Abstract
Manual therapy and other physical and manipulative treatments are often recommended for patients with recurrent, benign headache disorders. The evidence base supporting their use, though, is not as firm as that for many pharmacological headache treatments (1). In fact, the authors of a recent systematic review concluded that more and better quality research is needed to establish the place of these therapies in headache care (2). The randomized controlled trial reported by Castien and colleagues is therefore a timely contribution to headache science.
The authors randomized participants with chronic tension-type headache (CTTH) to receive either manual therapy or “usual” general practitioner (GP) care for the duration of the study. The premise of the intervention was that improvement in craniocervical muscle function should lessen headache, so the manual therapy intervention was designed to improve cervical mobility, stability and posture. The components of usual GP care were not as clearly specified but consisted of elements outlined in Dutch headache treatment guidelines.
At both 8 and 26 weeks, the manual therapy group experienced a statistically significantly greater decline in headache days compared with the usual care group, and was more likely to achieve a 50% or greater reduction in headache frequency. The manual therapy group also did better than the usual care group on a variety of secondary outcome measures such as pain intensity and disability. Based on these findings, the authors conclude that “manual therapy is more effective than usual GP care in reducing symptoms of CTTH”.
Two things about this study are noteworthy. Firstly, this was a comparative effectiveness trial, which means that the study treatment was compared to another active treatment rather than to placebo or sham treatment (3). Secondly, this was a pragmatic trial, in which the investigators aimed to evaluate the intervention, as far as possible, in conditions that approximate those of real-life medical care.
Both of these attributes are virtues when viewed from the perspective of clinical decision making. In the real world of headache care, doctors and patients do not usually choose between “treatment” and “no treatment” but instead choose among available treatments. A study that compares two active treatments, then, provides more useful information for clinical decision making than does a placebo or sham-controlled study. The pragmatic aspect of this study, with both treatments tested in headache sufferers drawn from clinical practice and delivered under conditions typical of routine practice, further increases its clinical relevance.
Viewed from a scientific perspective, though, these virtues are not unalloyed. The tricky thing about comparative effectiveness research is the choice of comparator. Ideally, the treatment match-up should compare therapies that have been first been shown effective in comparison to placebo or sham treatment. This is not clearly the case for either manual therapy or for usual GP headache care following the Dutch treatment guidelines (4). Comparative effectiveness studies should also stage a “fair fight”, in which both treatments are delivered at a reasonable intensity and dose.
In this case, although study GPs followed Dutch treatment guidelines, their care does not appear to have been standardized to the same degree as the manual therapy intervention. In addition, GP care was not of a similar intensity, on average, to the treatment received by the intervention group. Usual care participants had an average of only two or three visits during the study compared to six for the manual therapy group. This “pragmatic” aspect of the trial undoubtedly reflects the real and busy world of “usual” GP care, but it does not allow us to judge how well optimized GP care would perform in comparison with manual therapy.
Thus, although the results of this study tell us that of the two tested treatments one appears better, they do not allow us to judge the absolute magnitude of the benefit of either treatment. Nor does the study design allow us to judge whether the benefits of one or both of these therapies might be due to nonspecific effects such as the time and attention subjects received from the manual therapists or GPs. Additionally, the unblinded nature of the study makes it difficult to judge the extent to which subjects’ prior expectations and beliefs might have influenced study outcomes. Further research, perhaps using sham manual therapy and optimized GP care, is needed to answer these important questions. Ideally, such research will also provide information on treatment costs and harms, and the durability of treatment benefits. In the meantime, we have much-needed preliminary evidence suggesting that manual therapy may hold promise in the treatment of TTH. It seems better than usual GP care, but we do not yet know why or by how much.