My struggle with our professional bias around exercise and knee osteoarthritis

We have a new clinical care standard for the management of knee OA (link here).  It’s something that is fully consistent with how I practice and what I teach.  It couldn’t be more bias confirming and I should not be happier.

But, I’m uncomfortable.  It advocates things I love like:

Good education, appropriate imaging, person centred care, addressing the entire person, optimism around joint resiliency, reframing beliefs about “wear and tear”, nutrition, weight loss and of course exercise for pain and function.

 That all seems good right?  Yes, but I feel sheepish advocating for exercise for pain and function without advocating for other conservative approaches that have similar clinical results.  We consistently do not hold exercise to the same rigorous standards of evidenced based practice as we do every other intervention.   And this lack of consistency is what makes me uncomfortable. For me, to hold a strong opinion here would be academically dishonest. Permit me to briefly critique the research base around exercise and make a case for how our biases might even negatively influence patient care.

Exercise is held to a very low standard compared to other interventions

I would argue that researchers and clinicians made the decision that exercise and physical activity should be a primary treatment for knee OA BEFORE the research was conducted.  When we evaluate an intervention the gold standard is hopefully a trove of RCTs.  In those RCTs we must have a control group.  We must control for the non-specific effects of our intervention.  This never or rarely happens in the exercise and knee OA world and no one seems to care.   An example of this would be the GLAD exercise and education trials (link here).  These are pragmatic studies which are not designed to test whether exercise and education is effective.  They have no control group they are unable to make that conclusion.  Just reading the abstract you can see that the authors state that the GLAD intervention led to reductions in pain and reduced medication usage. You can’t say that. Without a control group you can’t make that conclusion. We would not accept these types of trials for other interventions like acupuncture or electrotherapeutics.  In fact, many of the clinical practice guidelines will explicit discount the literature on passive modalities because of a lack of a control group.

For a more detailed read on this topic see the editorial by Dr Martin Englund “The Emperor’s New Clothes” which make this same argument. You can read the editorial here

When we do have direct comparisons between exercise and other interventions we see that exercise does not do that well.  This best seen in the recent paper by Messier 2021 (link here) where heavy resistance training did not outperform a very low load exercise program nor an attention control group.  We also saw the same thing where the GLAD program was compared against a saline injection with no difference found between groups (link here)

 Why am I sheepish?

I want to be the biggest cheerleader for exercise and knee OA.  I’m happy that this standard advocates all types of exercise and its therefore very patient centred in that it gives lots of options for physical activity.   This is a massive improvement over the original GLAD program that advocated for “proper alignment” and “neuromuscular control” exercises - puke. And people at the time cited all the uncontrolled GLAD trials as evidence for the need for that type of exercise - they never should have done that but people did. Fortunately, we’ve moved on. 

But…how can I be academically rigoruous and honest if I know the literature well?

How can I just advocate exercise when I know how poorly it has been tested?

How can I advocate a standard that omits other common interventions because we have held those interventions to a far higher standard?  Or we say that there is more research conducted on exercise therefore it should be included in the CPG when that type of reasoning is circular and self fulfilling.  We keep conducting this weak research and it keeps getting repeated in the CPGs. We will be told there is limited research for other interventions (manual therapy, passive modalities, acupuncture) because if we keep minimizing these interventions in CPGs we will naturally have fewer trials that study them. It just becomes self fulfilling. Exercise just gets accepted without rigour and other forms of care are dismissed because there is no research.

Are we really person-centred?

A foundation of person centred care is the provision of options.  Where we must be honest about the options people have and the research underpinning those options.  Clinical Care Standards and CPGs like this have simply culled the options based on a hugely biased interpretation of the literature.  Here is a quote from the standard “Passive manual therapies, such as therapeutic ultrasound and electrotherapy, do not play a significant role in the treatment of knee osteoarthritis”.10,11

 This again is based on an interpretation of the research on these passive modalities.  If you actually look at the papers investigating pain and function you will see that the passive modality group has comparable reductions in pain and function.  However, in many of those papers there won’t be a control group.  So what happens, the CPG committees will discount those papers.  In other papers, there is a sham comparator group.  And what do we find?  Both groups see reductions in pain and improvements in function.  But because there is no difference we say the result is not real.  But again, what have we done?  We’ve held these interventions to a different standard.

We see the same thing with manual therapy or acupuncture.   When you see exercise directly compared against manual therapy you will not see a difference in pain reduction or a difference between the two in function.  But, because there is so much more research (albeit weak) on exercise the guidelines will advocate for exercise saying its better supported.  Which doesn’t mean its better. Just that there is more research. Which as I mentioned early is a self fulfilling prophesy by the CPGs.

If we are really honest with our patients I think its fair to say that they have many options to help with pain and function. We can’t, at least I can’t, in all honesty say that exercise has been “proven” to be superior.  And if we can’t say that how are these guidelines saying that?  Why are they taking away options from people? How is this person-centred care?  Sure seems like gatekeeping to me.

What would a more honest take on exercise and other options be?

Hey, exercise is one of the options that might help your knee pain.  If you want to keep exercising and being active you absolutely can.  Its totally safe and not harmful for you.  It will not lead to more OA.  If you want to start an exercise program I can absolutely give you some options and guidance.  Exercise also has a tonne of health benefits.  Fall prevention, healthy aging, reduced all cause mortality, etc.  But, if you don’t want to exercise and want to try something else we have some research and lots of anecdotes where people report improvements in pain and function as well.   Things like acupuncture, physio-modalities, manual therapy, bracing or even gait retraining might help you.  But, one of the knocks on those things is that its something someone does to you.  Its very passive.  You have to come into the clinic, it tends to be more expensive and you don’t get any of the great health effects of exercise.

Going Forward - what’s next?

I don’t know. You tell me. Do you think we are consistently academic rigourous?