Time to weed the CPG garden

In One Hundred Years of Solitude1 Gabriel GarcÍa Márquez describes a plague of insomnia and gradual memory loss among the inhabitants of a town known as Macondo. To get around the problem that, eventually, no one will be able to remember anything, the residents mark everything with its name — table, chair, clock, cow — along with instructions for use. But the vigilance required to maintain the system is beyond the moral strength of the inhabitants, so they succumb to “a spell of imaginary reality” in which facts are conjured vaguely by the reading of cards. One of the town leaders builds a memory machine, but even that fails.

Clinical practice guidelines (CPGs) are a memory machine of facts and instructions. But, as with Márquez's memory machine, they are more gratifying in their promise than in their outcome. Guidelines, alas, have failed to encourage proper use of diagnostic tests and therapeutic manoeuvres.2

This is so for lots of reasons. To begin with, there are too many CPGs, and many are of dubious quality. The CMA, an enthusiastic champion of guidelines, maintains the most comprehensive database, providing an online gateway to 1755 CPGs (or sections thereof) at www.cma.ca/cpgs/index.asp through the CMA Infobase. One would think that such a resource would provide busy practitioners with a painless route to the clinical information they need. Yet, as Ian Graham and colleagues report (see page 157),3 there are serious problems of quality. Of 217 drug therapy guidelines developed or endorsed by Canadian organizations in the period 1994–98, only 15% met half or more of the authors' criteria for rigour in the developmental process, and only 7% were rated by independent reviewers as sound without modification. Also in this issue, Walter Rosser and colleagues describe their efforts in Ontario to appraise CPGs and to make high-quality guidelines more user-friendly (see page 181).4

As Steven Lewis suggests (see page 180),5 perhaps we don't need more guidelines: we need fewer, and better, ones. The current method of CPG development depends on the missionary enthusiasm of an increasing number of subspecialists, the we-want-our-own-guideline provincial health ministries, and the not-to-be-left-out research and advocacy foundations — not to mention the commercial zeal of drug manufacturers. The result is too many CPGs on the same topics that vary in quality, make conflicting recommendations and fail to inspire trust.

The CMA Infobase ought to be a garden, not a warehouse. It needs regular and judicious weeding and clear labelling with regard to the grade of the produce. Perhaps more than anything, it needs the introduction of beneficial species: for each subject, a single national guideline, developed by impartial experts, that meets current quality criteria, has practical means of implementation and is supported by incentives for correct use. This will make for hardy, wilt-resistant plants, but they will still need regular maintenance. As demonstrated recently by Pérez-Cuevas and colleagues6 and emphasized by Davis,7 we need dynamic, “living” guidelines that grow along with science and experience. — CMAJ