Did the pharmacist UTI prescribing trial ever secure ethics approval?

Is the continuation of the UTI pharmacist prescribing trial in Queensland a scandal?   

Are the adverse events doctors say they’ve dealt with – including a missed ectopic pregnancy, a missed 15cm pelvic mass and a missed ruptured ovarian cyst – sufficient for health regulators to pull the plug?

And should someone start knocking on the door of QUT, the university that not only helped design the trial but gave ethics approval to the evaluation methodology, which on the surface seems to have missed the harm caused to a significant number of women?

The heavy politics surrounding this major expansion of the pharmacist’s role in patient care, which also removes the prescriber-dispenser divide, has resulted in the shutters coming down.

Transparency and openness – those elements that are meant to protect medicine from going bad – have been lost in a bureaucratic darkness where faceless officials fail to answer the questions asked, usually by citing confidentiality agreements or by replying to irrelevant questions that have not been put to them.  Or they simply don’t bother replying at all.

The evaluation of the trial – the one that, according to the Queensland Health minister Yvette D’Ath, shows it is a success – is still locked in a drawer somewhere inside Queensland Health.  

No-one has been told why.

Two months ago, Australian Doctor submitted a Freedom of Information request to Queensland Health for the evaluation report.

We received a message the day it was due to be handed over to us saying an unnamed “third party” had lodged an objection to its release, and that officials needed another 15 days to assess the rights and wrongs of this demand.

The evidence that something has gone wrong lies in AMA Queensland’s survey of doctors across the state, in which some 184 doctors said they had treated women for some 240 complications following a diagnosis by a pharmacist working under the trial.

The survey concluded that cancerous conditions had been overlooked on at least nine occasions, with patients being treated for a UTI when the symptoms related to cancer or precancerous conditions, including bladder, gut, cervical and vulval cancers.

We don’t know for sure if the QUT evaluation failed to pick these up.  

An article in The Australian on a leaked draft of the report said the evaluation’s findings were ‘overwhelmingly supportive’ of the trial.

Yet, more than 4300 of the 6750 women treated were lost to follow-up.  

Four women said they had ended up in ED, but following assessment by the trial’s steering committee – the committee whose members include the Pharmacy Guild and Queensland Health – the pharmacists were found to have followed treatment protocol.

But from what we know, the trial’s design was always going to struggle to fully capture adverse events or poor care.  

No urine tests were used to establish whether any of the 6300 diagnoses made by pharmacists were correct.

And there was no control group of usual care (ie, through GPs) to compare and contrast.

The mechanism used to capture adverse events was the treating pharmacist making a “reasonable attempt” seven days after dispensing antibiotics to contact the patient for feedback on what had happened and whether they were happy with the outcome.

If care had gone wrong – for whatever reason – the patient would give the details to the treating pharmacist, who was then expected to feed back the details of the adverse events to researchers.  

The conflicts are obvious.

The absence of control groups and a confirmed diagnosis also meant there was no easy way to assess if pharmacists were fuelling antimicrobial resistance in the community, specifically if the financial benefit of being a dispenser was having an undue influence on their treatment decisions.

QUT’s human research ethics committee signed off on this approach.

Why?

In April, AMA Queensland wrote to the committee’s chair, Dr Conor Brophy, asking her to explain.

Her response published here revealed that her committee had never given ethics approval for the trial itself.  

The committee’s role was limited to giving ethics approval to the QUT evaluation.  

Dr Brophy did not go into whether she felt the evaluation methodologies were likely to miss adverse events.  

She did not say much beyond the fact that pharmacists were subject to additional training – although not spelt out, this was a three-hour online module with a 20-question multiple choice exam at the end of it.

After outlining the key features of the evaluation, Dr Brophy admitted the AMA’s survey results were “concerning”.

“Any information on outcomes and the survey findings seem very relevant to a decision on the continuation of the clinical service, and I urge the AMA to share them with Queensland Health,” she said.

After QUT engaged in some initial stonewalling, the executive dean of the Faculty of Health, Distinguished Professor Patsy Yates, did speak to Australian Doctor about the university’s role in the pilot.

Why wasn’t the evaluation structured to identify potential harms to women?

“This particular project is a service evaluation. It’s not a clinical trial,” Professor Yates said. ”So again, it’s evaluated appropriately as a service evaluation through the standard Human Research Ethics Committee processes…

“A clinical trial, by definition, is really looking at new treatments or interventions. So, if something is a new treatment or intervention, it would be typically evaluated through a clinical trial process.”

But this was a new intervention involving Australian pharmacists diagnosing uncomplicated UTIs and prescribing S4 antibiotics for the first time?

No, she said. A clinical trial was unnecessary. The protocols under which the pharmacists were working had already been “proven” safe and effective.

“[The] intervention has previously been evaluated and trialed, and there’s published evidence of proven methods … There’s evidence that’s been published overseas, and it’s [been shown to be] an effective method.”

So, the QUT ethics committee didn’t have to worry about the safety of patients when it gave its approval for the evaluation?

“In this case, the evaluation was really about the experience of service provision by pharmacists and patients,” Professor Yates said.   

But the evaluation couldn’t determine whether a patient’s diagnosis by a pharmacist was even accurate?  

“Well, [the pharmacists] are following standard protocols, approved protocols … They’re following the protocols that were approved, and they’re proven. They’d been proven.”

We asked her for a copy of the QUT ethics committee approval for QUT’s evaluation – maybe it did examine in detail the safety risks to patients.

Is it publicly available?

“No, it’s not typical that those documents [would be publicly available],” said Professor Yates.

“Probably they would be if it was a clinical trial – you would register the trial on the trial’s registry – but otherwise, no, it’s not…

“And I think, again, you’re sort of referring to information that is really bound by that agreement [with Queensland Health] in terms of what … documents can be released.”

So, the question shifts to those “proven” protocols – the evidence showing pharmacist UTI diagnosing and prescribing is safe.

Professor Yates could not provide references to the overseas literature on the safety and efficacy of pharmacists prescribing antibiotics for uncomplicated UTI.

Therefore, after a Pubmed search came up with one result, we contacted three experts on the subject.

The first was Professor Ross Tsuyuki, chair of the department of pharmacology at the University of Alberta, Canada who appeared at this year’s Pharmacy Guild of Australia annual conference on the Gold Coast to talk about the future of pharmacy prescribing.

The other two were Dr Natalie Gauld (PhD), from the University of Auckland, whose research focuses on antibiotic prescribing by pharmacists in New Zealand; and Professor Lisa Nissen, head of the school of clinical sciences at QUT, who helped to design the Queensland trial and authored the evaluation.

Professor Nissen was unable to help.

But we were sent a number of papers by the other academics. There were just two studies that seemed to deal with the efficacy and safety of patients treated.  

It didn’t seem much.

The first was published in 2013 in the British Journal of General Practice. It involved pharmacists in Glasgow and Clyde in Scotland prescribing trimethoprim to patients with symptoms suggestive of moderate-to-severe UTI.

The number of women treated by a pharmacist turned out to be just 56, of whom 47 were prescribed antibiotics. Again, no urine test to confirm diagnosis. Again, follow-up collected via patient surveys and interviews.

The study, which at the time was the first to look at pharmacists prescribing antibiotics for UTI, did not attempt to track if there were any adverse events.

Among the limitations cited was this: “A small sample size and limited clinical information is recognised.”

“Despite participating pharmacists being aware of the need to supply the antibiotic with caution, adherence to the [prescribing] criteria required improvement,” it concluded.  

“Further research is required to investigate the underlying causes of pharmacist behaviours when empowered in an antibiotic stewardship role.”

The second study took place in Canada.

According to the resulting paper published in the Canadian Pharmacists Journal in 2018, it was designed to evaluate the “effectiveness and safety of, and patient satisfaction with, pharmacist prescribing and care in patients with uncomplicated UTI”.

Australian Doctor has written about this research before.

The diagnosing and prescribing model used by the researchers was very similar to the one now being used in the Queensland pilot.  

The study wasn’t blinded or randomised, but it did have two trial arms.

The first comprised 94 patients who presented to the pharmacy with a new prescription for the treatment of UTI from another healthcare provider (ie, a doctor).

The second comprised 656 patients who presented to the pharmacy without a prescription with symptoms suggestive of a UTI.

Like in Queensland, pharmacists made their diagnoses based on patients’ subjective reports. There was no urinalysis for pyuria or a urine culture, nor was there an attempt to identify whether the reported symptoms were the result of an STI.

The treatment options were: prescribe antibacterial therapy; modified antibacterial therapy; education only; or referral to a doctor.

Again, as happened in Queensland, pharmacists conducted follow-ups to find out if the symptoms resolved, if women adhered to the therapy, and if there were any adverse events.

But there seemed to be the same potential conflict of interest issues of hoping that outcomes were reported to the treating practitioners, with no independent mechanism to track adverse events.

Just to be clear, in the pharmacy arm, “clinical cure” was achieved in 88.9% of the patients – very similar to the rate in the physician arm (91%).

Some 64 patients were lost to follow-up.  

But adverse events were reported by 7.2% of patients, with no major difference between the two arms of the trial.

Most were gastrointestinal-related (59% of those reported) and transient, followed by secondary vaginal infections (15% of those reported).

A total of five reported adverse events resulted in a physician or ED visit: three in the pharmacist arm of the trial and two in the physician arm.

“Our findings demonstrate that the management of uncomplicated UTI by pharmacists is effective and safe, and that patients express a high level of satisfaction with this care,” the authors concluded.

Professor Tsuyuki is one of the co-authors.  

In response to our request for references to studies, he said he could not find very many trials on the management of uncomplicated UTIs by community pharmacists – pointing to a 2018 review which found five studies in total, excluding his Canadian study.  

One was the Glasgow study. The other four don’t deal with clinical outcomes in patients and rates of misdiagnosis and adverse events.  

One is about prescription rates of antibiotics by pharmacists, the others more focused on surveys of patients and pharmacists about whether pharmacist prescribing is a good thing.

One study did interview GPs about a local pharmacist UTI prescribing initiative in Scotland called Pharmacy First which found 68% thought it a useful service because it lowered pressure on GP appointments. 

However, in response to AusDoc’s scepticism about the scale and quality of evidence, Professor Tsuyuki said this: “I would like to point out that there is no evidence for physician treatment of UTIs.

“As you know, modern medicine is based upon the principle of evidence-based medicine.  

“One should have evidence for the effectiveness of interventions, and no-one is exempt from that principle.”

If Professor Tsuyuki’s study, despite its limitations, is indeed the only real published evidence on the safety of pharmacist antibiotic prescribing for uncomplicated UTIs, many doctors would argue it does not provide sufficient justification to roll out the Queensland trial without more robust measures to track clinical outcomes for women and identify any harms that may result.

Should QUT’s ethics committee have approved the QUT evaluation?  

Perhaps one argument in Professor Yates’ favour when declaring the protocols have been proven, is New Zealand, where pharmacists with specific training have been dispensing trimethoprim without a prescription to women with symptoms suggestive of uncomplicated UTIs since 2012.

Again, it appears there is little high quality research, if any, on the accuracy of diagnosis and adverse events. But it is real-world patient care and it has survived long-term [1].

The problem goes back to transparency. The rights and wrongs of the QUT ethics committee’s decision cannot be judged without its report being made public.

In Australia, there are some 220 human research ethics committees recognised by the NHMRC and which have to operate according to the NHMRC’s guidelines.

Those guidelines run to 116 pages and cover issues of consent, genomic research and conflicts of interest.  

They don’t go into much detail about what should be classed as a clinical trial or a service pilot, which seems to be the preferred term for the UTI prescribing trial.

A clinical trial is defined by the NHMRC as a “form of research designed to find out the effects of an intervention, including a treatment or diagnostic procedure”.

It adds: “[The] distinction between research and innovative clinical practice is unclear.  

“For example, innovative clinical practice occurs on a spectrum from minor changes at the border of established practice that pose little change in risk to patient safety, to novel interventions that should only be introduced as part of an ethically approved research protocol.”

But the point is that even if human research is not defined as a clinical trial, the NHMRC guidelines make clear there is still a duty on ethics committees not to approve human research that puts participants at unacceptable risk.

“When risks have been identified, gauged and minimised, and the research has been approved, the risks must then be managed,” it says.

“This requires that researchers include, in their research design, mechanisms to deal adequately with any harms that occur; and a monitoring process is in place and carried out.

“The greater the risk to participants in any research for which ethical approval is given, the more certain it must be both that the risks will be managed as well as possible, and that the participants clearly understand the risks they are assuming.”

Through the interview with Australian Doctor, Professor Yates stressed that many of our questions could only be answered by Queensland Health, which is ultimately responsible for the Queensland trial and its evaluation.

Eleven days ago, in an attempt to secure enlightenment, Australian Doctor submitted once again a long list of questions to Queensland Health’s media unit.  

They included perhaps the most important question  of all: whether the trial itself, not just the QUT evaluation, had ethics approval, and if so, who provided it if it wasn’t QUT.

We also requested a copy of the ethics approval given by QUT to the evaluation.

And we asked the basic question of whether the potential risks to women in the trial were ever assessed, given pharmacists in Australia had never diagnosed, prescribed and dispensed S4 medications for the treatment of UTI.

We also requested the names and roles of the steering committee that put together the pilot.

We are still waiting for a response.

The NHMRC said complaints about ethics approval should be made to the relevant institution.

“NHMRC has no remit to receive complaints that relate to decisions made by individual Human Research Ethics Committees (HRECs).

“HRECs are the responsibility of the institution that establishes them.”

Professor Yates defends QUT’s involvement.  

She can’t say how much QUT was paid for its work. That’s commercial in confidence.

Does she feel comfortable with what’s been done? Does she think this has been the best approach taken to evaluate this intervention?

“Yes. Based upon the existing evidence about this approach to managing UTI, based upon the fact that there was training [for pharmacists] and an oversight committee and the fact that there was follow up, I think that the evaluation of this pilot was appropriate.”

[1] In reference to pharmacists in NZ prescribing trimethoprim, Dr Gauld told Australian Doctor the following: “Anecdotal feedback has been that there is a huge relief from women at the ability to access a UTI treatment in this way, but also a reasonable number of women being referred [by pharmacists] to the doctor without treatment, as is appropriate.”

She added: “There needs to be review of the appropriateness of reclassifications [of trimethoprim as a non prescription medication] over time to ensure they are working well and to reflect changes in management such as a move towards nitrofurantoin in certain circumstances given resistance with trimethoprim.”