A recent article from the Archives of Internal of Medicine has caused a number of media sources to discuss commercial support of continuing medical education (CME). The article, written by academics from the University of California, San Francisco (UCSF), focused on the results of a survey of under 800 participants at a series of 5 live HIV CME activities. The information held mixed results as some of the questions focused on the potential for bias in single supported CME activities up to 86% of all participants and the consensus was that commercial support of CME was essential 56% of physician participants. The survey included 55% physicians, the remaining participants included nurses, physician assistants, and PhD’s.
It should also be noted that a similar study conducted by researchers from UCSF (including Michael Steinman author of the archives paper) and published last year, found that 97% of respondents stated that the activity they attended was free of commercial bias regardless of whether the activity received commercial support or not.
The authors perhaps overstated their position in the very first sentence with this statement “pharmaceutical and medical device industry funding supports up to 60% of CME costs nationwide.” But the authors clearly know this data is from 2006, and does not reflect the significant reduction over the past several years in commercial support of CME. Specifically, since 2007, commercial support has declined $355 million or 29.3%. Moreover, the total percent of commercial support as part of the overall CME budget dropped from 47.5% (2007) to 39.0% in 2009. Fewer companies are funding CME programs as well. Thus, do the author’s clearly frame their article with bias using out of date data, which they intentionally chose to use to bolster their claims?
Consequently, the authors then go into a discussion about reports from the Josiah Macy Jr. Foundation and the Institute of Medicine (IOM), which have recommended that all commercial support of CME activities be eliminated within the next 5 years. These reports recognized “that such a system may involve higher costs for physicians and require cost-cutting steps by education providers.” As a result, the authors sought to determine whether participants at live CME activities understand some of the costs involved, whether they would be willing to pay higher registration fees or accept fewer amenities to avoid or decrease the need for commercial funding, and how that willingness relates to their perception of bias related to commercial support.
The authors conducted surveys at a series of live 1-day CME courses delivered by the International AIDS Society–USA (IAS-USA) from January through June 2009. The IASUSA is a not-for-profit medical organization that delivers CME programs for human immunodeficiency virus (HIV) specialists. The organization requires that commercially supported programs receive unrestricted educational grants from several companies with competing products in the field.
One of the problems with this study is that the respondents to the survey are clearly not representative of physicians as a whole. Specifically, these were respondents who deal with HIV/AIDS patients, a very niche practice. These are not general practitioners who deal with a wide range of diseases, conditions, and treatments, and who rely on CME to stay up to date on a vast array of medical information and data. Moreover, we are not told how many of the 400 physicians are in private practice vs. those in training or in academic centers. Knowing this is important because those in training and those in academic centers do not rely on CME for continuing education.
Such an unrepresentative sample makes the results from this study skeptical. Especially considering three other studies from last year used more significant samples sizes and a more representative sample of physicians. All three studies found a lack of commercial bias in industry-sponsored CME (Cleveland Clinic; Medscape, and UCSF).
The survey items targeted 4 main areas (1) beliefs about commercial funding and potential for bias (10 items); (2) willingness to offset the cost of commercial support (10 items); (3) knowledge about some of the costs associated with providing a CME course (9 items); and (4) demographic information, including years in practice and types of interaction with industry in the prior 3 years. The survey items targeting the beliefs about commercial support and potential bias in CME were rated on a 4-point scale (eg, “no potential for bias” to “large potential for bias”).
Respondents reported that the greater the percentage of commercial support for an event, the greater was their perception of potential bias. For example, only 7% of physicians thought there was substantial (moderate or large) potential bias in activities without commercial funding, whereas 46%, 80%, and 86% reported substantial potential bias when a single commercial supporter provided 20%, 60%, or 100% of the activity total costs, respectively.
“Substantial potential bias.” What does this even mean? While the authors have not disclosed the exact questions they asked respondents, one can only imagine: Is it possible that when a CME program is funded with commercial support, there is a potential for bias? Of course people would answer yes, ANYTHING is possible!
Is there potential that a car driving today hits you? Yes. Respondents saying there is a “substantial potential bias” equates to meaning, “bias is possible.” There is always bias in everything, and thus measuring its potential is insignificant.
Moreover, this study did not ask participants about whether a particular activity was biased. Instead, they just asked whether there was a “potential” for bias in any commercially supported CME. Hardly a scientific question or way to measure what, if any, bias there could be in a CME program.
Overwhelmingly, physicians exposed to commercially supported CME do not perceive bias, even though they may well acknowledge the abstract possibility of bias. They may even acknowledge the possibility that a presentation may be biased and that they might misperceive it. However, what the authors need, but don’t have, is evidence that physicians are poor detectors of bias AND that the investigators, by contrast, are excellent detectors of bias. Surveys and subjective opinions cannot give such evidence.
The authors also noted that 70% of physicians perceived substantial potential bias when 2 or more commercial supporters completely funded the activity, whereas 86% reported bias when a single commercial supporter did so. In addition, 73% of physicians perceived moderate to large bias from faculty members on commercial speakers bureaus and from faculty receiving research support from industry 68% compared with faculty who receive no funding from pharmaceutical/medical device companies 5%.
The authors also found that fewer than half of the physicians (42%) were willing to pay increased registration fees if it meant commercial support would be less or nonexistent. In addition, physicians said that in order to reduce the amount of money industry spends on CME:
- 56% supported posting course curriculum online instead of in print
- 50% supported moving educational events to "less desirable" venues
- 50% supported ditching free food and snacks
The least desirable strategies for decreasing costs were to provide fewer topics and speakers (11%) or to credit fewer CME hours (15%).
Moreover, 56% of physicians agreed or strongly agreed that commercial support is essential for accredited CME and should not be eliminated, while only 17% favored eliminating such support. With over a majority of physicians agreeing that commercial support is “essential” for accredited CME, the authors seem to ignore this result and focus mainly on the “potential for bias.” Clearly, the authors had predetermined what they wanted to prove, and were only interested in why commercial support leads to potential bias, and not why 56% of physicians find value in commercially supported CME.
While this unrepresentative, small sample of physicians attempts to show that participants perceive a “potential” for bias in commercially supported CME, the authors attempt to connect this to physicians unwillingness to offset costs by paying more for CME seems misplaced.
The idea that physicians are unwilling to pay more for CME, and thus not reduce the “potential for bias,” is one way to state it. A different way at looking at the data is that because 56% of physicians agree that commercial support of CME is “essential,” there is no need to pay more for CME because doctors find value in commercially supported CME.
Moreover, as the authors recognize, there are several approaches that are currently used to minimize the presence of bias in CME activities. The Accreditation Council for Continuing Medical Education (ACCME) sets standards and guidelines for monitoring of commercial influence, including tracking of declared of conflict of interest as well as for resolution of that conflict. In addition, tools have been developed to help CME providers anticipate bias in activities that are at greater risk of bias.
Despite these strong safeguards, the authors again use ACCME data from 2006 to suggest that 60% (now 40%) of revenue for commercial support of CME raises concerns “whether it is possible to prevent substantial commercial influence with such a funding structure.” With a 20% decrease in funding over the past 3 years, such questions and concerns can hardly remain, especially given the strict ACCME guidelines.
What is also problematic about the author’s paper is they clearly admit that there “is little direct evidence about the degree to which commercial support of CME activities introduces bias.” If there is little evidence, and since their study produces none, on what basis do they assert their claims that commercial support of CME introduces bias? None. More importantly, they point to no evidence that commercially supported CME has ever harmed a patient.
Instead, the authors cite to “indirect evidence” that commercial support of CME shapes “both topic selection and presentation of information favorable to a company’s products or unfavorable to their competitors’ products.” Such indirect evidence seems misguided for a peer-reviewed article, and could hardly be called evidence-based.
Consequently, the authors discuss ways to provide quality CME either with alternate funding or at reduced cost. Some have suggested reducing costs by holding meetings and events at less expensive facilities and locations, or reducing speaker honoraria. However, as the authors note, “a rapid reduction or elimination of funding might be unacceptably disruptive, and some have postulated that such a change will result in the disappearance of live CME as we know it and the development of other forms of CME.”
The authors also point to an IOM study that estimated that if commercial supported CME is eliminated, the cost for physicians would go from $1,400 to $3,500 annually. However, we are not told how IOM came up with these numbers nor what factors they took into consideration.
For example, eliminated commercial support will likely reduce the geographic options for many physicians, meaning physicians will have to travel farther, take more time off from their practice, spend/lose more money. In addition, less commercial support may mean that doctors have less programs they want to attend that address the specific issues in their field or training. And of course, it does not address the issue of less education and training, and the negative impact this could have on patients.
Finally, a commentary to this article noted that while studies have shown that physicians “perceive little commercial bias in the material presented, this does not mean that commercial bias did not exist, just that the attendees did not perceive significant commercial bias.” What does this mean? These authors talk themselves into circles.
Perceived bias that exists or does not exist, what kind of medical degree or training do you need to assess whether there is bias in medical education? Do we not put our physicians through enough training and education or is this bias so subtle we need a microscope or PET scan to find it?
The problem with this study is that the researchers from UCSF put surveys into the hands of doctors who were not only unaware of the costs of putting together a CME program (and specific costs such as accreditation, compliance, oversight, logistics, etc.), but were unaware that the present accreditation system and the firewalls from FDA, ACCME, and HHS/OIG are more than adequate to protect from bias.
While the commentary calls for “minimizing bias,” the study did not prove there was any bias in CME programs, only that some physicians thought there was a “potential for bias.” In fact, this survey did not prove there was bias in any CME at all because it did not ask about a specific CME program or activity. Therefore, the authors of a study that says that healthcare professionals are aware that there might be a “potential” for bias, that does not in any way demonstrate that there is in fact any bias, jump to the conclusion that “we need to explore alternative models for CME programs and funding sources.”
This is a little like saying, “There is the potential for airplanes to crash, therefore I should take another form of transportation the next time I need to get from New York to Los Angeles.” All the authors did in this study was make inferences about “potential bias.” Such inferences were not reasonable because they had no evidence and were based on pure conjecture and speculation.
The future of CME depends on collaboration and innovation with multiple stakeholders to implement many of the goals and initiatives contained within the Affordable Care Act (ACA). At a time when we are experiencing shortages of primary care physicians and our health care system is bleeding money due to inefficiencies, we should be spending our resources on finding ways to produce high quality care at an affordable cost, and focusing on outcomes. The growing epidemics of obesity and diabetes will continue to cost our country billions of dollars, and educating our physicians throughout their lifetime will be critical to improving these horrible diseases.
Working with industry to provide education and CME is valuable for a number of reasons, none of which these authors examined. Had they asked these questions in the survey, they would have found out why 56% of physicians feel commercial support of CME is “essential.”