“Subscription science” and real conflicts of interest

Financial conflicts of interest (COIs) have long been recognized as a problem in science and medicine, and, over the last decade or so, increasing efforts have been made to minimize the potential bias caused by COIs. For the most part, rather than banning gifts and payments from big pharma, the approach has been transparency, with journals, conference organizers, and the like requiring disclosure of such payments received and the federal government tallying how much physicians have received from pharma and publishing the payments in the form of a searchable database. (As an aside, I searched for my name, and all that came up was a Florida optometrist with the same name. I was relieved that the database accurately reflected my knowledge that I have not received any payments from industry.) Unfortunately, as I’ve lamented from time to time, a depressing number of my fellow physicians seem to think themselves “above it all” in that they deny that COIs affect them, even though research is very clear that even small gifts can affect decision-making. Physicians are human, as much as we might like to think ourselves as better (or at least able to resist subconscious influences to which humans are prone) and thus just as susceptible to influence as anyone else, be it through financial or social incentives, no matter how much some of us might portray ourselves as medical monks, immune from base considerations of finance and human influence.

Although concern about financial COIs, in particular payments from pharmaceutical companies and device manufacturers, have resulted in significant action to provide at least some transparency, there is another form of COI that is often financial that has received less attention. I was reminded of this form of COI as I cursed Benjamin Mazer and Michael Rose for having beaten me to getting a commentary published in the peer-reviewed scientific literature about it, Subscription science: how crowdfunding has become a conflict of interest. I’m not sure I agree with the name that Mazer and Rose came up with for the phenomenon, but, with a minor quibble here and there (and, me being me, you know I’ll include the quibbles in my discussion), I’m in major agreement.

Let’s first set the stage:

Conflicts of interest are an ongoing threat to medical practice. Studies have shown that doctors’ treatment decisions can be directly swayed by industry payments.¹ Pharmaceutical and medical device companies also assert indirect influence through the funding of thought leaders and guideline authors.²³ Governments and professional bodies have attempted to rein in the impact of conflicts of interest in recent decades. Most medical journals now require standardised conflict of interest declarations.⁴ The US Physician Payments Sunshine Act requires manufacturers of drugs and devices to publicly report payments to physicians.⁵ These efforts have curtailed the overall number of US physicians receiving payments from industry, although large payments of more than $50 000 (£39 644; €45 560) might have increased.⁶ Financial relationships, even substantial ones, are not always nefarious. But patients, doctors, and regulators are right to be wary of the power of the purse. Crowdfunding has become another potent but overlooked source of revenue for physicians who want to share their views on medical topics. Crowdfunding refers to payments that are voluntarily contributed by individuals who would like to support an author’s work. Digital platforms such as Patreon, Substack, YouTube, and Twitter allow fans to offer recurring payments to healthcare professionals who produce opinion articles, explanatory videos, and podcasts. Although crowdfunding is often seen as a principled alternative to industry influence, these payments may also create meaningful conflicts of interest.

Surprisingly, for the most part, professional societies, conference organizers, and journals appear not to consider these sources of income as significant financial COIs. At least, I haven’t seen physicians and scientists publishing in peer-reviewed journals or speaking at conferences disclosing these sources of income. (Feel free to post counterexamples in the comments.) However, as I’ve been saying for a long time now, these sources of income can be substantial. In particular, Mazer and Rose single out Substack, which is entirely reasonable given its subscription model and why I noted a year and a half ago how Substack seems to be where all the quacks go. At that time, I characterized Substack as another example of how “everything old is new again,” noting how, in essence, in essence Substack is a monetized blogging platform, noting how it billed itself as “the home for great writing” and the “simplest way to start a subscription publication, and the best place to grow it.“ To me, the only difference between Substack and Blogger or WordPress was its option to monetize one’s posts through a subscription model, with Substack getting a 10% cut of the proceeds (although Substack also charges an additional 3% for its credit card processing service, which is necessary, you know, to actually process subscription payments). Moreover, even as a lot of quacks, antivaxxers, and cranks have flourished on Substack, so have a lot of excellent blogs, such as Timothy Snyder’s history blog, which I highly recommend.

I hasten to add that, in general concept, there’s nothing inherently wrong with Substack’s business model—on the surface, at least—which is to provide a platform that is easy for bloggers to use and, if they so desire, to try to monetize their work through a subscription model, from which Substack apparently takes what seems to be a reasonable cut. Moreover, you don’t have to monetize your blogging if you don’t want to. Like Blogger and WordPress, Substack does host free blogs. I’ve also noticed that nearly all bloggers who offer subscriptions publish a mix of “free” content that nonsubscribers can access and subscription-only content that requires payment.

I’ll even mention a little “secret” (although maybe it’s not such a secret anymore given how many times I’ve mentioned it): I’ve periodically toyed with the idea of moving this blog over to Substack as a free blog, mainly to eliminate the monthly expense of hosting it. As I like to remind readers I make zero money from this blog. Quite the contrary! When one of my posts goes viral, I end up paying Cloudflare for the added bandwidth charges out of my own pocket. Also, I am at best a WordPress/HTML amateur. I really don’t know what I’m doing on the technical end, and when things go wrong with the website—as they nearly inevitably do with any website, with this blog seeming to get glitchier with age and my often not knowing how to fix it—getting things running again is difficult. What’s usually stopped me is a combination of inertia and a dislike of the “look” of Substack blogs. Truth be told, I find Substack’s website to be ugly, with few ways to make it less ugly. True, the available templates have improved. They’re not as bad as they were a year ago, but they’re still pretty ugly, by and large.

Subscription science as a COI

With that background, let’s see what Mazer and Rose say about “subscription science”:

Physicians do not always agree about which third party payments are unacceptable. Yet two variables typically come into play: the size of the payment and how strongly the payer is associated with a specific scientific viewpoint. Crowdfunding poses a risk in both domains. For example, Substack, an online newsletter platform, is increasingly used by physicians to write medical commentary, with some newsletters reaching 10s of thousands of subscribers.⁷ Substack estimates that 5-10% of readers will upgrade to a paid tier, and paid subscriptions on the service cost a minimum of $5 per month.⁸ Although $5 sounds negligible, consider a newsletter with 10 000 total subscribers, 1000 of whom pay a $5 monthly fee. After subtracting Substack’s 10% cut, a doctor could expect $54 000 in annual payments. If physicians accrue 5000 backers, they can expect $270 000 in revenue. This is greater than the $265 000 average salary of primary care physicians in the US.⁹ In contrast, writing commentaries for newspapers, magazines, and medical journals typically pays a token sum, often less than $500, or nothing at all.

I’ve done this math before for some prominent Substacks promoting COVID-19 minimization and even antivaccine views. For example, take Dr. Vinay Prasad (please). Dr. Prasad’s Substack brags about “thousands” of paid subscribers. The minimum subscription fee is $5.00 a month, but he charges $7 a month or $70 a year, which works out to $5.83 per month. Substack’s cut is 10% plus credit card processing (~3%), so let’s say that 13% of gross subscription revenue goes to Substack and its credit card processing service, the rest going to Dr. Prasad. For illustrative purposes, let’s further guestimate that Dr. Prasad is at the low end of “thousands” of subscribers, even though I suspect he has at least the mid-range. For example, 2,000 subscribers would generate between $11,660 and $14,000 a month in gross income. Minus Substack’s total ~13% charge that works out to somewhere between $10,144 and $12,180 a month, or between $121,730 and $146,160 a year. That’d leave $8,700/month net income, or $104,400 a year, and it could be a lot more than that. For instance, if he had 9,000 subscribers, those same numbers would translate into somewhere between $547,787 and $657,720 a year. So basically, it is likely that Dr. Prasad is earning somewhere between $100,000 and $600,000 a year just from his Substack, given that his Substack says “thousands of paid subscribers,” not just thousands of subscribers, a small percentage of whom might be, as Mazer and Rose describe in their commentary, paid subscribers.

As we used to say in the 1980s, “Righteous bucks!”

None of this includes Dr. Prasad’s podcast or other online social media ventures.

Dr. Prasad is, of course, one of the upper-mid-level physician influencers on Substack. I could go on and cite many more examples in order to point out that even several hundred paid subscribers can produce a significant income stream in the tens of thousands of dollars per year. If a Substack writer can get into the tens of thousands range, then it is easily conceivable that income streams of over $1,000,000 a year are within reach. Come to think of it, maybe I should move my personal blog over to Substack. There’s just one problem. Skepticism, supporting science, and debunking pseudoscience and conspiracy theories are not the sorts of content that generally inspire readers to pay for access.

Mazer and Rose get close to stating the problem, without quite nailing it:

The ideological influence exerted through crowdfunding is less obvious than what might occur from an industry payer, but these payments are influential all the same. Consumers who choose to financially support physician-creators are not a random selection of the population or even of a doctor’s overall audience. Paid subscribers represent the most engaged fans, many of whom hold a specific set of strongly held beliefs. Substack, for instance, encourages content creators to assess their audience “in terms of a shared perspective” or “passion point”¹⁰ suggesting that successful writers “bring together like-minded people.”¹¹ The crowdfunding platform Patreon similarly recommends that creators make “the stuff your audience is craving.” If a piece of content is “getting a lot of attention online . . . this is a good sign that you’re on the right track.”¹² These are solid marketing tips but not a recipe for diversity of scientific thought. This advice implies that patrons want their beliefs reinforced, not challenged. Healthcare professionals who initially sought to change public opinion might find themselves changed in the process of acquiring an audience. Sophisticated analytical tools now allow creators to easily identify which content garners the most engagement and subscriptions. Real time feedback encourages physicians to pursue topics and angles that are most popular with their audience even in the absence of an explicit mandate. We call this phenomenon, where fans shape scientific content and conclusions, “subscription science.”

There’s a term for this phenomenon, of finding oneself “changed in the process of acquiring an audience.” It’s a term that has been discussed in the past by both Dr. Jonathan Howard and me: Audience capture. In fact, as both of us described, the money might not even be the most powerful motivating factor of audience capture, although certainly it makes a big difference. We’ve certainly seen many examples of audience capture, in which acquiring an audience first leads a formerly ostensibly science-based physician astray and then deeper and deeper into COVID-19 and antivax conspiracy theories, almost without the physician not realizing it. Again, physicians of a certain personality type seem to be unable to accept that they are prone to human foibles with respect to social and financial influences; so those who have been captured by their audience are generally in deep denial that that is what, in fact, has happened and will react very negatively to any suggestion that they have been influenced by their audience to such an extent.

Substack: The perfect platform to turn audience capture up to 11

It’s been a while since I first discussed audience capture. For a fuller discussion, you can go back to Dr. Howard’s or my posts from last year, but it is worth briefly discussing here again audience capture, this time in the context of Substack subscriptions. I will also reiterate that audience capture is not a new phenomenon. Back in the “old” days, TV commentators and newspaper columnists could fall prey to it based on the feedback to their commentaries. It’s just that social media in general, and Substack in particular, have made avoiding audience capture so much more difficult, particularly if you don’t consider yourself easily influenced by base considerations of popularity, praise, and, yes, money.

To review, I once again will reference an excellent post on social media influencers and audience capture published by Gurwinder Bhogal, ironically enough, on Substack and entitled The Perils of Audience Capture, subtitled, How influencers become brainwashed by their audiences. Bhogal introduces the concept of audience capture by relating the story of a man named Nicholas Perry, who started out posting videos on YouTube about his passions, playing the violin and the virtues of veganism. His videos went largely unnoticed. Later, he abandoned veganism and discovered that posting mukbang videos of himself consuming various dishes while talking to the camera, “as if having dinner with a friend.” These videos took off, and his audience ballooned, but:

…as the audience grew, so did their demands. The comments sections of the videos soon became filled with people challenging Perry to eat as much as he physically could. Eager to please, he began to set himself torturous eating challenges, each bigger than the last. His audience applauded, but always demanded more. Soon, he was filming himself eating entire menus of fast food restaurants in one sitting.

Resulting in this:

Nikocado, moulded by his audience’s desires into a cartoonish extreme, is now a wholly different character from Nicholas Perry, the vegan violinist who first started making videos. Where Perry was mild-mannered and health conscious, Nikocado is loud, abrasive, and spectacularly grotesque. Where Perry was a picky eater, Nikocado devoured everything he could, including finally Perry himself. The rampant appetite for attention caused the person to be subsumed by the persona.

As much as I agree with Mazer and Rose, this is the key area where I found their discussion lacking. They mention audience capture, albeit without really naming it explicitly, In fairness, I realize that they had a low word count and that their purpose was to emphasize that there are financial COIs every bit as important as pharma and device manufacturer COIs that are seldom considered—but should be. Consider my purpose here to be to add the important observation that it’s more complicated than that, even as I acknowledge that Substack appears to be an almost perfect instrument to combine audience capture with perverse financial incentives to become more dramatic and “contrarian” and thereby produce legions of physician-influencers who undermine public health and science while considering themselves “brave maverics.”

But back to audience capture. Here’s how it’s more complicated:

Audience capture is an irresistible force in the world of influencing, because it’s not just a conscious process but also an unconscious one. While it may ostensibly appear to be a simple case of influencers making a business decision to create more of the content they believe audiences want, and then being incentivized by engagement numbers to remain in this niche forever, it’s actually deeper than that. It involves the gradual and unwitting replacement of a person’s identity with one custom-made for the audience. To understand how, we must consider how people come to define themselves. A person’s identity is being constantly refined, so it needs constant feedback. That feedback typically comes from other people, not so much by what they say they see as by what we think they see. We develop our personalities by imagining ourselves through others’ eyes, using their borrowed gazes like mirrors to dress ourselves.

Over time, even without the financial incentive, audience capture can happen thusly:

When influencers are analyzing audience feedback, they often find that their more outlandish behavior receives the most attention and approval, which leads them to recalibrate their personalities according to far more extreme social cues than those they’d receive in real life. In doing this they exaggerate the more idiosyncratic facets of their personalities, becoming crude caricatures of themselves. The caricature quickly becomes the influencer’s distinct brand, and all subsequent attempts by the influencer to remain on-brand and fulfill audience expectations require them to act like the caricature. As the caricature becomes more familiar than the person, both to the audience and to the influencer, it comes to be regarded by both as the only honest expression of the influencer, so that any deviation from it soon looks and feels inauthentic. At that point the persona has eclipsed the person, and the audience has captured the influencer.

Dr. Howard noted:

Their fans often lavish them with praise, telling them they are “fearless” for challenging the orthodoxy. And you can see how some doctors get cornered by their followers and are forced to adopt more extreme positions. The cardinal rule is to never stop being “heterodox”, no matter what. This is how doctors end up praising viruses. Young doctors take heed. Be careful not to develop a brand and cultivate followers you must continually impress – instead of leading your followers, you might become their follower.

Now add a reliable income stream to the mix. Bucking the “brand” can lead subscribers to cancel their subscriptions, while more extreme writings can attract more subscribers and thus more income. It’s possible to run afoul of this dynamic even as you think you’re remaining on brand and “reasonable,” as Dr. Prasad did recently, when he “fact checked” the health claims in an interview with antivax extremist leader Robert F. Kennedy, Jr., concluding, hilariously:

RFK Jr was correct and incorrect on some matters. In several places his argument could be refined and made stronger. All In is a great podcast. The hosts have real courage. I encourage everyone to listen.

What matters did Dr. Prasad consider RFK Jr. to have been “incorrect” about? Two things, mainly. Dr. Prasad thinks that RFK Jr. is wrong to promote ivermectin and hydroxychloroquine as cure-alls for COVID-19. (Wow, something that Dr. Prasad and I agree on. Pinch me to make sure I’m awake!) The second is RFK Jr.’s claim that the expansion of the vaccine schedule in the 1990s is responsible for a wave of chronic disease, neurologic disease, and autoimmunity, a claim that he finds implausible and without supporting evidence. (Pinch me again!) Unfortunately, apparently feeling the need to placate his audience, Dr. Prasad undermines his takedown of old antivaccine misinformation by agreeing with RFK Jr. about the dangers of myocarditis from COVID-19 vaccines. He just doesn’t go as far as RFK Jr. does.

Amusingly and demonstrating the perils of audience capture from the most extreme fans, let’s just say that many of Dr. Prasad’s fans were…not pleased. Even though Dr. Prasad agreed with RFK Jr. on regulatory capture, remdesivir, “censorship,” and almost everything else in the interview, he was the subject of a video by Jimmy Dore featuring Dr. Jay Bhattacharya, with Dr. Paul Alexander saying that he likes “Prasad, smart, I do not agree with all he says here,” adding:

I think that Vinay needs to step back and not summarily dismiss Kennedy. Kennedy is standing up while others have their head up their asses. Kennedy is getting his teeth kicked in but standing up, huge praise.

In the comments, Dr. Prasad was referred dismissively to as a “mainstream straddler,” which is actually not a bad description of him from a science-based standpoint. At least, it wasn’t in the beginning before Dr. Prasad was nearly totally captured by his audience. He wanted to be seen as scientifically rigorous while also being a “maverick” who “questioned” the consensus as a True Scientist. Meanwhile, the comment section of Dr. Prasad’s post itself is full of people unhappy about his take on ivermectin and hydroxychloroquine. He even garnered the notice of RFK Jr. himself, who showed up in the comments to push back on Dr. Prasad’s (correct) assessment that ivermectin and hydroxychloroquine do not work against COVID-19. I note that RFK Jr. has never shown up in the comments here or at my not-so-super-secret other blog, which tells me that he views Dr. Prasad as promoting a similar message as his and wasn’t happy to be “betrayed.”

Others objected to Dr. Prasad’s (correct) dismissal of RFK Jr.’s equating correlation with causation with respect to the vaccine schedule and neurodevelopmental disorders. While it’s true that there are occasional comments calling Dr. Prasad to task for not objecting overmuch to RFK Jr.’s statements regarding how most vaccines are detrimental, which is a perfectly legitimate thing to call Dr. Prasad out for, most comments ranged from praising him for being willing to “engage” with RFK Jr. to taking him to task for having pointed out two areas where RFK Jr. is unequivocally wrong.

In any event, the pushback for going against the brand you’ve created is not always nasty or violent. Sometimes it’s just a shift in tone, or a change in the ratio of positive versus negative comments, or even something like this:

But the alleged fact check left much to be desired, and represents a disappointing attack from a COVID narrative skeptic against another skeptic when a conversation between the two would have no doubt been much more productive.

Go and read Dr. Prasad’s “fact check” for yourself. It was anything but an “attack.”

Mazer and Rose are correct to note:

The political polarisation of the covid-19 pandemic has fuelled subscription science and its resulting conflicts of interest. Many physicians and scientists who were sincerely sceptical or supportive of public health measures, for example, have acquired large, paying audiences over the last three years. Members of the public who were outraged by perceived government over-reach or apathy sought out professionals who would bolster their political views with scientific justification. Physicians have been encouraged by their devoted admirers to draw assertive conclusions in lieu of exploring epistemic uncertainty. Subscription science can lead to doctors promoting anti-vaccine views at one extreme or fear mongering about SARS-CoV-2 at the other.

I will, however, note that there is a rather obvious bit of false equivalency here. Whether it was demanded by the BMJ editors, given how bad BMJ has become with respect to science, or part of the original manuscript, I do not know. However, I am hard-pressed to find nearly as many examples of COVID-19 fear mongerers compared to COVID-19 minimizers and antivaxxers, in particular ones with large paid audiences. Reading that passage made me wonder just whom they had in mind and what they considered COVID-19 fear mongering. That being said, I won’t argue that one might be pushed into more extreme stances in terms of public health interventions through audience capture. I’m just hard pressed to think of an example, apart from perhaps Eric Feigl-Ding. This is an assymetrical problem right now; it is the deniers and antivaxxers who are a much larger problem.

What is the solution?

Mazer and Rose note that physicians “do not transform into perfectly objective and politically neutral beings simply by writing for the mainstream media or scientific literature” but point out that such traditional outlets “offer useful curbs on ideological bias,” such as editors and fact checkers, while journals perform peer review, which, for all its shortcomings, does tend to tamp down the most extreme statements based on studies. To them, correctly, crowdfunded media is “popular partly because of its absence of editorial controls, which enhances the content’s intimacy and immediacy.” To this, I would add that all social media, even social media in which a physician receives no payment, can result in audience capture, as can legacy media. I like to point to Dr. Drew Pinsky as an example of a physician who has gone from just a bit dubious to promoting the worst of the worst in terms of COVID-19 misinformation.

Which brings us to this suggestion:

Physicians and scientists have a right to publicly disseminate their views, even in exchange for money. Yet the legitimacy conferred by professional credentials should come with some responsibility for accuracy and fairness. Publicly disclosing crowdfunding revenue should be the minimum required. Journal editors and professional societies should also take crowdfunding payments into account when choosing the authors of editorials and clinical guidelines. Most importantly, doctors who wish to remain trusted arbiters of medical science, in all its complexity and ambiguity, should seek to avoid any financial scheme that explicitly or implicitly circumscribes the conclusions they reach.

As for Mazer and Rose’s latter suggestion, I fear that that ship has sailed. As for the former conclusion, I agree that the best that we can hope for now is to treat income from social media and crowdfunded publishers like Substack the same way that we treat pharmaceutical and device manufacturer income: Disclose it on every paper, presentation at professional society meetings, or talk for anyone.

The bottom line is that COIs do matter. Because science is a human endeavor, it will never be perfectly pristine, because nothing humans do is perfectly pristine. Moreover, science- and evidence-based medicine hasn’t always understood or handled COIs, both disclosed and undisclosed, real or perceived, very well, to the point where new regulations by the government may well be necessary. Against that background, it was correct of professional societies and journals to increasingly emphasize disclosure of COIs from industry. What we as a profession have done less well is to deal with non-industry financial COIs and the much more difficult to quantify COIs based on physician-influencer brand. I would argue that we should deal with those as well, but first let’s deal with social media and publishing income, which has the advantage of at least being quantifiable.