Cardiologist Milton Packer, MD recently published an editorial criticizing the medical community for being too reluctant to accept findings from well-designed treatment trials as a basis for clinical decision-making -- seizing on minor issues such as underrepresented populations or lack of long-term safety data as reasons to maintain current practices and to call for yet more trials.
We contacted experts from several specialties via email to ask:
Do you agree that, when a trial gives robust evidence that a treatment is effective and reasonably safe, that physicians drag their feet in adopting that treatment?
When seeing results from a seemingly definitive clinical trial, how does one find the appropriate middle ground between being too quick to change practice and being too slow? How frequently does physicians' famous cautiousness work to patients' disadvantage?
How about Packer's charge that the medical community is too eager to see slivers of hope in trials with overwhelmingly negative results?
The participants this week are:
Ajay J. Kirtane, MD, SM, chief academic officer, Center for Interventional Vascular Therapy, and director, Cardiac Catheterization Laboratories at Columbia University Medical Center/New York-Presbyterian Hospital in New York City
Marilyn M. Rymer, MD FAHA, vice president, neuroscience and professor, neurology at the University of Kansas Hospital in Kansas City, Kan.
D. Ross Camidge, MD, PhD, professor, medicine/oncology and Joyce Zeff Chair in Lung Cancer Research at the University of Colorado at Denver Anschutz Medical Campus in Aurora
Pamela Douglas, MD, professor, medicine and Ursula Geller Professor for Research in Cardiovascular Disease at Duke University School of Medicine in Durham, N.C.
Anne B. Curtis, MD, FACC, Charles and Mary Bauer professor and chair and UB Distinguished Professor, Department of Medicine at the University at Buffalo in Buffalo, N.Y.
Examining Magnitude of Benefit
Kirtane: In my mind, the reason for the relatively slow adoption of clinical trial results relates to the absolute magnitude of benefit observed in trials of cardiovascular medicine these days. Unfortunately the numbers needed to treat based upon many trials' results are so great that this that can explain some of it. For therapies with clear and large benefits (e.g., TAVR for aortic stenosis), adoption can be very rapid. I also think that physicians are trained to be skeptical, and there are definitely those who just are more comfortable with what they know. But that's not always a bad thing, because it's important to separate the "hype" from actual results that have the promise to dramatically alter patients' lives.
Rymer: When physicians are seen to be "dragging their feet" in embracing new therapies after a positive clinical trial, there are often underlying reasons for slow adoption. The controlled use of a drug or device in a clinical trial is very different from widespread prescription to a diverse population of patients. It might be quite reasonable for a physician to wait and see how the drug or device performs outside the trial before recommending it to patients. Another reason for slow adoption may be that use of the drug or device depends on complex systems of care and/or specialized personnel that take time to develop. That has certainly been the case for the use of IVtPA in acute stroke. Then there is the substantial cost of new therapies that might cause physicians to consider relying on older, less expensive treatment even though the trial results indicate improved efficacy for the new drug.
Camidge: The only exceptions -- where new drugs don't penetrate the market -- reflect a growing dissatisfaction with new drugs that deliver little benefit, no matter how statistically significant, or do so at exorbitant costs, or both. Take necitumumab for example: a positive phase III trial improving overall survival in squamous lung cancer receives an FDA license, but there is a strong recommendation from the NCCN guidelines committee that the benefit is too low, toxicity too high, and cost too out of keeping with benefit to utilize it. Ironically, new drugs with the same low level of benefit seem to be what are being discussed in cardiology as disappointingly slow in adoption, reminding us that personalized medicine may be the direction cardiology needs to move in to change the world, more so than funding trials that have to contain 8,000 patients in order to show "benefit."
Striving for Middle Ground
Douglas: We should strive for a middle ground between infatuation with each new medication before it is fully proven, and being so consensus driven that we unduly delay adoption. New models of research and regulatory policy that cautiously introduce novel therapies while continuing to collect effectiveness and safety data would convert our healthcare system into a "learning environment" while providing patients with quicker access to promising medications. We are already doing that to some extent with CV devices with the NCDR registry requirements for ICDs and TAVRs; other strategies that bridge the gap between commercial development and widespread adoption are being discussed and tried -- and should be encouraged.
Camidge: Perhaps oncologists are just more optimistic, or have fewer choices in the cupboard already so something new stands out more, or at worst perhaps the preponderance of intravenous drugs for cancer generates financial incentives that encourage the uptake of newer and more expensive drugs by practitioners who share in their infusion center's profit, but either way it seems that oncology contains a lot more early adopters than cardiology does. The almost exponential rise in nivolumab sales provides ample evidence of this, even for a drug that usually has only about a 20% response rate in most cancers.
Curtis: It can be difficult to find the middle ground. There are clearly examples of a single clinical trial showing a result that could change practice, only to have it not be replicated in other studies or other problems found when the treatment approach becomes more widespread. Such an example would be drugs that pass through clinical trials, are approved by the FDA, but then are withdrawn from the market later when serious side effects are documented that did not become apparent during the clinical trials. One has to weigh the strength of the benefit expected and the other alternatives available to treat the patient in determining when to adopt a new approach.
Importance of Caution
Rymer: Although there may be some physicians who are overly cautious in prescribing new therapies, in general the decision to use a new treatment based on trial results depends on the risk and benefit to an individual patient. If mortality, morbidity, or suffering can be substantially improved or alleviated by early adoption of a new therapy, then it is worth the risk of unexpected, and sometimes serious, side effects that could result and the increased cost of care.
Curtis: I believe one has to distinguish between physicians practicing in the community and what is written in guidelines. It is true that guidelines sometimes change too slowly. The delay stems from the cumbersome process of assigning experts to writing groups, reaching a consensus on new treatments/findings, writing or rewriting the document, and circulating it broadly for comment and revision before publication. Some societies have adopted a "focused update" approach, whereby if there are important new clinical trial results, the guidelines for that section of the document can be updated more quickly without have to address the entire document each time. As to physicians in the community, their willingness to change practice may relate to their familiarity with the results of new trials, the strength of the findings in the trials, and how much change the new recommendations might require from their current practice.
The Role of Sub-analyses
Kirtane: There can definitely be a temptation to overemphasize the importance of subgroups, but some of this is because of a desire to individualize and personalize clinical decision-making for patients, rather than simply homogenizing trial results across a pool of patients.
Curtis: When one has expended a great deal of effort on a clinical trial, which involves expense as well as patients' and healthcare professionals' time, it is disappointing when the hypothesis of the study is disproven. While there can be an honest effort to see if anything important has been missed by doing sub-analyses, there can also be a reluctance to abandon a cherished opinion about what treatments should work for patients.
Friday Feedback: Do Physicians Drag Their Feet in Taking New Research into Practice?
Aucun commentaire:
Enregistrer un commentaire