viernes, 1 de abril de 2011

Guiding the guidelines

The Lancet, Volume 377, Issue 9772, Page 1125, 2 April 2011

doi:10.1016/S0140-6736(11)60445-5

Guiding the guidelines

On March 23, the US Institute of Medicine (IOM) released two complementary sets of standards to improve the quality of systematic reviews and clinical practice guidelines. As the IOM points out, clinicians can be overwhelmed by the quantity of reported clinical evidence, and often need to rely on aggregated evidence to guide decision making (ie, systematic reviews and guidelines). But how well such reviews and guidelines can or should be done was the question posed to the IOM. Their answer includes 21 standards for systematic reviews and eight for guidelines.

For standards for systematic reviews, the IOM acknowledges the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), which it drew on to formulate its recommendations. Some of the standards are obvious: include experts in the clinical field under study and in systematic review methods. Some are worth highlighting: include expertise in searching for the evidence—an information scientist—a point sometimes omitted (or not reported) in published reviews. Some are strict, but welcome. For managing conflicts of interest, the IOM recommends not only full disclosure but also exclusion of those with clear financial or intellectual conflicts. The standards include the need to consider the non-English literature, the so-called grey literature and abstract booklets, and unpublished data. Some medical publishers will be shocked to read the final standard: publish the report with free public access. The Lancet welcomes that recommendation.

Clinical Practice Guidelines We Can Trust is the tandem document from the IOM—tandem in the sense that one of these standards states that a systematic review must precede guideline formulation. The IOM points out that the database of the Guidelines International Network includes more than 3700 guidelines from 39 countries. The standards were formulated to overcome common deficiencies—particularly in bias and how bias is handled—in existing guidelines. These deficiencies include the variable quality of evidence, poorly done systematic reviews, methodology that is not transparent, a development group that does not reflect all stakeholders, and unmanaged conflicts of interest. The IOM recommends that before the development group is formed, conflicts of interest need to be fully disclosed—and their definition includes financial, intellectual, and institutional interests. The prospective panel members should also explain how any conflicts they might have could influence the guideline process. The IOM is clear in some specific areas: members of the group need to divest relevant financial investments they or family members hold, and not take part in advisory boards or marketing activities with relevant companies. Additionally, the chair and co-chairs should have no conflicts of interest, members with conflicts should form a minority, and funders should have no role in guideline development. Overall, tough talk from the IOM—but much needed talk. Apart from expert members in the group, the standards include involvement of patients or their representatives.

For recommendations in the final clinical practice guideline, the standard proposes explanation of the underlying reason for the conclusion, including benefits and harms, a summary of the evidence and its quality (and any gaps in the evidence), and a rating of the level of confidence in the recommendation. The guideline should also include explanation of any part played by opinion or clinical experience, and description of any differences of opinion. The concluding standard concerns the need to update the guideline when new evidence of benefits or harms emerges. The final translation, as the IOM puts it, is to ensure a clinical practice guideline is actually used by practitioners through, for instance, clinical reminders, decision aids, and quick reference guides.

The perennial problem with systematic reviews and practice guidelines is that they remain population-centred not patient-centred, which makes individual treatment decisions not as easy as might be hoped. The IOM acknowledges this problem: “particularly under conditions of uncertainty regarding optimal decisions, clinician experiential knowledge and skill (the ‘art of medicine’) and patient values and preferences remain essential contributors to quality health-care practice, in a complex interplay with science”. It is within such interplay that the IOM hopes that its new standards will aid clinical decision making rather than dictating a one-size-fits-all approach.

To borrow from Donald Rumsfeld—in medical treatment, there are known harms and benefits; there are harms and benefits known to need further investigation; and there are harms and benefits yet to be discovered. The two new sets of standards from the IOM will hopefully aid in the finding and synthesising of those knowns and unknowns.