The title of this rant is almost disingenuous, as my main premise is that we have let guideline production “drift” and therefore my goal is return guideline development to its original intention. Our approach to guidelines reminds me of my favorite Led Zeppelin song – “Dazed and Confused.” I hope that a thoughtful return to fundamental principles will stimulate debate, and perhaps even revive the guidelines movement.

• 1. We must choose clinical issues which deserve guidelines carefully. Every problem does not need a guideline. Guideline development should marry clinical need and informed research.
• 1a. As a corollary, we should have two levels of guidelines – guidelines for generalists and guidelines for subspecialists. As a generalist I need a concise evaluation of the key common issues related to a problem For example, when considering atrial fibrillation, I want to know several things: clinical evaluation for etiology, when to cardiovert, how to achieve rate control, when to anticoagulate, when to refer. A cardiologist may want guidelines which involve advanced electrophysiology or indications for antiarrythmics.
• 2. Guideline panels should reflect intended users. Thus, if we want guideline to inform generalist practice the panel should have a majority of generalists. The panel could include subspecialists, but the generalists should insure that the guidelines remain practical and usable.
• 3. Guidelines should consider patients with other medical conditions. Thus, a diabetes guideline should include an understanding of the impact of CAD or CHF. Guidelines which consider one disease may or may not pertain when patients have multiple diseases.
• 4. Guidelines should prioritize the recommendations. For example, if I were developing a CHF guideline, I would prioritize in this approximate order: #1 establish the diagnosis (Echo), #2 treat systolic dysfunction with ACE-I or ARB (and a discussion of dosing), #3 add beta blockage (usually carvedilol), #4 diurese to keep patient euvolemic – avoid volume contraction, #5 consider adding sprionolactone in selected patients, #6 indications for AICD, #7 indications for biventricular pacing, #8 when to consider referral for possible transplantation.
• 5. Produce readable guideline documents. Generalists do not have time to read long tomes.
• 6. Have guidelines sponsored through an independent agency. This could be a government agency, or an independent group. We should divorce funding from sources which have a stake in the guideline (i.e., no company which provides drugs for treating hyperglycemia should fund diabetes guidelines).

These represent my first thoughts. I suspect you readers will make some good suggestions for modification or expansion. Please share these thoughts, and I will gladly modify my reinvention plan.

Thanks for the many thoughtful comments as I have explored these issues over the past 2 weeks!