Guideline fail?

What if adherence to a guideline was associated with worse outcomes? How much independent validation of guideline recommendations should occur? Is it realistic to expect guidelines to be subjected to “real-world” testing prior to adoption?

If these questions intrigue you, you’ll be interested in a Pfizer-pfunded four-center performance improvement initiative that included education around the ATS-IDSA pneumonia guidelines and a prospective assessment of outcomes. As the authors report in Lancet Infectious Diseases this week, patients who received “guideline compliant” empiric therapy were more likely to be dead at 28 days than were those who received “guideline non-compliant” therapy.

Before you submit your resignations to IDSA and ATS, you should know that there were plenty of problems with this observational study—most of them are well-summarized in an accompanying editorial. The failure of the authors to consider appropriate de-escalation of therapy in their determination of guideline compliance is an especially big problem, given that antibiotic overuse has been linked to increased mortality in the ICU.

Despite the limitations, I think this study raises important questions about guideline-driven approaches to complex infections. It is interesting, for example, that guideline non-compliant empiric therapy was equally or even more likely to cover the eventually-isolated pathogen than was compliant therapy (85% vs. 81% of the time, respectively).