Reflections on the 2018 “FASHION” Trial
In Australia, between 2010 and 2013, the number of people undergoing hip arthroscopy increased by over 50% (Medicare Australia, Medicare Benefits Schedule Item Statistics 2014: http://medicarestatistics.humanservices. gov.au), and in the United States it increased 6-fold between 2006 and 2010.  However, high-level evidence of the efficacy of both surgical and nonsurgical interventions for patients with femoroacetabular impingement syndrome (FAIS) and/or labral tears is lacking.
Rigorous research designs involving placebo controls, randomization, allocation concealment, and double blinding are considered gold standards for medical interventions, but unfortunately are less common within surgical interventions than in non-invasive interventions.  Surgical procedures have many factors that could contribute to placebo responses. [3, 4] In a recent systematic review, large effects of placebo surgery and other invasive procedures, particularly in the field of pain related conditions, were observed. 
Many patients undergoing hip arthroscopy have not been informed of the absence of evidence supporting a beneficial effect of surgical procedures compared to other treatment options for FAIS.
While evidence is lacking to support the use of surgical intervention, its biological plausibility is also under question, given that completely asymptomatic individuals often have morphology typically seen in FAIS (see image below), and there is a lack of association between morphological changes of the hip, symptoms, and dysfunctions. [5, 6] And despite this, the number of hip arthroscopies has risen sharply.
So why are so many hip arthroscopies being performed?
As Morgan  describes, in the field of medicine, there has historically been the requirement for significant evidence to disprovesomething (i.e. remove an ineffective treatment from the market); but merely analogy, expert opinion or physiological narrative to adoptsomething. 
This is precisely what is wrong in the field of expensive and invasive treatments for chronic pain. Low standards for adoption, despite ample history showing the folly of this approach.
The role of industry:
Device manufacturing companies have important roles in shaping the opinions of physicians and patients. The mandate of device manufactures is to sell as much product as possible for the highest possible price.  For publicly traded companies, the Chief Executive Officers’ incomes are largely dictated by share prices, which are affected by perceived future profit streams and not by the magnitude of the benefits of their products to patients.  The influence of device manufacturing companies permeates medical education, physician thought leaders, the clinical research agenda, and practice guidelines — especially when guidelines are developed by professional societies that receive substantial funding from industry.
The device manufacturing industry has the technical expertise and financial resources to dominate the clinical research agenda and design the majority of clinical trials, including the choice of outcomes that sound important but are often not (such as the use of surrogate endpoints) . Industry ‘mega-trials’ have the power to detect very small gains. Selective reporting of trial results and ‘spin’ of marginal effects overstate the benefit of small gains.
For example, last year a poorly designed surgical trial, with no placebo control group, and a strawman conservative group, was heralded as a success.
“The FASHION Trial”
The Positive Spin:
The lead author kicked off the social media PR, followed by enthusiastic proponents of the procedure:
The positive spin put on this trial was interesting.
As it turns out, only 1 of 10 outcomes favoured arthroscopy, even with a biased trial design.
All serious adverse events related to treatment were in the arthroscopy group, including post-surgical infection.
Introduction of bias:
One of the inclusion criteria for the FASHION trial was that you were deemed appropriate for surgery by the surgeon. In other words, participants were told by the surgeon “you would benefit from surgery” but subsequently randomly allocated to physiotherapy led conservative management.
Patient’s expectation of improvement would have been significantly reduced if patients were told;
- They had a structural abnormality that is causing their pain
- They would benefit from surgery to correct the abnormality
- But were instead allocated to “exercises”
As we can see from the data, just because a person with pain has FAI morphology on imaging does not mean the imaging finding is the reason for the pain. But if the patient believesthat is the reason, psychologically, the idea that non-surgical treatment will “fix” their “problem” may seem illogical. Words, images, as well as beliefs and the overall narrative of the pain experience are significant. This is a salient bias that this trial completely ignored.
Crafty cropping of graphs:
In addition to the bias allowed to freely enter, the difference in outcome was also visually amplified by cropping the graph to make the one of the left appear as though the effect size was larger. 
Strawman control group:
The non-surgical intervention was described as “Best Conservative Care”.
However, the exercises given primarily target hip and trunk ‘stability’ with body weight or resistance bands. The full exercise database can be found here. I have shown below some examples of what was involved.
This is precisely the sort of reductionist, biomedical “stability” focused messaging that we are trying to move away from as a profession. Telling someone that they have something structurally wrong with them based on scans, and then prescribing “stability” exercises, while ignoring psychosocial factors, is not evidence based treatment. Chronic pain is a biopsychosocial condition. The authors claim, in the title of their paper, that the conservative group received “Best Conservative Care”. This is factually inaccurate.
This article by Cheryl Lee on the Pain-Ed website explains precisely why the exercises given in the FASHON trial were not only ineffective, but may potentially contribute to chronicity in susceptible individuals:
“Hold up your fist in front of you and clench it hard.
Keeping it clenched, how far you can roll the wrist forwards and backwards? How does that feel? Comfortable? Easy? Of course not. Would you be comfortable walking around all day with the muscles of the hand, wrist and forearm held in co-contraction? Of course not.
I did the prescribed exercises, even though they made things feel tighter. I got worse and was doing less, never feeling any stronger or more stable. Physiotherapy could do nothing more for me. The chronic pain clinic told me to start using a wheelchair. It was a hell that lasted 6 years, but that is another story.
I was not broken. My core was not weak. My spine was robust and strong, not unstable. I was scared to bend, squat, or move quickly, so avoided it. My breath was shallow and ragged. I was scared about the future. This ongoing fearful situation had ramped up my body’s defences and jacked up the tension in my body, splinting me more than the exercises I had been doing.
The advice I had been given, the words that had been said, the beliefs I had taken on through my own learning had turned out to be nocebos — noxious messages with a detrimental effect on my health and well-being — that kept me locked in a prison partly of my own making.
The constant co-contractions, the fear, the movement avoidance — these had been huge factors in the pain symptoms. I learned to let go of my belly and back, and create space in my body; move with fluidity and ease; breathe fully into a soft, yielding belly; slouch and relax at the same time. I learned that I could contract my muscles with control, but also release them FULLY; that I could run, jump; ride a bike with a rounded back. I learned my spine and body were strong, and that I could get better.
I’m not an isolated case. This is universal. Let’s stop peddling old-fashioned ideas and bullshit about health. Stop selling movement and exercise by telling people they are weak and unstable. Instead, let’s empower, educate and help people to make sense of their pain, shake the labels, improve their function and move fluidly and fearlessly.”
The conservative arm was setup to fail (a strawman), not only because they didn’t receive an elaborate theatrical placebo with a convincing narrative (like the surgical group), but because the treatment itself is not reflective of best care, and potentially iatrogenic.
Financial conflicts and industry influence:
As outlined in the trials declaration of interests, the lead author is personally funded by device manufactures:
Stryker is one of the top 5 leading device manufactures in the world, with a net annual revenue of $1.02 Billion. 
Multiple studies have shown that doctors who are personally funded by private industry tend to have more favourable views of products, and their subsequent practice behaviour is affected. 
By purposely designing a trial like FASHION, without a sham control group, the effectiveness of surgery is overestimated, and question marks over causality still remain.
The lead author, personally sponsored by one of the world’s largest device manufactures, personally;
- Designed the trial (knowingly introducing bias that could have been eliminated with sham control)
- Wrote the grant application
- Drafted the final report
- Was the guarantor for the study
- “Decided to submit the manuscript”
- Was personally funded (outside his submitted work) by a group set to financially benefit from a positive result from the trial.
In light of the evidence, these facts above should allow clinicians to view this study in context.
Ideally, it should be academics, free from financial conflicts of interest with industry, who design trials, apply for grants and have the final decision regarding submission of the manuscript.
Low levels of scientific rigor required for surgical interventions:
If arthroscopy were a pharmaceutical drug, the procedure would be required to undergo scrutiny of testing its safety and feasibility in phase 1 and 2 blinded placebo trials. This is obviously not the case in the field of surgery.
In other words, if hip arthroscopy for FAI was a pharmaceutical, the current standard of evidence would not permit it to market.
Additionally, once a practice has achieved market share, it is very difficult to get physicians to halt or abandon the practice, even if it is later shown to be ineffective in unbiased clinical trials. Some studies estimate that it takes up to 17 years for “medical reversal” (abandoning or scaling back on a practice after evidence says it’s inappropriate). 
The “last resort” argument:
In spite of this clear lack of reliable evidence, this is the sort of argument I hear quite frequently from proponents of the surgery:
“Some patients with hip shape of FAI are asymptomatic and certainly don’t require treatment. Some patients with FAI improve with physical therapy, and don’t need surgery. But the group that remains in pain can benefit from properly done hip arthroscopy.”
There is no logic to this. Failed conservative management is not an adequate justification for surgery that has no evidence.
In the cited example, if we wish to engage in evidence based practice, “group that remains” needs to be enrolled in a sham RCT to determine if any benefit from “properly done” hip arthroscopy actually relates to structurally altering the hip.
Its asymptomatic prevalence alone should challenge the notion that structure equates to pain. All prior negative sham RCTs for chronic pain should also question the biological plausibility of the rationale.
The history of medical science demonstrates that invasive interventions based on bioplausibility for chronic pain are more likely to fail, rather than succeed, when subjected to blinded placebo controlled randomised trials .
Now we have to wonder, how did this clinical trial get published in the topmost journal?
Stay tuned for part II – it will be released soon!