring physician performance based on adherence to recommendations derived from limited clinical evidence.”

However, there were concerns about recommendations that were omitted between revisions without explanation. “As each of these potential reasons for omission carries distinct implications for practice,” Neuman et al wrote, “our findings stress the importance of communication on the part of guideline-producing bodies regarding the reasons that specific recommendations are removed from guidelines, as well as any changes in practice that might be implied by their removal.”

Both Shekelle and the Neuman research team agreed that evidence, speed and adequate communication of changes were paramount in moving forward with Class 1 recommendations in future published guidelines, in particular to addressing why changes and omissions occurred.

Shekelle suggested, “As the preferred method of accessing information moves increasingly to the internet, guideline developers ought to be able to produce electronic versions of guidelines that have hypertext links to take users directly to the information they are most interested in, and one of these links should be to a table outlining ‘what’s new?’ and ‘what’s different?’”

Expediting updates when new findings change Class 1 recommendations, determining the strength of the argument to update recommendations and best practice for communicating the changes and why are all key to ensuring clinical practice stays up-to-date. “While our results highlight the overall durability of cardiovascular disease guideline recommendations, they also emphasize that particular subsets of recommendations may be more fragile than others as a basis for changes in practice and policy,” Neuman et al stated. To that end, considerations should be made about the confidence level of Class 1 recommendations as they are issued, to ensure the best possible care for patients.