9 Scientific Research
of side-effects. So far the results do not hold any surprises. Their
importance lies in the comparison between true and placebo
acupuncture (acupuncture at non-acupuncture points).
For the treatment of lower back pain, both studies (GERAC and
ART) showed no significant advantage for true acupuncture; the
effect of acupuncture therefore seems to be independent of the
location of the acupuncture points. For the treatment of
gonarthrosis, the ART trial observed a significant short-term
effect in favour of true acupuncture. However, without consis-
tent treatments the disorder will progress, further data showing
no significant difference between true and sham acupuncture.
Results of the GERAC trial confirm the significant short-term
results. For the treatment of migraines and tension headaches,
the relevant ART trials could not establish a significant differ-
ence between true and placebo acupuncture. The migraine sec-
tion of the GERAC trial found a significantly reduced number of
days with headache in subgroup analyses. Regarding tension
headaches, up to now the GERAC reports show significant dif-
ferences in favour of true acupuncture.
How can we explain the good results of placebo acupuncture,
especially for the treatment of lower back pain? And how can the
differing results between the ART and GERAC trials for
gonarthrosis, migraines and tension headaches be explained?
Shouldn’t the trials, which are both designed according to the cri-
teria listed in Section 9.1, be very reliable because of the high
power of their design? The key to the answer – besides the already
well-known deficits of the GERAC trials (by international stan-
dards, with only 140 hours – lowest standard – of training, ques-
tionable choice of true and sham acupuncture points) – lies in the
so-called ‘internal validity’ of the collected data. The external val-
idation, especially of the GERAC trials, was carried out in an
exemplary manner: patients and physicians were separately ques-
tioned about the treatments by telephone interviewers. This fil-
tered out a high number of patients receiving unacceptable
additional therapies, thus increasing the validity of the data. How-
ever, it was hardly possible to control the internal validity of the
trial. With treatments taking place in hundreds of clinics, no onewas able to control if true or sham points were needled. One
peculiarity of both trials suggests that physicians treating patients
in the placebo group selected true acupuncture points: generally,
the participating physicians weren’t research scientists, but family
doctors who, by participating in the trial, were able to give their
patients treatments paid for by their obligatory health insurance,
thus securing a part of their own income. Therefore, participation
in the trials not only contributed to the income of these physicians,
it also meant that the crucial performers of the trials had other
goals in mind – patient satisfaction, their income – rather than the
trial itself (i.e. collecting reliable data regarding the effectiveness
of acupuncture). These conflicting goals particularly affected the
placebo-acupuncture, which had a higher potential of negatively
affecting patient satisfaction and the income of the participating
physicians, and which could not be balanced by external quality
assurance through telephone interviewers. The part of the
GERAC trials investigating tension headaches provides another
interesting insight. This section of the trials prohibited needling
of points located on the head and it was here that the treatment
protocol was disregarded most often (according to GERAC sym-
posium, RUB 16 November 2005). And it was exactly in this
externally validated part of the trial that GERAC observed signif-
icant differences between true and placebo acupuncture!
An important lesson to be learnt from the ART and GERAC tri-
als is that size and adherence to biomathematical standards alone
will not guarantee quality. If, in a trial above a certain size, the
data collection is compromised due to conflicting interests of the
initiators of the trial and the treating physicians, this will have a
considerable adverse affect on the validity of the study.
The above reservations should result in a rather cautious
interpretation of the results of large-scale trials such as GERAC
and ART.
Even conventional medical procedures, considered as proven
beyond doubt, recently failed spectacularly in controlled, dou-
ble-blinded trials (Moseley, N Engl J Med 2002), which means
that the efforts for scientific proof regarding specific details of
acupuncture will be seen in a different light.