Walking back the WalkBack Trial

Should healthcare professionals recommend walking for back pain? Should patients take that advice? If you’ve got back pain, can you “walk it off”? Does walking prevent back pain?

No one really knows, because it’s amazing how under-studied this topic is — as usual. But science has recently thrown us a bone, a study that justifiably inspired some good-news headlines … albeit with disappointing fine print. As usual!

At least two other recent back pain trials by some of the same researchers had Catchy Names® that brand them like products: the OPAL trial of opioids and the RESTORE trial of Cognitive Functional Therapy. Both were criticized by some for being over-hyped.

This time it’s the “WalkBack Trial” of the effect of walking on back pain.

Why study walking?

Exercise is often recommended to back pain patients as a “can’t go wrong” option, with a strong fitness upside. It’s worth exercising even if it doesn’t benefit backs, even though basically everyone believes that backs benefit. And yet no one has ever checked to see if a nice easy exercise like walking actually helps, as opposed to more technical and expensive options like, say, Pilates.

Now the WalkBack Trial seems to show that walking works.

Pocovi et al. tested a combined walking and education program over three years, comparing it to a neutral non-intervention group. They concluded that patients who participated in “an individualized, progressive walking and education intervention” enjoyed “significantly reduced low back pain recurrence,” and that the juice was worth the squeeze (“cost effective”).

Sounds great, and it probably is at least partially. But, as with both OPAL and RESTORE, the interpretation and conclusions may reach beyond the evidence — even if the data itself can be taken at face value. For instance, the study purports to show that walking “prevented” back pain. But that might be going too far.

You say “prevented,” I say “slightly delayed”

Most of WalkBack’s participants eventually got more back pain. Most of their back pain did recur. Is back pain “prevented” if it’s just a bit delayed? And not for all that long?

The results of WalkBack are statistically significant, the authors report, but they do not look “significant” — literally. Anyone can plainly see this in the paper’s figure 2B, the most relevant graph, which shows … a blatantly unimpressive visual contrast between recurrence with and without walking!

Does that 15% gap show that walking “prevents” back pain?

This is figure 2B from Pocovi et al., a graph of the “cumulative recurrence incidence” of back pain over about three years. In the early days, hardly anyone experienced any recurrence. As the months marched by, the control group started to suffer slightly more relapses than the coached walkers. By day 200, the gap was as big as it was ever going to get: roughly 80% of the control group had suffered more back pain, compared to about 65% of the coached walkers. Is a group of people enjoying “prevention” of back pain if 65% of them have had a relapse within the first year? Just 15% less than if they weren’t being coached & walking constantly? After that, the gap narrows steadily to almost nothing after three years: about 80% of both groups had experienced a relapse by the end. It barely looks like “prevention” in the short term & it really doesn’t look like it in the long term.

I accept that “prevent” is technically correct, from one perspective. But I think it’s important to emphasize that most people are not going to get “prevent” from this data. No one would prefer to live on the red top line. But no one is going to feel like they won the back pain lottery if they get to hang out on the blue bottom line. It’s more like a door prize.

And here’s an awkward question: if coaching/walking “prevents” back pain, shouldn’t it stifle relapses more over time? Instead, it looks more like it slowly but steadily loses potency, ultimately becoming indistinguishable from doing nothing in particular! (Which sounds a lot like everything else that supposedly helps back pain, doesn’t it? See “The greatest hits of back pain science are a disappointment.”)

And so this appears to be a classic example of a study result that is statistically “significant” enough to call it “significant” without that word being a lie … but also one that fails to clarify that it’s not very clinically significant result. (This is a word game that is all too common in the science business, which I am weary of pointing out.)

What if you already walk? No WalkBack for you!

WalkBack wasn’t a trial for active people. Pocovi et al. excluded anyone who was even moderately active before the experiment began. So this was a trial of walking for beginners. And exercise benefits are notoriously more robust for beginners! Whatever the goal, it quickly gets harder to make progress.

If WalkBack showed only modest benefits even for beginners, it’s likely that it would have shown even less benefit for active people with back pain. It’s not even clear from WalkBack that “walking prevents back pain” for beginners, and it almost certainly does not do the job for walking veterans.

None of this means it isn’t a great idea for inactive people with back pain to get more active … but it negates the value of WalkBack’s results for millions of people who were active before their back pain began.

The education elephant in the room

WalkBack was also a test of coaching, of receiving attention and encouragement and information over six sessions. This is already more than most walkers will ever get, but there’s even more unusual substance. Those sessions weren’t just about walking! Subjects were educated, taught a particular way of thinking about back pain.

The WalkBack gang probably spent six sessions training their study subjects to believe, among other things, that psychology may be just as much a problem for backs as pathology, if not more. That is probably the substance of what they describe as “modern pain science.” In other words, they were probably not just encouraging people to walk, but presumably to embrace unproven hypothesis that pain intensity and persistence is a partly a function of threat-perception — a hypothesis has become extremely popular in Australia and abroad, and is now arguably being overconfidently applied well in advance of strong empirical support (and often clumsily as well, so there's a theory-versus-practice problem too).

This is all straight from the RESTORE playbook, where the substance of Cognitive Functional Therapy (CFT) was spelled out in more detail last year. Here’s what I wrote about it then:

It’s really all about the fighting the fear. This is not just saying “unhelpful thoughts, emotions, and behaviours” are the problem; it’s even an more aggressively psychological focus on “thoughts, emotions, and fears.”

You don’t have to boil that down any further. It’s spelled out. … If you’re spooked into stillness, your back pain experience will be worse. If you can master your fear, it’ll be better.

… CFT is unmistakably rooted in a psychogenic pain hypothesis. But it sounds a lot more serious and professional to say “targeting unhelpful pain-related cognitions, emotions, and behaviours” than “you’re in serious chronic pain because you’ve put the fear-whammy on yourself and you’re too spooked to move.”

Which isn’t necessarily wrong, and there’s more to CFT, much of which I like. But we should be clear about what we’re talking about: WalkBack participants were likely taught an idea like this, which arguably is not actually established “modern pain science.”

Major confounders of overstated conclusions

I find it weird that this was framed as a trial of “walking” when it is so clear that there was another very specific educational ingredient. But even without that curve ball, the education obviously involved plenty of “good bedside manner” stuff, which is well-known to improve outcomes all by itself.

So the WalkBack results may not have that much to do with walking.

Consider: how well would the WalkBack walkers have done if they’d gotten insults instead of encouragement? If walking is good medicine for back pain, then it should perform anyway — but I wouldn’t bet on it!

Both walking and education were probably active ingredients here, and yet even the combination produced an underwhelming effect, damning walking with even fainter praise than it already looks like on its face. There’s just no way this study persuasively shows that “walking prevents back pain,” and yet that’s the big idea we’re supposed to walk away with.

Fortunately, walking doesn’t need to prevent back pain to be worthwhile: it’s a lovely way to exercise, and there are a lot of other health and social reasons to do it. Walk away!

I added a whole new chapter about walking to my immense back pain book. This blog post was basically just an excerpt from that, focusing on the WalkBack Trial. The book has a good-sized free intro. The whole thing is basically a collection of everything I’ve written about back pain over the last twenty years, and customers get permanent access to it. So someone who bought my back pain book in 2008 can login today and get that new chapter — and everything else I’ve added and changed over the years, which is a lot.