To our knowledge, this is the first controlled study comparing a GPR program to a physical therapy SE program in patients with persistent LBP. Our findings support the hypothesis that the GPR intervention (specifically the posture adopted) is effective in treating persistent LBP with low disability levels, when compared to SE. Patients allocated to the GPR group showed significant improvement in functional disability and pain intensity as compared to the SE group. Of note, the improvement obtained at short-term follow-up was maintained at mid-term follow-up for each outcome.
Our results are similar to those obtained in two RCTs by Fernandez-de-las-Peñas et al. [19, 20], who showed better results of the GPR as compared to a program of analytical exercises in patients with ankylosing spondylitis, and to those obtained in other studies on LBP cited in the review by Vanti et al. . Therefore, it seems that global reeducation is more effective in reducing pain and disability in subjects with LBP than segmental techniques. However, these results are different from those by Cunha et al. , who did not find different outcomes comparing conventional static stretching and muscle chain stretching in chronic neck pain. The reasons for these differences may be related both to the areas affected by spinal pain, and to the fact that GPR might be more effective when compared to analytical stabilization or mobilization techniques, although not superior to other stretching techniques.
When considering the clinical impact of our research , we can state that the GPR program produced a clinically meaningful improvement. In fact, 48% of subjects in the GPR group obtained a reduction of at least 30% in their RMDQ and VAS scores, compared to the 12% in the SE group. It should be noted that our results seem to demonstrate the effectiveness of the GPR program at relatively low disability levels.
In our study, the SE group obtained a slight improvement in functional disability, pain intensity and mobility of the whole spine and pelvis at the 3- and 6-month follow-up. The mean effects of the SE program were less relevant than those reported in some previous trials [39, 40], but are in line with those obtained in some other recent studies [12, 41]. An important aspect of the lack of agreement among these studies is the absence of subjects' subgrouping. According to Hayden et al., clinical trials should investigate the effectiveness of specific exercises in well-defined LBP subgroups . However, the identification of subgroups is a difficult process, since it cannot yet be guided by a coherent theory of causation of back pain . Moreover, the disability levels of the subjects included in our study are relatively low compared to previously published disability scores in 'chronic LBP' populations.
The main strength of our study was that patients undergoing a specific program were unaware of the presence of another training, because patients recruited by the same centre underwent the same treatment approach. Moreover, both treatments consisted of a one-to-one supervised exercise program actively involving the patient: according to the literature, these kinds of management are effective to reduce disability and improve function in chronic LBP [43–45].
Apart from the robust clustering, analysis could be adjusted only for patient factors (e.g. because of the differences in socioeconomic factors between patient arms we adjusted for socio-occupational status). On the other hand, there are other factors that might possibly explain, at least in part, the differences in outcome measures. Considering the differences in the physical therapists' experience, it has to be underlined that the more experienced physical therapists were those involved in the SE program. It is generally assumed that greater experience is associated with better clinical outcomes [46–48]. However, some recent studies evaluating the relationship between physiotherapists' years of experience and patient outcomes reported that years of experience were not associated with improved patient outcomes in outpatient rehabilitation [49, 50]. Moreover, in our opinion the size of the centers does not affect care quality, since this is strictly related to the close relation between patient and physical therapist.
The main limitation of this study was the absence of randomization. This practical choice could have led to some potential sources of bias: the two groups were similar at baseline, but slightly differed for socio-economical characteristics; this issue was considered in the multivariate analysis, taking also into account the possible different aggregation of patients within the different centres. The absence of randomization could have also influenced the patient's allocation to a specific physical therapist. Therefore, the possible influence of each physical therapists' experience and the psychosocial aspects of the patient-physical therapist relationship must be considered, as some physical therapists could have had the capabilities to establish a better relationship with their patients, thus positively influencing the effects produced by the treatment and the motivation for self-management.
Another important limitation was the elevated number of dropouts in the SE group (28%). Furthermore, dropped-out patients differed on critical baseline characteristics (RMDQ and VAS) from those who completed the study; on the other hand, they were equally represented in the two groups (SE and GPR). The high number of dropouts was mainly due to the determination of patients to give up treatment in the absence of expected results. This aspect was managed with the intention to treat approach, which confirmed the improvement obtained in the GPR group with respect to the SE group at the per protocol analysis.
Moreover, patients' adherence to the home exercise program could not be monitored. Indeed, adherence to the therapeutic program represents a crucial aspect of chronic LBP treatment [51, 52] and seems to be related to professional behaviors and explanations and to the total number of exercises prescribed [53, 54], as well as to individual and psychological characteristics . Finally, since the possible influence of physical and sport activities or cognitive-behavioral aspects was not considered, the presence of a complicated condition as a result of physical and psychosocial factors cannot be excluded (yellow flags) .
As a consequence, although we cannot definitely state that GPR alone is effective for patients with chronic LBP, GPR can be considered an important approach in the management of patients with persistent LBP and low disability levels. Because dropouts from both groups had higher levels of pain and disability than the subjects who completed the trial, we cannot apply our results to subjects with chronic, more disabling LBP.
Several physical therapy methods appeared effective in LBP; most of them have been studied by rigorous clinical trials. The recent trend of research on this topic allows to identify clinical prediction rules that can be applied to LBP subgroups. Some of the proposed classifications for subgrouping patients with LBP are related to techniques as manipulation, stabilization, specific exercises, or traction. Global postural treatment is not considered within that subgrouping. This is mainly due to a general lack of knowledge and to the little evidence of effectiveness of this method. Our study may represent the first step in this direction, but it should be followed by further, higher level studies. Accordingly, some suggestions for future research are the measurement of the clinical effects of the GPR on specific LBP subgroups, with respect to the age of patients, the phase of disorder (e.g. acute, subacute or chronic pain), and the clinical characteristics. Moreover, the effectiveness of GPR should be compared with other techniques, as manipulative therapy, cognitive behavioral therapy, etc.