In part I of this series I talked about myths associated with methods used alongside spinal manipulation (and got fired because of it). In part 2 (not available in english) I discussed research on the short-term effects of spinal manipulations (tl;dr: pain relief, maybe some range of motion improvements). In this third installment I’ll discuss the long-term effectiveness of spinal manipulation treatments.
Most people seeking physiotherapy aren’t looking to receive a palliative therapy for the foreseeable future, they view it as curative – a part of really getting better.
Answering a question such as “do spinal manipulations work?” is a fairly complex endeavor. When doing a randomized trial one needs a large group of people to demonstrate a small treatment effect, but a small group will suffice to demonstrate a large treatment effect. It’s fairly easy to establish a very large treatment effect on an otherwise stable condition (like thyroid hormone therapy where there is no placebo effect on serum hormone levels). So the fact that this question is still even being asked today already implies that spinal manipulations likely have a small or non-existing treatment effect. Basically, if they worked really well for a significant number of people (i.e. they worked as well as spinal manipulative therapists will tell you), one would only need a couple of medium or small sized studies to settle the issue.
Further, if one follows two groups of people who have a condition that is unstable – people frequently improve on their own, or get worse on their own – it’s fairly likely that the groups will be unequal at the end, regardless of interventions. Large powerful intervention studies are expensive, and so only a few of them are performed. Small studies with conflicting results give little or no information individually, but allow cherry-picking the data. So those small studies need to be combined into a meta-analysis, and it’s one of these that I will center this discussion around, a study published in 2014, authored by Menke.
A meta-analysis like this pools the treatment effects from multiple clinical trials and uses statistics to compare interventions. The strength of the Menke analysis is in how inclusive of studies it is (more difficult to cherry-pick), and the fact that he does a comparative analysis of different treatment groups from spinal manipulation trials. His analysis is highly informative and extremely interesting, even if the paper is in many ways unorthodox. The main conclusion to be drawn from the paper, sorry for the spoiler, is that spinal manipulative therapies provide no overall benefits long-term.
I will present here some key points from his analysis, that help to answer the question of how well spinal manipulations work clinically in the long-term. I say clinically, because there is a great paradox of why these therapies remain popular despite the evidence of their ineffectiveness and their weak theoretical foundation. Some parts of his analysis reveal components of the “formula for success” of how and why spinal manipulative therapists thrust cash in their pockets. The interested reader is of course advised to read the original meta-analysis.
#1 – A waiting list is significantly worse than doing nothing.
Most treatment groups had comparable results if they were manipulations or electrotherapy or what have you. But treatments that required any human contact outperformed those without human contact. People put on a waiting list (a waiting list control) were the only group that worsened in the short term. So it’s better to purposely do nothing than to wait for a therapy. It’s reasonable to assume from this that going from a waiting list to any intervention at all will always result in an improvement, so a waiting list is a good strategy to improve patient satisfaction without improving treatment effects.
#2 – The effect size of spinal manipulations remains unchanged since the first trials
Menke converts the results of individual studies into effect sizes which may be compared between studies – think of effect sizes as a unit of comparison – this is similar to converting pounds and stones into kilograms to compare people’s weights. What this reveals is that the effect sizes of spinal manipulation have remained the same from when trials on them started. Improved study quality, better health care, MRI etc. have not lead to a bigger effect of the spinal manipulations. As spinal manipulation outcomes are unchanged despite better research and healthcare, it’s reasonable to assume that they will be unchanged in the foreseeable future as well.
#3 – The prognosis is actually good, with or without interventions
One of the difficulties in researching back pain treatments is that the condition is unstable. This means that spontaneous improvements are common, and the pain and disability vary greatly due to natural fluctuations throughout the year. Most people seek treatments (and register into clinical trials) at their worst, and will then improve regardless of the intervention. So for this reason, most interventions that cause little or no direct harm will have good clinical results.
#4 – Spinal manipulation success has the same statistical chance as a throw of the dice
If treatment wings within trials were classified according to whether or not they outperformed sham treatments, the chance of spinal manipulations being the best treatment was equal to a throw of dice where the sides on the dice equal the treatment arms. Two treatment arms – 50% chance. Three treatment arms – 33% chance, etc.
When two treatments are compared on an unstable condition with small groups, it would be expected that the groups were unequal at the end of the study despite treatments. Any treatment under study needs to outperform this natural variability to show any meaningful results. Spinal manipulations fail to do this.
#5 – It makes no difference who performs the manipulation, or how, or why
In 30+ years of spinal manipulation research, no group of therapists has proven more effective than another. They’ve tried chiros, physios, bonesetters, you name it. It also makes no difference if the specific technique is chosen by the therapist based on an examination or by a researcher based on nothing.
#6 – Out of all the comparison groups, only one intervention stands out
You guessed it. Exercise. Something which actually has a demonstrable long-term effect on how the body functions is the only intervention that works better than sham therapies, including manipulations. It may not matter much which type of exercise or training system is chosen, but it’s important to get the patient moving. This is in-line with most clinical guidelines for treatment of back pain (1), so it’s hardly surprising.
Discussion
Strength:
While being far from conclusive for all patients, under all circumstances at all times the Menke article has many strengths to it. The biggest in my mind is that it answers one very important clinical question that is worth emphasizing:
If a person with low back pain decides on his/her own to see a therapist that does spinal manipulations, the results will be in line with Menke’s review.
Recent research in spinal manipulative therapy has come up with clinical prediction guidelines to predict which patients respond most favorably to spinal manipulations (1). Applying these guidelines must be done clinically, preferably by the first contact doctor or physio. If the person sees a spinal manipulative therapist first, the benefit will be non-existent.
Limitation:
It would not be fair of me to only discuss the strengths of the review without any of the limitations. The review does not discriminate between patient groups. It’s possible that future studies will identify means to effectively apply spinal manipulations, perhaps due to clinical prediction rules or sophisticated use of emerging technology such as ultrasound imaging. However, the key point remains that currently the lack of long-term effectiveness of spinal manipulative therapies are largely known for the broad strokes of clinical populations.