A randomised control trial found that sending audit letters to doctors who requested a lot of pathology tests stopped them from doing that. But was it ethical?
Some of Australia’s most prestigious primary care researchers are currently at odds over the ethics of studying compliance interventions on a population – namely, general practitioners – who never consented to be studied.
On one side, study authors argued that gathering consent from roughly 5500 de-identified GPs would have been “impracticable” and that the potential harms of the intervention were minimal.
On the other, GP researchers said that there was an inherent harm in denying study participants the right to autonomy.
The research
The inciting study, titled Nationwide audit and feedback to reduce overuse of pathology tests among high-requesting general practitioners in Australia: a factorial, cluster-randomised, controlled trial, was published in The Lancet Primary Care in April.
Its author list included evidence-based medicine proponent Professor Paul Glasziou – the former director of Bond University’s Institute for Evidence-Based Healthcare – and Monash University researcher Professor Denise O’Connor.
The paper itself looked at an “audit and feedback” intervention aimed at GPs who were in the top 10% of providers requesting pathology test combinations which were “commonly overused and unlikely to support diagnosis or patient health outcomes”.
Those in the intervention group received personalised written feedback via pamphlet or letter, signed by chief medical officer Professor Paul Kelly. Some included information about the cost of pathology testing or an invitation to CPD-accredited education activities.
The Bond University Human Research Ethics Committee approved the study and waived participant consent because the main aim of the research was to evaluate audit and feedback variations and therefore carried only “minimal” risk of discomfort or inconvenience.
All up, the letters went out to more than 5000 GPs.
Broadly, the intervention worked.
In comparison to the control group, there was a 36% reduction in the overall rate of requests among the intervention group over the following six months.
The response
The July issue of The Lancet Primary Care featured a letter from Professor Louise Stone, Dr Michael Tam, and Associate Professor Oliver Frank – all well-known GP researchers in their own right – flagging concern about the participant consent waiver.
“General practitioners were not informed that they were included in a trial and were not given the opportunity to withhold their consent,” the correspondence said.
“The researchers argue that gaining specific consent would make the study unfeasible and that government endorsement made consent unnecessary. We disagree.
“Governments may intervene in clinical practice without specific consent, but researchers should not.”
Professor Stone and colleagues wrote that they believed it was possible to disclose the nature of the study and include opt-out consent.
“Respect for autonomy, dignity, and fundamental humanity should be non-negotiable in research,” they wrote.
“We ask the team to consider clinicians’ ethical rights in future studies.”
Speaking to The Medical Republic directly, Professor Stone argued that the paper understated the potential harm of the intervention.
“Although the various programs emphasise that GPs need to ‘reflect’ on the letters and consider the clinical relevance of the advice, nudge letters use the language of influence and marketing to subconsciously change behaviour,” she said.
“There has been considerable concern about the impact of the letters on the GPs, particularly as the teams acknowledge that using the Chief Medical Officer as a signatory produces a ‘strong authority cue’.
“Given government’s role in regulation, a letter from the Chief Medical Officer is not a neutral communication.
“If a GP responds to a ‘nudge’, reduces their investigation into a patient’s condition, and is reported to AHPRA for refusing to investigate, the influence of nudge letters is unlikely to constitute sufficient defence.”
The GP researcher also said that there was “inherent harm” in GPs being denied the right to autonomy.
While the letter published in The Lancet Primary Care only focussed on consent, Professor Stone also questioned the merit of nudge letters in general.
“Interventions are only ‘low value’ if they are inappropriate for the individual patient, so the value of an intervention should be considered in relation to the population to which the individual belongs,” she said.
“While back x-rays may be low value for non-cancer back pain, they are very appropriate for potential osteoporotic fractures, so a GP with an aged care population or a GP dealing with eating disorders may well be using a high volume of back x-rays that are clinically appropriate.”
The outcomes that studies on nudge letter interventions tended to measure, Professor Stone pointed out, was the change in prescribing or requesting or referring volume, rather than a clinical outcome.
“… Researchers do comment on the ‘beneficial’ effect of the nudge letters, implying that the outcomes are desirable,” she said.
“I would argue that looking at a reduction in test ordering without considering clinical outcomes reduces the merit of the studies, because the outcomes may well be undesirable in certain contexts.”
The rebuttal
A second letter, this one co-signed by the original research term, was published alongside Professor Stone’s correspondence in July.
In it, they reiterated that the study was approved by an accredited Human Research Ethics Committee in 2021 and that the waiver was granted on the basis that the study satisfied the criteria in clause 2.3.10 of the Australian National Statement on Ethical Conduct in Human Research.
The researchers also said that they supported the national statement’s values of respect for human beings, research merit and justice.
“The study evaluated feedback variations (i.e., invitation to continuing professional development-accredited education, provision of cost information, and feedback format) that carried no more than minimal risk of discomfort or inconvenience to participants compared with usual feedback,” they wrote.
“General practitioners routinely receive peer comparison feedback from the Department [of Health, Disability and Ageing] and participants were provided only with information about their own practice relative to peers, alongside access to best-practice educational resources.
“The feedback acknowledged that the clinical reasons for requests were not captured by the data and multiple factors could influence request rates. Participants were invited to reflect on the information and make their own judgements regarding their future clinical practice.”
Professor O’Connor, the study lead author, told TMR that the study was “designed and undertaken in line with best practice, legal and regulatory requirements”.
“We apply the highest standards of ethical and research integrity to ensure the most rigorous outcomes while upholding the safety and rights of both GPs and patients,” she said.
