The End of Placeboes Is in Sight

 

Matthew Chapman

 

Five years ago Alison Stell woke up at her friend’s house at Christmas time in excruciating pain and crawled to the bathroom. Unfortunately for 34-year-old Stell it was only the start of her ordeal. She ended up in hospital for three months as doctors attempted to diagnose and treat her.

 

Since then no treatments other than pain killers have helped improve what doctors now believe is chronic irritable bowel syndrome. “I have had biopsies taken, I’ve got ulcerations in my stomach right now, I’ve had infections, I’ve been in and out of hospital constantly for the last five years, and I am a lot of pain killers,” says Stell. “I’m very desperate.”

 

It is that desperation that led Stell to sign up to take part in a clinical trial involving placebos to test the safety and efficacy of IBS drug Blautix. Because the trial is double-blind, neither the trial participants nor those running the trial know which group of people is being given a placebo and which is receiving the experimental treatment. Stell – who struggles to sleep and can no longer work because of her condition – may never find out if she was in the placebo group.

 

Being placed in the placebo group is a source of great fear for Alison, but she feels like she has no other option. “They know people are desperate to start with and willing to take the risk of being in the placebo group,” says Alison. “It is unfair, I don’t see the point in placebo groups, and why you would put people through them and why you would waste people’s time.”

 

But now a new scientific movement is planning to do away with placebos altogether. If successful it would redefine the long-held ‘gold standard’ for clinical trials and could prove hugely beneficial for both patients and pharmacy companies alike.

 

Placebo groups have long been viewed as a necessary evil by the medical profession when no other option is available. The first known placebo trial dates back to 1863, when the American doctor Austin Flint compared a treatment for rheumatic fever with a placebo derived from heavily-diluted plant extract.

 

Nowadays, wherever possible, clinical trials will involve an ‘active’ control, where the new therapy is measured against the current standard of care available rather than just a placebo. This is a result of the 1964 Declaration of Helsinki, which states a new treatment should be tested against the best current therapeutic methods. However, a clarification in 2002 by the World Medical Association stated placebo-controlled trials are ethically acceptable even when a proven therapy is available, if there are compelling enough methodological reasons or when patients will not be subject to additional risk or serious or irreversible harm.

 

Most cancer trials will now avoid placebos, but in other areas of medicine such as depression – and experiments involving controlled or illegal substances – placebos remain much more commonplace. In Stell’s case, she must tell the study team about all medicines, supplements, probiotics and natural remedies she is taking because some treatments are not allowed during the course of the study.

 

“The primary purpose of a clinical trial is not to treat the individual who participates, it is to treat future patients,” says Charlotte Blease, a Keane Scholar at research initiative OpenNotes, which seeks to make healthcare more transparent. “Most patients are not debriefed after clinical trials and nobody even bothers to tell them afterwards what happened to the drug or what they were allocated, and that’s really a terrible way to treat clinical participants.”

 

Fortunately, a potential solution to the ethical quandary raised by placebo groups is on the horizon. The digitisation of health records means real world evidence such as doctors’ notes and pathology reports could be used to create a synthetic control arm to replace clinical trials’ existing control arms.

 

This means rather than having patients taking a placebo in a clinical trial, the same results can be achieved by identifying patients in the real world at the same stage of illness and within a similar demographic. The same methodology that would usually be applied to analyse a control group can be applied to real world data to generate results that can be compared against the new treatment. For instance, this could involve analysing the efficacy of the real-world treatments being received by patients with a certain stage of cancer.

 

The promise of real world evidence to create synthetic control arms was part of the rationale for pharma giant Roche buying oncology data company Flatiron Health for $1.9 billion (£1.5bn) in February 2018. Flatiron Health has been working with the US Food and Drug Administration (FDA) since 2016 to better understand how real world evidence can support control arms.

 

In April 2019 the FDA gave its first approval for a drug based purely on real world data. A collaboration between Pfizer and Flatiron Health allowed the FDA to use real world evidence from electronic health records to give the green light for the breast cancer drug Ibrance to also be used to treat men.

 

“We are at a tipping point with the ability to systematically collect patient level data in the real world,” says David Harland, medical head of personalised healthcare at Roche UK. “We're moving from paper based patient records into electronic health records. We’re now at the place where we can start to collate and curate that data to a very high regulatory grade.”

 

The concept of creating synthetic control arms is not a new one. Nigel Blackburn, director of Cancer Research UK’s Centre for Drug Development, first heard of the idea from Medidata, a company that provides software as a service for clinical trials. Medidata proposed interrogating its huge dataset of clinical trials to see if they had a sufficiently robust dataset data to construct a synthetic control arm. Using historical data to replace control arms could save pharmaceutical companies “an absolute fortune in both time and money,” Blackburn says.

 

The ultimate aim, however, is to use real-time data to use as a synthetic control arm. “If we could get the NHS electronic health records sorted out, it would be amazing what you could do with that data,” says Blackburn. Roche is already working with Moorfields Eye Hospital in the UK to create a “holistic data-record” in eye disease and oculomics.

 

Harland says the NHS is the near “perfect place” to build synthetic control arms because of the breadth of the data available. “In many respects the US is still a number of years ahead of the UK, but each individual centre essentially operates its own data systems,” says Harland. “Whereas in the UK there is more interoperability of that data.”

 

The sensitivity of health data means privacy will always be a strong concern when it comes to the prospect of opening NHS data up to private companies. Harland says patients will always be explicitly asked if they are happy for their data to be used in clinical trials.

 

“If you are being treated at a cancer centre in the UK you may well be asked to say you consent to use your healthcare data in clinical trials going forward as part of a virtual control or synthetic control arm,” he says.

 

The pace of innovation in the oncology sector has led Harland to predict that in five years’ time over half of regulatory submissions will include a synthetic control arm or real world data component. Synthetic control arms could also help eliminate the placebo effect – the psychological boost patients sometimes receive when taking an inactive treatment.

 

“What you are theoretically able to see is how your intervention compares with how patients are performing in the real world as opposed to an artificially good response that you might see in a clinical trial because of the placebo effect,” says Harland. “In the past real world data has had a reputation problem because it is not being considered robust enough, but now we are able to increase the quality through technology, the advantage of it will start to outstrip using a placebo in a clinical trial because it will be more reflective of how patients actually respond to their standard of care.”

 

Blease believes that anything that can avoid people being placed in a placebo group is a beneficial from an ethical standpoint, but warns of some potential pitfalls of synthetic control arms built on real-world data or historic clinical data.

 

She believes the adage of “garbage in, garbage out” is applicable and a great deal of care must be taken to ensure the data is robust, otherwise issues with the original dataset will be carried through to the clinical trial.

 

Researchers would also have to find a way to replicate the Hawthorne Effect: phenomenon where the subjects of a study alter their behaviour when they are aware they are being observed. In the case of a clinical trial this could involve overstating the effectiveness of a treatment in order to please the doctor.

 

“You have got to recreate the Hawthorne Effect (in the synthetic control arm) in some sense because otherwise you’re going to overestimate the drug effect,” says Blease. This would be no easy task but could involve an algorithm simulating the Hawthorne Effect within real world data.

 

The future of real world data in medicine is set to explode and Google’s acquisition of Fitbit could have major ramifications for the healthcare sector. Speaking before the Google acquisition, Harland said the future of real world evidence could include the incorporation of data from a smartwatch or Fitbit. “Realistically we are not there yet, but that is absolutely the future,” says Harland. “The amount of data that will be available will grow exponentially.”

 

The growth of real world evidence can not come fast enough for Alison Stell and others participating in clinical trials involving placebo groups. But until then, their desperation is still driving them to put themselves forward for clinical trials, never knowing if they will be the lucky ones to benefit from a revolutionary new medicine.