Randomised controlled trial of physiotherapy compared with advice for low back pain

Despite my limited clinical experience in this particular area, I
think there is one fundamental fact that is missing from the article in
question.
Whether we are providing a full-on, multiple-treatment service to a
patient with back pain, or a simple one-off consultation of assessment and
advice, surely we as physios are one of the ideal healthcare professions
to provide either or. This is what we spent three years at university for,
after all...
Initial physiotherapy assessment should provide the revevant information
about each patients condition, so the clinician can make the right
decisions about both the nature of each individual case, as they arise, as
well as how is best to go about solving the problem.
Having written a 3rd year literature review about how is best to combat
lower back pain, I think a more relevant question we need to ask is not
whether physiotherapy treatment/advice is effective or not, but rather is
the resource well-managed by 'the people upstairs', so that it is cost-
effective and readily available when it is needed?

Competing interests:
None declared

Due to the delay in my husband receiving his British Medical Journal
(BMJ) here in Saudi Arabia I am afraid this is a ‘not-so-rapid’ response.

Randomised controlled trial of physiotherapy compared with advice for
low back pain (BMJ Vol 329, pp 708-711)

Hoping that Dr Domhnall MacAuley’s editorial (BMJ Vol 329, 25 Sept
2004, pp 694-695) had a facetious rather than serious character, it
nevertheless filled me with concern about how most of his colleagues were
going to interpret and use the results of this study. I hope that with my
letter, even without going into the statistics of the paper, I may give
some of the readers food for thought and would like to take the
opportunity to recommend Trisha Greenhalgh’s excellent book on the basics
of evidence based medicine “How to Read a Paper”, in which the questions
presented here can be found.

While I recognize that Helen Frost et al’s paper has not been
published in its entirety I would like to limit my reply to the article in
the BMJ because this is what I presume to be the literature (and maybe the
tool for decision making) of the larger number of the subscribers.

I was surprised to see that the paper had been published without the
usual reviews and requests for alterations and corrections with subsequent
re-submissions being declared. These reviews are frequently announced as
footnotes in other journals and serve to prevent a publication of the
standard that I wish to discuss here.

As a physiotherapist I firstly have to take issue with the terms
‘routine’ and ‘standard’ physiotherapy treatment. I understand the term
‘routine medical check-up’ to be used in a scenario where patients
attend assessments in regular intervals that contain exactly the same
components (eg. blood pressure, weight, height, certain lab tests, ECG
etc), every time for all attendees.

While physiotherapy treatments contain standard techniques
(mobilizations, manipulations, various soft tissue techniques etc) and
modalities such as ultrasound, interferential, shortwave I would like to
think that any Physiotherapist worth his/her salt will baulk at the term
‘standard treatment’. Following careful assessment we should decide on a
specific approach in response to the problem posed by a particular
patient.

I also disagree with the dissection of physiotherapy treatments that
the authors have undertaken in their paper, separating electrotherapy from
advice and those two from ‘routine physiotherapy’ without giving a
definition of what exactly constitutes 'routine physiotherapy'. I would
like to stress here that advice and education are integral tools of
physiotherapy in long-term management, especially of low back pain.

I suggest that the research presented in the article has various
flaws, some of which I endeavour to point out in the following to the
doctors who may want to use the article in their decision making where the
use of physiotherapy in the management of their patients with low back
pain is concerned.

When we read a research paper we need to ask certain pertinent
questions in order to allow us to judge whether this particular paper is
relevant to us in our clinical practice and whether the information given
in it has been derived in a scientifically acceptable way. In the
following I would like to cover some of the questions that I asked myself
when I read the article and then leave the reader to draw his/her own
conclusions

1. Did the authors formulate a problem? Did they define the desired
outcome?

Even though the paper states the objective of the study it neither
formulates any actual problem, which would have made the investigation
necessary nor is a motivation (such as cost saving in the NHS) made known.
The tentative literature review of management of low back pain only
insufficiently familiarizes the reader with the topic.

2. Was a hypothesis formulated and was it tested by the study?

No hypothesis or null hypothesis is stated so we are left in the dark
as to what the researchers were actually trying to prove or disprove. In
good research a hypothesis should be formulated, which the researchers
then set out to disprove. If it withstands the rigours of good research it
can be regarded as not wrong. According to Richard Feynman (The Character
of Physical Law, 1965) the key to science lies in initially guessing a new
law, computing the consequences of the guess, comparing the result of the
computation to nature and if the guessed law disagrees with experiment it
is wrong. He warned that experimentation must be checked for errors and
oversights, which can easily creep into research at all stages.

3. With the information given, could the study be reproduced? Is
there any information (a table or paragraph) that allows for sufficient
comparison of the two groups?

The methods section of this paper is extremely short and sketchy in
that insufficient information has been given about the exact conditions
admitted and excluded, not enough has been said about the exact
instructions given to physiotherapists or about the precise nature of the
advice given to patients. For example, we do not know whether patients
with spondylolisthesis were admitted to the study and if the handout
covered advice for this condition. Further, no information has been given
about how the deviations from the original protocol were handled. We also
have very little information about the therapists, that is to say if all
therapists were of comparable standard. Their qualifications and years of
experience are not stated in the study.

4. Was this the best study design for the patients or for the
outcome? Was appropriate treatment given to all cases?

Despite having received ethical approval I am wondering whether this
study is entirely ethical. Should treatment be withheld from patients?
Even though the researchers would like to make us believe that their
research concludes that their protocol of ‘routine’ physiotherapy
treatment is not superior to ‘advice only’ this was not known beforehand.
This means that effectively NHS approved treatment was withheld from half
their subjects. My own doubts are reflected in the fact that some of the
participating patients complained about not getting treatment and some
participating therapists failed to adhere to the protocol for ethical
reasons.

5. Did difficulties compromise the study design?

Because neither all the patients nor all the therapists were
satisfied with the protocol, significant changes over which the
researchers had no control were undertaken while the study was in
progress. Out of 142 patients in the ‘advice only’ group 14 ended up
having hands on treatment of varying intensity. This makes up almost 10%
of the sample. We are neither informed about how the researchers handled
deviations from their protocol nor how they dealt with the results of
those deviations. The study design is therefore severely compromised.

6. Was a sample size calculation given in the paper? Was the sample
size sufficient?

No calculation was given but commonsense must tell us that, if
annually 1.3 million people receive treatment for low back pain in the UK,
a sample size of 144 and 142 subjects for each group over a period of
about three years cannot be sufficient to make a clinically or
statistically significant statement. Also only 70% of all patients
provided follow-up at 12 months. Trisha Greenhalgh states in her book that
this is at the bottom end of the scale before a study becomes invalid.
Hence the power of the study is low. And would not ‘last value carried
forward’ in lieu of actual data give a distorted view, especially if there
is only a 70% follow up for the final value?

7. Was the study performed under real life circumstances? Was the
design appropriate and sensible?

Trisha Greenhalgh warns that if a study has not been conducted under
real life circumstances (ie shorter waiting time for subjects,
interventions that are not normally available) this may cast doubt on the
applicability of the study to our own clinical practice. The first
deviation from a real life scenario is that in the trial the patient-
information leaflets differed from those usually given. However, the main
issue here is that patients were randomly assigned to the two groups while
in reality physiotherapists should carefully select which patients are to
be given hands-on treatment (as well as education and advice) and which
patients are suitable for advice only. Due to the random assignment of the
patients to the groups the result of this paper is not really applicable
to our clinical practice.

8. Did the analysis include only predefined ‘objective’ endpoints,
which may exclude other important aspects of the intervention? What
outcome was measured and how? Have the authors drawn justified conclusions
from their paper?

In my opinion the study failed to look at some of the important
objectives of hands-on physiotherapy, some of which are: faster pain
relief, quicker return to work, sport and activities of daily living, in
short reduced morbidity. As they are short-term outcomes they were not
considered since the main outcome measure of the study was the status of
patients at 12 months post intervention (with 2 and 6 months as secondary
outcome measures).

In the ‘therapy group’ 118 patients had six or fewer sessions. There
is no explanation as to the reasons, apart from adherence to the study
design. However, considering that some of the therapists used their
discretion and did not adhere to the protocol by giving more treatments if
they deemed it necessary this could also be interpreted as: no more than
six treatments were needed to achieve satisfactory immediate results in
82% of the patients of the ‘therapy group’.

This falls in line with the paragraph which states that patients from
the ‘therapy group’ were more likely to report benefits at two and six
months as well as in the 0-10 rating scale at all time points than the
'advice only' group.
Further, the fact that both groups were given advice as part of the
protocol diminishes the difference between the two groups and thus between
the expected outcomes of the two groups. In my opinion we therefore should
not expect any significant difference in the outcomes of the two groups at
12 months. My interpretation of the result here is that both groups
adhered equally to the advice given.

I will leave the decision whether the authors drew correct
conclusions from their paper up to the reader.

9. Is the evidence given in this article politically desirable?

Without wanting to sound cynical I have to ask whether this paper can
be regarded as evidence of anything but I suppose that the answer to this
question would depend on where one stands. Apart from the fact that the
outcome is in my opinion not clinically relevant for the reasons stated
above we still do not know about the motivation of the study and about who
commissioned it. If one wished to save cost in the NHS it certainly seems
that what the authors regard as the outcome of this study is politically
desirable as it would support an argument in favour of abandoning
Physiotherapy in the NHS.

One statement that Trisha Greenhalgh made in her book, and which I
have come across many times during my own clinical practice is: “What is
important for the doctor may not be important for the patient and vice
versa”. This is why I do not understand the discrepancy in the result as
reported and the statement made regarding reported benefits. It seems as
if noone was listening to what the patients really reported. At the end of
the day we should ask ourselves what we as clinicians are trying to
achieve. As a physiotherapist I regard myself as accountable to my
patients first and foremost and strive to act in their best interest to my
best knowledge and ability.

Friederike von Rabenau MSc Physiotherapist

Al Khobar
Saudi Arabia
Phone: ++ 966 3 8827711x 2221

Competing interests:
None declared

A very interesting topic to discuss. In particular it looks that the
authors have left the discussion to the readers as the paper is suffering
from a comprehensive discussion. They did not discuss the significant
benefit of physiotherapy in short term (two months). They did not
highlight the patients' perception of treatment benefit and why it was in
conflict with the outcome measures. As we know when the perception of
patient is positive, then it is more likely to return to normal activity
and work as a common goal in treatment of low back pain. Furthermore, the
treatment in this study is traditional and routine physiotherapy which is
not necessarily the optimal physiotherapy intervention, therefore, I don't
agree to generalise the findings to all physiotherapists in NHS. The
sample is not a good represetative of chronic low back pain population
because the mean of Oswestry disability index (ODI) at baseline is low in
both groups (21.12 and 21.60) in compare with nomative data for ODI (43.3
for chronic low back pain) (Ref. 1)

In general that is a good paper which cause us (as physiotherapist)
to re-evaluate our practice more critically. According to literature and
guidelines, when chronic back pain is a concern, psychosocial factors
should be addressed and more than 70% of the sample in this study are
chronic. That is surprising that the authors did not discuss about
biomedical and biopsychosocial approaches and their role in obtaining such
a result. The reader would like to know why there was not a difference
between two groups and what is the authors suggestion about it but the
athours keep silence about these important subjects.

Traditional medical model is mostly tissue oriented and focuses on
pain, it does not consider the factors involved in development and
persistence of the problems such as psychosocial factors (Ref 2). It has
been emphasised that physiotherapists will not be successful in their
treatment unless they address the patient's fear avoidence beliefs (Ref.
2). It has also been shown that physiotherapist's beliefs may have an
influence on management of chronic low back pain (Ref. 3).

Frost et al pushed the physiotherapists to choose a "standard
protocol reflecting routine practice in NHS" but it has been recommended
in the lierature to choose an individual problem solving approach,
individual plan with personal goals and individual monitoring of progress
by physiotherapists (Ref 2).

How to integrate and implement the evidence and clinical guidelines
into practice is another issue that should be taken into account (Ref 4)
and the gap between research results and actual practice should be closed
(Ref 5). Koes et al (2001)clarified that there is a need for systematic
implimentation strategies to change the behaviour of health care providers
(Ref 6).

Still it looks better to entitle the study in this way: Good
assessment and advice by a physiotherapist is as effective as traditional
physiotherapy. By this title the article does not mislead the media and
the media do not mislead the patients and health care professionals.

Abdolkarim Karimi-Physiotherapist

References

Ref 1 Fairbank, J. C. T. & Pynsent, P. B. 2000, "The Oswestry
Disability Index", Spine, vol. 25, no. 22, pp. 2940-2953.

Ref 2 Watson, P. 2000, "Psychosocial predictors of outcome from low
back pain," in Topical Issues in Pain 2: Biopsychosocial Assessment and
Management: Relationships and pain, L. Gifford, ed., CNS Press, Falmouth,
pp. 85-109.

Ref 3 Daykin, A. R. P. & Richardson, B. P. 2004,
"Physiotherapists' Pain Beliefs and Their Influence on the Management of
Patients With Chronic Low Back Pain", Spine, vol. 29, no. 7, pp. 783-795.

Ref 4 Muncey, H. & Watson, P. 1999, "Efficacy of a
Unidisciplinary Outpatient approach for patients with musculoskeletal
pain. Abstracts, Annual Scientific Meeting of The Pain Society,
Edinburgh".

Ref 5 van Tulder, M. W. P., Koes, B. W. P., & Bouter, L. M. P.
1997, "Conservative Treatment of Acute and Chronic Nonspecific Low Back
Pain: A Systematic Review of Randomized Controlled Trials of the Most
Common Interventions.", Spine September 15, vol. 22, no. 18, pp. 2128-
2156.

Ref 6 Koes, B. W. P., van Tulder, M. W. P., Ostelo, R. M., Kim
Burton, A. P., & Waddell, G. D. M. F. 2001, "Clinical Guidelines for
the Management of Low Back Pain in Primary Care: An International
Comparison.", Spine, vol. 26, no. 22, pp. 2504-2513.

Competing interests:
None declared

Dr. Nehrlich's response to my response seems to me,
in the circumstances, to be flippant, at best. As a
teacher/facilitator in the world of art and creativity I am
all too familiar with 'charming theories'; indeed the
entertainment of such theories is an important part of
'thinking outside the envelope'. Such theories could
properly be posted in some whacky, off-beat, fringe
journal. But the BMJ is none of those. It is, in effect, the
serious voice of British Medicine, and, as such, is
heard around the world, as this correspondence
confirms.

And so when the BMJ publishes something that will
likely have significant consequences for at least 1.3
million patients per year in Britain alone, it does so with
a considerable degree of responsibility. In my opinion
the conclusion that 'physotherapy is no better than
advice', what Dr.Nehrlich calls a 'charming theory', was
not based on solid ground; in fact it was pretty flimsy
ground. It might turn out, that, in the fullness of time,
this 'charming theory' proves to be true. But, as of now,
there are not sufficent grounds to say that it is.

Like Dr. Nehrlich, I commend the BMJ for having this
forum in which diverse opinions, including strongly
contrary opinions, may be heard. However it is my belief
that much damage has been done already by the
publication of this article; damage which will be very
hard to rectify, especially if, as I suspect to be the case,
the 'charming theory' proves to be false.

Neil Watson, MA, MD, FRCS
Artist and Writer
Formerly Consultant Hand Surgeon

Competing interests:
None declared