After more than 20 years, the conflict of interest (COI) movement has failed to substantiate its central claim that interactions between physicians, researchers and the medical products industry cause physicians to make clinical decisions that are adverse to the best interests of their patients. The COI movement's instigators have produced no solid evidence of harm commensurate with their extravagant allegations. At the same time, they have diverted resources away from more worthwhile pursuits, such as basic and applied medical research, clinical care and medical education towards onerous compliance exercises and obtrusive laws. Perhaps worst of all, they have made it respectable to ignore the epistemological foundations of medical science, diverting attention away from the scientific merit of the information presented and focusing it instead on the identity and motives of those who present the information.
The debate begins
Just over 20 years ago, two commentators warned the medical profession of insidious integrity threats. One  raised the hue and cry of corruption, supposedly engendered by financial ties with the medical products industry under the code name ‘financial conflict of interest (fCOI).’ The other  warned of a ‘new McCarthyism’ where scientific research would be judged on the basis of who funded it, rather than the rigour or validity of the research itself, a subversion of the epistemological foundations of science .
At the time, it was unclear which view would prevail. Would the relevant authorities demand a cool and considered evaluation of the evidence and a balanced, point-counter point debate? Or, would they allow one side to sweep the field, dominating the discussion and creating a propaganda echo chamber, granting one side a patina of truth while denying the other a voice?
The current state
We now know that the message of the fCOI movement instigators has triumphed and, as exemplified by Tagore's article  that the fCOI narrative has buried its opposition in an avalanche of one-sided rhetoric, forming what behavioural economists call an ‘availability cascade’  of industry vilification and unsubstantiated accusations . Yet, if we set aside the endless string of dubious opinion pieces, surveys and sensationalist editorials, the relevant facts in the debate may be relayed and considered in short order. You may find, as we have, that they do not support the unambiguous indictment of industry's motives and actions that has been promoted by the conflict of interest movement.
Their basic argument boils down to one dubious assertion: The medical products industry causes physicians to make prescribing decisions that conflict with the best interests of their patients.
Even fCOI narrative instigators admit that no data are currently available to ascertain the truth or falsehood of this statement . Many object to this influence on ideological grounds, claiming that physician's prescribing decisions should be based solely on information provided by totally disinterested parties. Yet purging medical product marketing is shortsighted since decision-makers make better decisions when they have more information, even if that information comes from biased sources. Chressanthis et al.  have verified this truism using real physician prescribing data to assess the impact of restricting drug representative (drug rep) interactions on physicians responses to the release of a new first in class diabetes medication, a black box warning and negative clinical trial results for a new combination lipid therapy. The results unequivocally showed that restricting drug rep access to very low levels resulted in slower responses to all three forms of news relative to physicians who allowed even low or medium access levels. The evidence indicates that drug rep access provides physicians with at least some useful information. Unfortunately, the Chressanthis study does not tell us about patient outcomes which, of course, are the only outcomes that really matter.
So, the debate about information-bias rolls on. We must necessarily dig into the various proxy measures for patient outcomes and purported mechanisms of influence that have been postulated by fCOI instigators.
The instigators narrative uses a variety of proxies to substitute for patient outcomes. They include variables such as the cost of medical products, their degree of novelty, and the emergence of unforeseen risks. While each of these relates to patient welfare, each proxy is problematic in its own way.
Costs invariably affect patients, but considering the cost of a medication in isolation is meaningless. New medications, which are still under patent protection, are more expensive than older generic medications, yet that fact alone does not make them a worse choice. Additional cost is often justified in particular patient populations because of improvements in the efficacy of the therapy, its side effects or its method of administration (pill rather than injection, fewer doses per day, etc.). Furthermore, large-scale econometric analyses [9-14] demonstrate that, on average, newer drugs pay for themselves by reducing non-drug healthcare expenses, such as hospitalisations. Even absent such savings, they frequently add value to the healthcare system by improving patients’ life expectancies or quality of life by reducing disease specific morbidity and mortality and by enhancing economic productivity [15-18].
As in many other industries, the drug industry makes much of its therapeutic progress through incremental innovations. Some academics deride incremental innovations as ‘me-too’ products, but by doing so they overlook the importance of incremental progress in any innovative area, the value of having multiple drugs within a single class for tailoring treatment regimes to patients and cost competition, and the fact that these drugs usually result from a race to develop new medications based on a common mechanism, rather than imitation of competitors’ products .
Critics of the medical products industry often claim that industry invents new diseases or medicalises normal life states in order to expand markets for their drugs [20-22]. These claims are dismissive of patient suffering. They ignore the near impossibility of convincing the FDA that a product is safe and effective for treating a non-illness. Companies would only hurt themselves if they selected business plans that included creating prescription-only treatments for conditions that do not exist . If a condition does not exist, patients will not seek treatment for it and physicians will not prescribe products to treat it. The most credible evidence provided by such critics are the rise in diagnostic rates for obscure conditions following the availability of treatments for those conditions . However, this rise in diagnoses is because of doctors having a new way to treat dysfunction and distress, and thus have a reason to diagnose, rather than companies creating illusory conditions.
For instance, contrast Goldberg and Greenberg who both comment on the rise in depression diagnoses following the emergence of Selective Serotonin Reuptake Inhibitors and Serotonin – Norepinephrine Reuptake Inhibitors. Goldberg explains the rise in diagnoses as a result of doctors having a set of new, safer tools with which to treat depression, facilitating treatment of moderate and even mild cases of depression. As diagnoses are generally only given when treatment is available, the availability of safer treatments may lead to a rise in diagnoses. In contrast, Greenberg claims that pharmaceutical companies gulled the public into seeking treatment for normal, transient sadness. We find Goldberg's explanation more plausible, and therefore more compelling, than Greenberg's conspiracy theory.
Financial conflict of interest instigators often criticise the use of new medications because of unknown risks. However, the rate at which problems arise is quite low. Of FDA approved drugs, 8.2% acquire black box warnings and 2.9% are withdrawn . As discussed earlier, econometric analyses suggest that the benefits of new drugs sufficiently improve patient outcomes to reduce use of non-drug medical services and offset somewhat their higher costs. As these improvements are aggregate over all new products, the low rate of unforeseen risks hardly justifies arbitrarily avoiding recently approved therapies.
The fCOI narrative dedicates much of its time worrying about how marketing effects physicians, without ever showing that this influence is itself harmful. While we believe that such a focus is misguided, we will review the various proposed mechanisms of influence here and the shortcomings of each theory.
Drug reps visit healthcare providers, primarily physicians, and provide them with the ‘details’ of their products, in an exchange commonly referred to as ‘detailing.’ These visits have been widely condemned by fCOI instigators (e.g. [26, 27]). Physicians, in contrast, report that these visits are quite useful  and the evidence from Chressanthis et al.  support this view. The difference is perspectival. Detailing provides patient-care doctors with convenient access to FDA regulated information about new products and product-updates through expedient face to face encounters. There is no evidence that these visits harm patients. In the absence of such evidence, the debate boils down to an ideological squabble.
When discussing detailing, critics often focus on small conveniences and reminder items used by drug reps to facilitate getting doctors’ attention during detailing encounters . These items formerly included pens, notepads, and free meals. Since the advent of the pharmaceutical and device company trade organisations’ (PhRMA and AdvaMed) ethics codes in 2008, modest meals are still provided but pens, notepads and other ‘reminder items’ are no longer distributed [30, 31]. The PhRMA and AdvaMed bans on reminder items arose in reaction to the accusations of industry critics that such ‘gifts’ represented a covert form of bribery, in which physicians essentially wrote scripts for particular products in exchange for more free pens, food and perhaps even subsidised travel to a conference. The bans were intended as a public relations manoeuver to improve industry's image, but were spun by critics into an admission of guilt.
The quote ‘advertising is repellant to people of delicate feeling,’  succinctly captures the motivation behind the fCOI movement's assault on industry advertising. The instigators recoil from commercial communication, favouring instead the views of ‘impartial experts.’ But the fCOI instigators have their own agendas, favouring generic products and thereby the interests of insurers – for example – who sponsor them. Their criticisms of branded-product marketing fall far short of the level of evidence needed to suppress free speech in the ways they have suggested, at least in the United States. It is true that all advertisements raise awareness of particular products independent of their therapeutic value. It is true that direct to consumer advertisements increase requests for particular products. Such requests increase prescribing of the requested products . However, because of substitution rules, many of the patients who request a branded product actually receive a different drug for the same condition . Barring the unlikely event that all individuals requesting drugs by brand name are already on products for the same condition, such requests serve the public good by increasing treatment rates for the conditions for which products are advertised, reducing the portion of the population which goes untreated for particular conditions and improving public health wherever such treatment is appropriate.
Advertising aimed at physicians also plays an important role in public health. It raises physician awareness of the existence of new therapies in a concise format. The fact that these advertisements are sometimes presented in a lowbrow or even misleading manner or fail to support all claims with high level evidence, while disappointing to purists, is largely beside the point. Advertisements are attention-getters and should not be mistaken for scientific evidence. More empirically, solid sources should be used to evaluate the quality of a new product before deciding whether to prescribe it.
Companies pay physicians to speak to other physicians about new indications and new products, usually drugs or devices, in a comfortable setting, such as a restaurant. The content of these speeches is regulated by FDA as a form of marketing, meaning that speakers may only discuss products’ on label indications and must spend equal time on positive and negative aspects of the products. These regulations mean that physician peer-to-peer speaking is among the most heavily regulated forms of speech in the United States.
Doctors attend these events voluntarily for the simple reason that they want to learn about new products and find it useful to obtain such information from other physicians. While such events provide educational content, they should not be confused with accredited Continuing Medical Education (CME), which we discuss next. Promotional speaking events are marketing activities which are designed to raise awareness of and stimulate interest in products. As with advertising, these events provide a valuable service by raising awareness of new products, but should not be treated as stand-alone sources of information. Physicians should always consult other resources before prescribing a new product.
The medical products industry provides 28% of ACCME-accredited CME funding (excluding income from advertising and exhibits income) in the United States, which is used to support 18% of CME activities . Of course, such funding does not flow from pure altruism. It is motivated by a desire to increase prescriptions for its products. However, such commercial interest is only allowed to manifest itself in a limited number of ways . The main route is through the selection of a general area of medicine which the CME event will focus on (e.g. depression, bipolar disorder, schizophrenia, etc.). Commercial funders are not permitted to select speakers or to fund CME activities that focus on particular products. Three large surveys of CME participants indicate the commercially sponsored CME is not perceived to be biased [36-38]. Despite such evidence, the instigators routinely attack commercially sponsored CME on the mistaken grounds that such education is biased [39, 40]. They draw on a nirvana fantasy to claim that commercially sponsored educational events should be banned and replaced by ‘unbiased’ alternatives. The actual results of attacks on industry funded CME demonstrate that reducing industry sponsored CME reduces the overall rate of CME activities produced, with disproportionate impacts on providers outside of major cities. Clearly then, these attacks are not serving patient interest, but are in fact undermining them by depriving providers of high quality educational content. As CME has been tied to patient outcomes (e.g. ), this reduction in CME because of ideological attacks translates into forgone patient benefits, if not patient harms.
Proxies in mechanisms
Much of the research on fCOI in medicine focuses on what are effectively proxies or surrogate end-points tied to the mechanisms discussed above.
Lawsuits and settlements
In recent years, newspaper headlines sensationally reporting large payments by the pharmaceutical industry to settle claims of off-label promotion and related charges have become commonplace. These headlines seem to tell us how the government caught the pharmaceutical industry breaking the law and punished it. The real story is more complicated. These lawsuits are based on a legal strategy that virtually guarantees that companies will settle, regardless of their merit. Most are ‘whistle-blower’ cases, giving a former employee a large financial stake in the outcome. His or her chances of a large payday are enhanced when the government decides to join the prosecution. The cases are brought under the False Claims Act on the premise that pharmaceutical marketing and industry interactions induce physicians to prescribe products off-label, allegedly creating a ‘false claim’ for reimbursement from federal healthcare programs. To be proved in court, the prosecution must show that the pharmaceutical industry caused physicians to prescribe off label, in the scientific ‘but-for’ causal sense. In other words, if the off-label prescribing might reasonably have occurred without the company's marketing, then the case should fail. Rather than bear this burden of proof, the government threatens companies with a fatally strong enforcement remedy – debarment. Were a company to risk trial and lose, it could be banned from selling its products to the federal government. Given the large share of the pharmaceutical market controlled by federal insurance programs from Medicare to CHIP and the VA system, such a ban would amount to a death sentence, effectively killing the US branch of any pharmaceutical company that received it. This strategy clarifies why pharmaceutical companies ultimately settle these cases, even when the cases are weak . The large settlements enrich the successful whistle-blower, but the associated costs are passed on to consumers (and to the federal government too, ironically) in the form of higher product prices.
These lawsuits mislead the public into believing that the medical products industry is exerting a toxic influence on physician prescribing decisions through mechanisms such as detailing and advertisements. Indeed, prosecutors sometimes editorialise to this effect, ‘Off-label marketing can undermine the doctor-patient relationship and adversely influence the clear judgment that a doctor's patients have come to rely on and trust’ (United States Attorney Zane David Memeger, as quoted in ). Despite such opining for the press, there is little evidence that these lawsuits reveal physicians making decisions that conflict with their patients’ best interests.
Academic journals publish articles based on discovery documents from the aforementioned lawsuits (e.g. ). These documents sometimes include copies of manuscripts prepared by professional medical writers for publication under the authorship of prominent academics in the field. These documents raise concerns that the authors listed on the publications are not truly responsible for the study results. Consequently, most journals and academic institutions have instituted rules banning ‘ghostwriters’ or undisclosed professional medical writers. While such bans represent an understandable reaction to allegations of ‘ghostwriting,’ and we agree that the contributions of anyone working on a paper should be disclosed, there is little to no evidence that the use of professional medical writers is associated with bias, much less research misconduct. The best available evidence on the accuracy of articles involving professional medical writers comes from a recent study by Woolley et al. , which examined retractions for misconduct from the PubMed directory and found only 3.8% of article retracted for misconduct had any industry involvement, only 1.4% involved declared professional medical writers and none involved both. Retractions for misconduct rather than mistake was significantly associated with absence of both industry involvement (odds ratio 0.25; 95% CI 0.11–0.58) and medical writers (OR 0.16; 95% CI 0.05–0.57). So, if both industry involvement and the presence of declared medical writers reduce the likelihood of retractions for misconduct, relative to retractions for mistakes, then it seems unlikely that these articles are on the whole less accurate than others. If they are similarly accurate to other articles, than there is no reason to conclude that the presence of medical writers, whether disclosed or undisclosed, distorted the literature in way that detracted from the quality of care physicians rendered to their patients.
The fCOI instigators have devoted significant effort to evaluating whether industry funded randomised clinical trials (RCTs) differ from RCTs funded by not for profits in their rate of positive outcomes, rate of adverse event reporting or the quality of the trial . The results of such analyses have been quite variable. Industry sponsored trials do appear to report a higher rate of positive outcomes than non-industry funded studies . This, in itself, is not evidence of corruption because the higher rate may be resulting from legitimate differences in study protocol such as industry's use of larger sample sizes or placebo vs. active comparator controls . These two factors should at least partially explain the difference in positive results as greater statistical power reduces your chances of rejecting true hypotheses and beating a placebo control is usually easier than beating an active comparator. Other factors may also help to explain this difference. For instance, because the financial solvency of the pharmaceutical industry depends on its ability to select only the most promising drug candidates for clinical trials, it is quite likely that the compounds entering industry sponsored clinical trials have been more thoroughly vetted and therefore are more likely to be safe and effective than drug candidates entering academic trials . This factor alone could explain the gap in positive findings without resort to accusations of biased reporting.
Regardless of whether these factors are sufficient to explain the difference in the rate of positive findings, showing that industry funded studies are more likely to have positive outcomes cannot answer the question it is meant to answer, which is ‘Are industry funded studies scientifically unsound?’ By alleging an affirmative answer, the instigators are committing what Daniel Kahneman calls substitution . They are substituting an easier question ‘Are industry funded studies more likely to report positive outcomes?’ for the harder question ‘Are industry funded studies scientifically unsound?’ The answer to the former question is at best a weak and non-validated proxy for the latter question. In the absence of the assumption that industry results are more likely to be positive because of some sort of misconduct, no reason exists to consider this finding relevant to the integrity of industry sponsored medical science. It might simply indicate that industry is selecting therapeutic agents which are efficacious.
Furthermore, the present evidence does not support fCOI instigators claim that the disparity in the rate of positive findings between industry funded and non-industry funded studies causes physicians to make decisions that conflict with the best interests of their patients. The critics are essentially arguing that because industry ‘biases’ the scientific literature through the disparity in published positive findings, it alters the evidence base on which physicians make their prescribing decisions. If the medical literature was corrupted by deliberate distortions produced by industry, as the critics allege, then physician prescribing decisions would like be altered. However, the critics have neither shown that the scientific evidence presented in industry funded studies is faulty or that it improperly leads physicians to make erroneous prescribing decisions. Their conclusions rest instead on flimsy evidence, rhetorical wizardry and an ideological animus against the medical products industry.
Financial conflicts of interests
In violation of the epistemology roots of science, which suggest that ideas and evidence should be weighed independently from the person who presents them, industry critics have encouraged a capricious discounting of all industry funded research and the expert opinions of anyone with industry ties. They justify this ad hominem dismissal of empirical data and otherwise credible expert opinion by citing the disparity between publicly traded companies’ fiduciary duty to pursue stockholder profits and physicians’ duty to place patients’ interests before their own. They claim that this disparity leads industry to deliberately deceive physicians so that they will prescribe medically unnecessary products and that it leads physicians with industry ties to place corporate profits above patient interests. They support these startling accusations with fallacious reasoning, superficial analyses, and misleadingly framed anecdotes.
One such anecdote, frequently cited by fCOI instigators as a ‘smoking gun’ demonstrates how the framing of these anecdotes leads to misleading conclusions. A 1998 study found that authors with financial ties to drug manufacturers were more likely to endorse the use of calcium channel blockers, which some believed to be unsafe, than those without such ties . This was interpreted as a case where financial ties caused physicians to endorse a product they otherwise might not have. Yet, the association is confounded since these authors’ views could also be explained by their greater seniority, prominence and experience. A separate study found that industry consultants are, in general, more productive across numerous metrics than their unaffiliated colleagues, suggesting that the presence of consulting relationships may indicate expertise, rather than bias . More importantly, the endorsement of calcium channel blockers has been vindicated by subsequent research, which shows that calcium channel blockers match diuretics and ACE inhibitors and surpass beta-blockers in safety and efficacy . So, what has been alleged as a case of fCOIs biasing authors may be explained by the authors’ superior experience and knowledge? The same knowledge and experience may also explain why companies consulted with these individuals in the first place.
Closer analyses of the other anecdotes cited by the fCOI instigators reveal that they are similarly flawed. Anecdotes are considered unscientific precisely because they are subject to such framing effects. Furthermore, by extrapolating from a handful of misleadingly framed anecdotes, the critics commit a severe form of denominator neglect in which they ignore the innumerable instances in which interactions between physicians and the medical products industry has produced undeniable good. In clinical settings, these benefits largely accrue through the transfer of information in a convenient and heavily regulated format from industry representatives to physicians. In research settings, the results accrue primarily from collaboration during innovation attempts and research funding.
In fact, nearly every major medical advance over the past century has involved industry and most have involved physician-industry collaboration. These collaborations are the lifeblood of medical innovation. We should not restrict or eliminate them based on vague feelings of unease or ideological arguments. Their benefits are tangible. Without tangible proof of harm, these relationships should continue.
Medical innovation is incredibly difficult. Only 1 in 5000–10,000 compounds that enter testing are approved by the FDA for therapeutic use in humans . Likewise, only 10% of drugs that enter clinical trials make it to market . The best available data on the cost of drug development indicate that it costs companies upwards of one billion dollars per product that reaches market, mostly because of the high failure rate . So given the difficulty of medical innovation and the importance of physician–industry interactions for that innovation, why are we making those interactions more difficult?
Things are getting worse
When we first started writing about this in 2005 [56, 57], the fCOI instigators were just starting to shift from a narrow focus on decrying physician-industry research collaboration to demonising all physician–industry interactions. In recent years, this shift has reached new heights and certain commentators have gone as far as calling for a complete apartheid between physicians and researchers and the medical products industry.
As this review reveals, the conflict of interest movement has failed to substantiate its central claim that interactions between physicians, researchers and the medical products industry cause physicians to make clinical decisions which are adverse to the best interests of their patients. After 20 years of impugning the motives of industry and demeaning the professional judgment of physicians, the instigators have failed to produce solid evidence of harm commensurate with their extravagant allegations. At the same time, they have diverted resources away from more worthwhile pursuits, such as basic and applied medical research, clinical care and medical education towards onerous compliance exercises and obtrusive laws and regulations. They have propagated an availability cascade that projects the superficially plausible message that the medical products industry is coopting patient care by corrupting physicians and researchers. Perhaps worst of all, they have made it respectable to ignore the epistemological foundations of medical science, diverting attention away from the scientific merit of the information presented and focusing it instead on the identity and motives of those who present the information.
We dedicate this commentary to the memory of the late Dr William Glazer, a leading mental health professional, who died last June. Unlike most physicians who have passively allowed the conflict of interest narrative to wash over them and succumbed to its resultant regulations, Bill challenged the narrative and urged his colleagues to resist [58, 59].
All authors participated in the design, writing and editing of the manuscript. All authors approved the final draft.
LS received no funding for this article. DB and TS's contributions to this commentary were supported by a grant from the Searle Freedom Trust. Only the authors were involved in this project.
DB and LS have no pertinent relationships to disclose. TS is a scientific founder and director of BioAegis Therapeutics and a director of Velico Medical Corporation. His employer, Brigham & Women's Hospital has licensed technologies concerning which TS has patents to those companies. He currently receives no financial compensation for these relationships.