With regard to clinical practice guidelines, clinicians want an authoritative resource that will clearly and concisely instruct them in most clinical scenarios. Guideline developers want to give them this, “but producing guidelines is not as straightforward as it might seem,” according to David Garcia, MD, Professor of Medicine at the University of Washington in Seattle, who has been involved in developing guidelines in hematology.
At the 2013 ASH Annual Meeting, Dr. Garcia cited several challenges that must be addressed during the development of guidelines: the participation of parties with conflicts of interest, the difficulty of calculating net clinical benefit, and the danger of including recommendations not based on high-quality evidence.
Conflicts of Interest Cannot Be Avoided
Potential conflicts of interest may be not only financial but intellectual, and not only real but perceived. Financial conflicts of interest are well recognized and can affect both the direction and strength of the guidelines. Even the perception of a financial conflict of interest can challenge the credibility of the guidelines, he pointed out, but intellectual conflicts of interest may be even more important, he contended.
“Writers of guidelines may have powerful intellectual conflicts of interest. If you consider that an expert in a field spends his or her career writing grants, and conducting trials to prove that treatment A is better than treatment B in a certain situation, it may be hard for this person to consider the totality of the evidence and come up with a recommendation going against his or her own hypothesis,” he suggested.
But it’s impossible to completely eliminate “conflicted authors,” since these tend to be the experts in the field, he pointed out. “If we exclude them, it’s possible the guidelines will miss the mark and fail to address key real-world dilemmas,” he said.
“World authorities” who find themselves excluded may, in turn, “find occasion to criticize the guidelines and call their validity into question,” he suggested. “This would likely lead to poor uptake of the guidelines in clinical practice.”
Dr. Garcia suggested these solutions: Include authors with potential conflicts of interest but disclose all conflicts of interest (financial and intellectual), restrict financial conflicts of interest (ie, set a dollar-per-year threshold), limit voting or discussion by authors with selected conflicts of interest, and include nonconflicted persons as independent safeguards (ie, methodologists trained in interpreting evidence).
How Strong Are the Recommendations?
The strongest recommendations are based on high-quality evidence that (1) the intervention impacts an important clinical outcome, rather than a surrogate outcome that may not matter to the patient, and (2) that the benefits likely outweigh any risks.
While these may seem obvious, physicians and patients often differ on these points. Future guideline writers need to focus on patient-important (not surrogate) outcomes whenever possible, he suggested, and be equipped with better knowledge about how patients view the trade-offs associated with treatment options.
Even when there is high-quality evidence, individual patient preferences as well as resource considerations (eg, widely varying treatment settings) may reduce the strength of a recommendation, he added.
“So, for those asking why the strength of so many guidelines is weak, there are many reasons,” Dr. Garcia said.
Perils of ‘Expert Consensus’
“When I am in a clinical situation where the evidence is inconclusive, I pick up the phone and call someone with practical experience in this situation,” he said. “Experiential recommendations are important to me in the hallway, but publishing expert opinions can lead to unintended consequences.”
While they sometimes cannot be avoided, guidelines based on expert consensus can be “unjustified, unreasonable, and stifling of future clinical research,” Dr. Garcia maintained.
“Once a group of experts makes a recommendation, even if it is labeled weak, the tendency of practicing doctors who just want the answer to what they should do is to adopt the practice—sometimes with such fervor that convincing them to enroll a patient in a clinical trial to definitively answer the question can be challenging, and even unethical,” Dr. Garcia suggested. “Sometimes, we recommend things that turn out to be harmful, such as hormone replacement therapy in postmenopausal women.”
Some of the weaknesses in guidelines can be ameliorated, he said, through transparency and by the way they are used. “We need to maintain transparency regarding the evidence and methods used to create the guidelines, and any possible conflicts of interest,” he said. “And we also need to provide key recommendations for use at the bedside, as a summary, checklist, or algorithm, with the strengths clearly labeled.” ■
Disclosure: Dr. Garcia reported no potential conflicts of interest.
