You reviewed highlights of recent uncertainty in the evidence on whether exercise meaningfully reduces pain and improves function in knee OA. What implications does this have for future research and clinical practice?

Recent high-quality systematic reviews and randomised controlled trials (RCTs) highlight greater uncertainty about whether exercise meaningfully improves knee OA outcomes. 

Exercise therapy probably improves pain and physical function; however, whether these benefits are clinically meaningful remains unclear based on established minimal clinically important differences. 

This evidence base has several limitations: many systematic review findings are informed by low-quality RCTs, designing a placebo exercise intervention remains challenging and the minimal clinically important differences used to interpret findings vary across research. 

Future studies should evaluate what outcome changes are meaningful from the perspective of people living with knee OA.

Future research should focus on a broader range of health outcomes that matter to people with knee OA (eg, quality of life) and identity the types of exercise patients enjoy or prefer. 

Despite the uncertainty we highlighted in our review, clinicians may wish to educate patients about the potential broader health benefits of exercise when helping them choose treatments. 

Given that there is no strong evidence that any one type of exercise is more effective than another for knee OA, clinicians can feel free to guide patients to choose the types of exercise they prefer or feel most able to do.

Your study found that exercise adherence was not associated with better outcomes. Why do you think that is and what implications does it have for clinical practice? 

Much of the observed effects of exercise may relate to natural history, placebo and contextual effects, and regression to the mean. This is supported by two key findings. 

First, evidence has largely failed to demonstrate a dose–response relationship between exercise and changes in pain and physical function in patients with knee OA. 

Second, the effects of exercise on pain and function are not meaningfully explained by changes in strength, range of motion or proprioception.

Given that exercise dosage appears weakly related to clinical outcomes in knee OA and the physiological effects of exercise mediate the benefits from exercise, higher adherence to exercise interventions (ie, completing a greater overall dosage) may not produce better outcomes compared to lower adherence. 

However, the current evidence base has considerable limitations. 

Very few trials have evaluated whether exercise interventions with targeted behavioural support outperform those without such support.

Many RCTs on knee OA do not adequately report adherence, preventing a large proportion of these trials from being pooled in meta-analyses and limiting the confidence in systematic review findings. 

Studies that do report adherence largely rely on self- reporting of session completion, which is prone to recall bias (ie, over- or under-reporting of actual sessions completed).

Due to these limitations, we cannot confidently conclude whether adherence is important for clinical outcomes or to what extent
it matters. 

Targeting adherence in clinical practice may be less important for reducing pain and improving self-reported function. 

However, adherence could be important for other health outcomes where dose–response relationships have been observed.

Your study highlights the need for better reporting of exercise adherence in clinical trials. What specific recommendations would you give to researchers designing future trials on exercise interventions for knee OA?

Future trials need to ensure that they adequately report participants’ adherence to exercise interventions and clearly outline how this adherence has been measured. 

There are currently no perfect solutions to measuring adherence and many trials rely on self-reported data. 

Compared to measures of adherence with a device (eg, using accelerometers with Therabands), self-reported adherence tends to overestimate exercise compliance. 

However, using device measures is not always possible, depending on the type of intervention and resources available to researchers. 

Advances in digital solutions will likely provide solutions in the future. 

For now, more work is needed to develop user-friendly, scalable approaches suitable for varying clinical populations, such as those with low health and digital literacy levels.

You reviewed research on remotely delivered exercise programs. What does the evidence say about the effectiveness of telehealth and digital platforms for knee OA care?

The number of RCTs evaluating digital programs has increased greatly over the past five years. 

Overall, these trials suggest that digital programs are likely to be effective and safe for managing knee OA. 

Notably, digitally delivered exercises by a physiotherapist have similar effectiveness levels to in-person-delivered care for pain and function—some patients may even prefer it. 

Digital programs, unsupervised by a clinician or combined with some asynchronous support, also appear to be effective and may help overcome cost barriers to exercise care for knee OA. 

Stepped interventions, where patients receive increasing levels of support from a clinician, could help reduce the burden of knee OA on the healthcare system by funnelling resources to patients who are unresponsive to unsupervised, digital care.

Click here to access the paper featured in this article.

>> Travis Haber is a physiotherapist and clinical researcher. He is a postdoctoral researcher at the Centre for Health, Exercise and Sports Medicine at the University of Melbourne. His research focuses on improving nonsurgical management, including exercise therapy, for hip and knee osteoarthritis.