By David Tuller, DrPH
When I first investigated the PACE trail in 2015, one of the features most shocking to me was the investigators’ blatant violation of a core human rights document that they had promised in their protocol to adhere to. That violation of the Declaration of Helsinki involved their failure to tell study participants about their close ties with the insurance industry. For years, the lead investigators had advised disability insurers that the interventions being tested in PACE for efficacy were in fact effective and could get claimants back to work.
For reasons that I don’t quite understand, this breach of the Declaration of Helsinki has received relatively little attention in the PACE debate–perhaps because PACE has so many disastrous methodological problems that “mere” ethical ones have taken a back seat. However, Professor MIchael Sharpe has recently performed an act of public service–through his decision to blame Guardian columnist George Monbiot for causing Long COVID, he has redirected attention back to the PACE trial. So I thought it would be useful to repost my first piece about these insurance industry ties.
This topic is also relevant now, since Berkeley is once again crowdfunding to support my project. Professor Sharpe in particular has whined about this, charging that I am being paid to trash his work. As I have explained, I worked on this project for 2-3 years with no support, so that proves I was happy to trash his work for free. At this point, my project is about investigating problematic research in this field, and that’s what I’m doing. My crowdfunding is public and sponsored by Berkeley. Everyone knows about it–unlike the naive PACE participants, who were not informed of Professor Sharpe’s cozy relationship with UNUMProvident, for example.
Below is an edited version of the piece that ran on November 17, 2015.
In my initial story on Virology Blog, I charged the PACE investigators with violating the Declaration of Helsinki, developed in the 1950s by the World Medical Association to protect human research subjects. The declaration mandates that scientists disclose â€œinstitutional affiliationsâ€ and â€œany possible conflicts of interestâ€ to prospective trial participants as part of the process of obtaining informed consent.
The investigators promised in their protocol to adhere to this foundational human rights document, among other ethical codes. Despite this promise, they did not tell prospective participants about their financial and consulting links with insurance companies, including those in the disability sector. That ethical breach raises serious concerns about whether the â€œinformed consentâ€ they obtained from all 641 of their trial participants was truly â€œinformed,â€ and therefore legitimate.
The PACE investigators do not agree that the lack of disclosure is an ethical breach. In their response to my Virology Blog story, they did not even mention the Declaration of Helsinki or explain why they violated it in seeking informed consent. Instead, they defended their actions by noting that they had disclosed their financial and consulting links in the published articles, and had informed participants about who funded the researchâ€“responses that did not address the central concern.
â€œI find their statement that they disclosed to The Lancet but not to potential subjects bemusing,â€ said Jon Merz, a professor of medical ethics at the University of Pennsylvania. â€œThe issue is coming clean to all who would rely on their objectivity and fairness in conducting their science. Disclosure is the least we require of scientists, as it puts those who should be able to trust them on notice that they may be serving two masters.â€
In their Virology Blog response, the PACE team also stated that no insurance companies were involved in the research, that only three of the 19 investigators â€œhave done consultancy work at various times for insurance companies,â€ and that this work â€œwas not related to the research.â€ The first statement was true, but direct involvement in a study is of course only one possible form of conflict of interest. The second statement was false. According to the PACE teamâ€™s conflict of interest disclosures in The Lancet, the actual number of researchers with insurance industry ties was fourâ€”along with the three principal investigators, physiotherapist Jessica Bavington acknowledged such links.
But here, Iâ€™ll focus on the third claimâ€“that their consulting work â€œwas not related to the research.â€ In particular, Iâ€™ll examine an online article posted by Swiss Re, a large reinsurance company. The article describes a â€œweb-based discussion groupâ€ held with Peter White, the lead PACE investigator, and reveals some of the claims-assessing recommendations arising from that presentation. White included consulting work with Swiss Re in his Lancet disclosure.
The Lancet published the PACE results in February, 2011; the undated Swiss Re article was published sometime within the following year or so. The headline: â€œManaging claims for chronic fatigue the active way.â€ (Note that this headline uses â€œchronic fatigueâ€ rather than â€œchronic fatigue syndrome,â€ although chronic fatigue is a symptom common to many illnesses and is quite distinct from the disease known as chronic fatigue syndrome. Understanding the difference between the two would likely be helpful in making decisions about insurance claims.)
The Swiss Re article noted that the illness â€œcan be an emotive subjectâ€ and then focused on the implications of the PACE study for assessing insurance claims. It started with a summary account of the findings from the study, reporting that the â€œactive rehabilitationâ€ arms of cognitive behavioral therapy and graded exercise therapy â€œresulted in greater reduction of patientsâ€™ fatigue and larger improvement in physical functioningâ€ than either adaptive pacing therapy or specialist medical care, the baseline condition. (The three intervention arms also received specialist medical care.)
The trialâ€™s â€œkey message,â€ declared the article, was that â€œpushing the limits in a therapeutic setting using well described treatment modalities is more effective in alleviating fatigue and dysfunction than staying within the limits imposed by the illness traditionally advocated by â€˜pacing.â€™â€
Added the article: â€œIf a CFS patient does not gradually increase their activity, supported by an appropriate therapist, then their recovery will be slower. This seems a simple message but it is an important one as many believe that â€˜pacingâ€™ is the most beneficial treatment.â€
This understanding of the PACE researchâ€”presumably based on information from Peter Whiteâ€™s web-based discussionâ€”was wrong. Pacing is not and has never been a â€œtreatment.â€ It is also not one of the â€œfour most commonly used therapies,â€ as the newsletter article declared, since it has never been a â€œtherapyâ€ either. It is a self-help method practiced by many patients seeking the best way to manage their limited energy reserves.
The PACE investigators did not test pacing. Instead, the intervention they dubbed â€œadaptive pacing therapyâ€ was an operationalized version of â€œpacingâ€ developed specifically for the study. Many patients objected to the trialâ€™s form of pacing as overly prescriptive, demanding and unlike the version they practiced on their own. Transforming an intuitive, self-directed approach into a â€œtreatmentâ€ administered by a â€œtherapistâ€ was not a true test of whether the self-help approach is effective, they arguedâ€“with significant justification. Yet the Swiss Re article presented â€œadaptive pacing therapyâ€ as if it were identical to â€œpacing.â€
The Swiss Re article did not mention that the reported improvements from â€œactive rehabilitationâ€ were based on subjective outcomes and were not supported by the studyâ€™s objective data. Nor did it report any of the major flaws of the PACE study or offer any reasons to doubt the integrity of the findings.
The article next asked, â€œWhat can insurers and reinsurers do to assist the recovery and return to work of CFS claimants?â€ It then described the conclusions to be drawn from the discussion with White about the PACE trialâ€”the â€œkey takeaways for claims management.â€
First, Swiss Re advised its employees, question the diagnosis, because â€œmisdiagnosis is not uncommon.â€
The second point was this: â€œIt is likely that input will be required to change a claimantâ€™s beliefs about his or her condition and the effectiveness of active rehabilitationâ€¦Funding for these CFS treatments is not expensive (in the UK, around Â£2,000) so insurers may well want to consider funding this for the right claimants.â€
Translation: Patients who believe they have a medical disease are wrong, and they need to be persuaded that they are wrong and that they can get better with therapy. Insurers can avoid large payouts by covering the minimal costs of these treatments for patients vulnerable to such persuasion, given the right â€œinput.â€
Finally, the article warned that private therapists might not provide the kinds of â€œinputâ€ required to convince patients they were wrong. Instead of appropriately â€œactiveâ€ approaches like cognitive behavior therapy and graded exercise therapy, these therapists might instead pursue treatments that could reinforce claimantsâ€™ misguided beliefs about being seriously ill, the article suggested.
â€œCheck that private practitioners are delivering active rehabilitation therapies, such as those described in this article, as opposed to sick role adaptation,â€ the Swiss RE article advised. (The PACE investigators, drawing on the concept known as â€œthe sick roleâ€ in medical sociology, have long expressed concern that advocacy groups enabled patientsâ€™ condition by bolstering their conviction that they suffered from a â€œmedical disease,â€ as Michael Sharpe, another key PACE investigator, noted in a 2002 UNUMProvident report. This conviction encouraged patients to demand social benefits and health care resources rather than focus on improving through therapy, Sharpe wrote.)
Lastly, the Swiss Re article addressed â€œa final point specific to claims assessment.â€ A diagnosis of chronic fatigue syndrome, stated the article, provided an opportunity in some cases to apply a mental health exclusion, depending upon the wording of the policy. In contrast, a diagnosis of myalgic encephalomyelitis did not.
Presumably, after White disseminated the good news about the PACE results at the web-based discussion, Swiss Reâ€™s claims managers felt better equipped to help ME/CFS claimants. And presumably that help included coverage for cognitive behavior therapy and graded exercise therapy so that claimants could receive the critical â€œinputâ€ they needed in order to recognize and accept that they didnâ€™t have a medical disease after all.
In sum, contrary to the investigatorsâ€™ argument in their response to Virology Blog, the PACE research and findings appear to be very much â€œrelated toâ€ insurance industry consulting work. The claim that these relationships did not represent â€œpossible conflicts of interestâ€ and â€œinstitutional affiliationsâ€ requiring disclosure under the Declaration of Helsinki cannot be taken seriously.
Update 11/17/15 12:22 PM: I should have mentioned in the story that, in the PACE trial, participants in the cognitive behavior therapy and graded exercise therapy arms were no more likely to have increased their hours of employment than those in the other arms. In other words, there was no evidence for the claims presented in the Swiss Re article, based on Peter Whiteâ€™s presentation, that these treatments were any more effective in getting people back to work.
The PACE investigators published this employment data in a 2012 paper in PLoS One. It is unclear whether Peter White already knew these results at the time of his Swiss Re presentation on the PACE results.
Update 11/18/15 6:54 AM: I also forgot to mention in the story that the three principal PACE investigators did not respond to an e-mail seeking comment about their insurance industry work. Lancet editor Richard Horton also did not respond to an e-mail seeking comment.