Key findings
To the best of our knowledge this study is the first to investigate Italian physiotherapists’ knowledge, attitudes, and beliefs about SM, and the influence of prior educational background on SM practice. Our results provide validation of previous surveys conducted in the USA and the Netherlands; more specifically, this survey confirms that physiotherapists’ beliefs on the safety and effectiveness of SM strongly differ between the cervical spine and other spinal regions [
27,
28]. Similar to British, Canadian, American, and Dutch physiotherapists, Italian physiotherapists reported that they regularly use SM in the management of their patients for a variety of musculoskeletal conditions; furthermore, Italian physiotherapists reported being comfortable in the delivery of SM to the thoracic spine, but less so for the cervical spine [
26‐
28,
31]. Notably, the utilization, the comfort, and perceived safety for upper cervical SM by Italian physiotherapists differs considerably from other spinal regions with several barriers being identified [
27]. Respondents with a background in traditional non-thrust manual therapy (i.e., Maitland) reported using SM significantly less often, perceived SM as less safe and less effective, and reported less comfortable with performing SM; furthermore, they reported more frequently performing additional screening prior to the use of SM.
Beliefs about safety and the influence of educational programs
Similar to previous surveys, we observed that most physiotherapists generally agreed that SM is safe and effective in all spinal regions [
26‐
28,
31,
51]. Although reviews with less methodological flaws have demonstrated that SM is safe and a causal link between SM and SAEs is still unproven [
18], physiotherapists from SA, UK, Canada, the USA, the Netherlands, and Italy reported cervical SM—especially upper cervical SM–the least safe and effective [
26‐
28,
31,
51]. Although extensive literature on the safety of SM relies mainly on communicated opinions and on poor-quality non-systematic reviews [
22,
34], one reason could be that most of the studies investigating SAEs following SM were focused on the cervical spine [
22]. Furthermore, these beliefs appear to be mainly based on anecdotal reports rather than high-quality clinical studies, and concerns about the safety of cervical SM are still consistently propagated within post-graduate programs and academic settings [
52‐
55].
Adams and Sim (1998) were the first to observe that educational programs have a direct influence on practice—i.e., many UK physiotherapists have a background in the Maitland paradigm (i.e., a traditional non-thrust manual therapy approach); hence, not surprisingly, they emerged as being less frequent users of SM [
26]. Similarly, in our survey, Italian physiotherapists with a background in a traditional non-thrust manual therapy paradigm (i.e., Maitland) tended to utilize SM less frequently for all spinal regions; furthermore, they reported SM as less safe, less effective, and being less comfortable with the technical delivery of SM techniques in clinical practice. “Lack of practice/training” have been consistently reported as one of the major barriers to performing SM [
26‐
28,
31]; that is, possessing a musculoskeletal specialization, having attended a continuing professional development course on SM, or an osteopathy program are significantly associated with higher levels of reported comfort with performing SM when compared to not having previous education/training on SM and having a background in a traditional non-thrust manual therapy approach. Interestingly, no difference was found between the latter; moreover, one reason for this could be that SM receives less emphasis and less formal training when it comprises just one small aspect of a traditional non-thrust manual therapy education program [
35]. Furthermore, in our study, physiotherapists with backgrounds in traditional non-thrust manual therapy paradigms reported the most frequent use of pre-manipulative testing for the upper cervical spine. Therefore, when physiotherapists with a background in traditional non-thrust manual therapy consider the use of SM, especially to the upper cervical spine, it appears their beliefs on the risks outweigh the perceived effectiveness.
For the reasons above, widespread confirmation bias may have influenced clinicians to consider non-thrust mobilization as a safer approach and led researchers to investigate non-thrust mobilization as an alternative intervention to SM [
1,
31,
56‐
60]. Moreover, perhaps due to the perceived risk that is mainly based on anecdotal reports, many researchers have avoided investigating the effectiveness of SM to the cervical region, and have instead focused on investigating alternative approaches such as thoracic manipulation for the treatment of neck pain [
61,
62]. In addition, all of the above may be some of the reasons why these two very different treatment techniques progressively became considered by some to be synonymous, interchangeable, or a progression of the same approach (namely, mobilization/manipulation) [
1,
31,
56‐
59], as observed by the preference to use mobilization as the first-choice treatment by Italian physiotherapists.
Another example of the influence of educational background concerns the phenomena of audible ‘popping’ sounds during SM. Although the etiology of the popping sounds is still under debate, the collapse of gas bubbles within the zygapophyseal (i.e., facet) joints (i.e., the “cavitation” phenomenon) has traditionally been accepted as the main mechanism [
63,
64], as well as for most Italian physiotherapists. Nevertheless, a more recent study provided tribonucleation as an alternative mechanism to the popping sound [
65]—that is, the use of cine MRI appeared to confirm that the audible popping sounds occurred at cavity inception, and no bubble/cavity collapse was ever visualized. Furthermore, several recent studies reported a mean of 3–4 audible popping sounds following thrust spinal manipulation to the upper cervical spine, cervicothoracic junction, and lumbosacral spine [
66‐
68]. Therefore, the cavitation hypothesis alone appears unable to explain the multiple audible popping sounds, and the different frequencies and waveforms associated with these sounds, following SM [
66‐
68].
The popping sound is still a controversial topic for both the effectiveness (i.e., by influencing clinical outcomes) and the definition of SM. Notably, IFOMPT included the popping sound in its definition of SM [
3]; furthermore, Evans and Lucas [
69] concluded that the popping sound is one of the five necessary criterion for a valid definition of SM. Nevertheless, in the current survey, most Italian physiotherapists did not consider the popping sound as an important indicator for the successful technical delivery of SM. The notion that the popping sound is not related to clinical outcomes originates from a few studies [
70‐
73] and traditional approaches [
35]. However, most practitioners anecdotally believe that the popping sounds are an indicator of the successful delivery of SM, explaining why researchers often repeat SM when the popping sound is not elicited on the first attempt [
10,
66,
74‐
82]. In addition, some studies have observed preliminary evidence that suggests a greater hypoalgesic effect (i.e., perhaps associated with proinflammatory cytokine secretion, temporal sensory summation, and/or supraspinal mechanisms) in subjects that experienced audible popping [
73,
83,
84]. Although no firm conclusions can be drawn about the clinical relevance, when the popping sound was a requirement in the methods, SM was found more effective in reducing short-term pain and disability than non-thrust mobilization (i.e., no audible popping sound) [
74,
82]. Interestingly, patients themselves appear to expect popping sounds to accompany SM [
85,
86]. Therefore, the assertion that the popping sound is not required for a successful SM is not supported by the three-pillars of evidence-based practice [
87].
Clinical prediction rules and additional screening prior to SM
Prescriptive CPRs has been designed to help guide clinical decision making to provide SM only to those patients that are likely to benefit from such treatments and to attempt to reduce the presumed risks of SM [
19]. Even though multiple systematic reviews have raised concerns about the value of these prediction rules and the validation of these tools [
19,
88], similarly to Puentedura et al., we found that being aware of spine CPRs impacted the beliefs on safety, the perception of effectiveness, and the utilization of SM [
28]. Notably, the fact remains that none of the prescriptive CPRs in physiotherapy are recommended for application in clinical practice as validation and impact analysis studies are still lacking [
89].
With the attempt to identify those patients at increased risk of having an SAE following SM, and according to researchers and academics that still recommend its use, [
20] a significant proportion of our survey participants agreed with using pre-manipulative testing for the upper cervical spine. However, the validity of pre-manipulative testing for screening purposes has been questioned due to the low sensitivity and low specificity of the test procedures themselves [
90‐
93]. Thus, because of a lack of construct validity, questionable safety, and an inability to predict SAEs, IFOMPT and many researchers have recommended pre-manipulative testing be abandoned [
25,
40,
50,
68,
94,
95].
These widely established beliefs seem to outweigh the impact of current literature on physiotherapists’ practice; hence, de-implementation of the clinical use of these pre-manipulative tests appears to be challenging. Undergraduate programs have been observed not being able to translate updated literature from academic training to clinical settings [
96]. However, we observed that possessing a musculoskeletal specialization decreases the odds of using CPRs and additional pre-manipulative testing prior to SM, positively impacting the implementation of current literature into practice.
SM utilization, comfort, and barriers
Half of the Italian physiotherapists reported regularly utilizing SM, but only one quarter would regularly provide upper cervical SM. Notably, this utilization rate is similar to Canadian and the USA physiotherapists [
28,
31]. However, these results are much lower than the rates reported by physiotherapists in the Netherlands and UK [
26,
27]. This may be explained by the fact that Kranenburg et al. [
27] and Adams and Sim [
26] limited their survey to musculoskeletal physiotherapists, whereas the current survey and prior USA studies included all physiotherapists, irrespective of advanced training [
28]. Accordingly, and as in the USA, physiotherapists possessing a musculoskeletal specialization were more likely to use SM and more comfortable when doing so. That is, it has been previously observed that SM is taught to a lesser extent within physiotherapy undergraduate programs and that advanced training has the potential to influence the reasoning, the decision-making, and the technical skills of professionals [
33,
97,
98]. It is also important to note that a large amount of variation exists between the training and the content provided within programs and between countries [
33]. Interestingly, using CPRs increases the use and the comfort of both Italian and USA physiotherapists, suggesting that using a decision-making tool that is supposed to help identify those patients that are most likely to favorably respond to the use of SM, decreases the level of concerns on SM safety.
Although no convincing evidence supports a causal link between cervical SM and SAEs [
17,
18], and SM is still recommended in CPGs as an effective intervention to treat neck pain [
8,
14,
15], similarly to the Netherlands and the UK [
26,
27], we found a significant difference in the use of SM to the upper cervical spine when compared to other spinal regions. Also, respondents reported experiencing barriers to the use of upper cervical SM three times more than in the other spinal regions. Thus, physiotherapists worldwide seem to possess different beliefs about the perceived risks of upper cervical SM. The main causes reported in the literature appear related to anxiety on the safety of SM and the lack of clinical expertise in this spinal region. Notably, these barriers seem to be dogmatically influenced by traditional theoretical constructs (e.g., Maitland) [
26]. The contrasting caution levels in the different spinal regions between SM and mobilization appear to be based on the empirical assumptions that upper cervical SM increases the risk of SAEs [
40]. Notably, Michaeli reported SAEs more frequently following mobilization in different cervical regions [
30]. Additionally, although Canadian physiotherapists commonly use less cervical SM because of a perceived association with SAEs [
31], they reported having experienced minor to moderate adverse events with the same average occurrence for both mobilization and SM [
21].
As well as physiotherapists from the Netherlands and the USA, Italian physiotherapists are generally comfortable performing SM [
27,
28]; however, their comfort level drastically drops for the use of upper cervical SM. Interestingly, anxiety on safety as a barrier to perform SM was a major difference between Italian, UK, USA, and Dutch physiotherapists [
26‐
28]. Our results showed that Italian physiotherapists were more focused on the patients’ beliefs and preferences instead of their anxiety surrounding safety, recognizing the patient-centered approach as a key feature of their practice [
99]. Our results also demonstrated that working in a direct access setting and having more years of practice is associated with an increase in comfort level; that is, giving the opportunity to practice their skills increases the physiotherapists’ confidence [
100‐
102]. In line with UK, USA, and the Netherlands, most respondents perceived “lack of practice/training” and “lack of clinical experience” as major barriers to performing SM for all spinal regions [
26‐
28].
Implications on clinical practice and future research
Although the general trend from contemporary systematic reviews suggests that SM is a valuable and cost-effective treatment for musculoskeletal pain [
12,
103], Italian physiotherapists do not consider SM as an important core skill set. In addition to the anecdotal beliefs mentioned above, one primary reason may lie in the debate on the abandonment of hands-on interventions seen in, for example, social media-based opinions [
12]. With the attempts to discredit SM, advocates find fertile ground to prove the reductionist juxtaposition between hands-on and hands-off approaches in the methodological flaws of SM primary studies. The results of the majority of SM clinical trials should be considered with caution, as they mainly rely on a traditional construct of SM—e.g., proper technique selection [
7,
12], expecting one single popping sound from a single targeted joint [
66‐
68,
75,
104], reliable/valid palpatory skills [
105‐
107], correcting peripheral impairments [
5,
7]–which has long since been outdated [
5,
7]. Another confounding factor is tending to average results across heterogeneous substrates; more specifically, SM is skill-dependent in both the application and the execution, and there are technical differences between SM treatments done by osteopaths, chiropractors, and physiotherapists [
1]. Nevertheless, the contention that SM is a real treatment with therapeutic effects and few harms, is even supported by the averaging of heterogeneous data [
8,
108].
Although the traditional features of manual therapy have been strongly challenged [
5‐
7], Italian physiotherapists still rely on biomechanical concepts to decide whether to use SM and to determine its effectiveness (e.g., evaluation of passive accessory inter-vertebral joint mobility). Importantly, SM has been found to involve both biomechanical and neurophysiological mechanisms [
5]. However, although the interaction of these mechanisms has been frequently neglected, their combination could explain the reason for the effectiveness of SM despite its implementation heterogeneity. In addition, it might explain the irrelevance of SM traditional features on outcomes [
5]. That is, continuing to attribute the effect of SM to peripheral biomechanical mechanisms is too simplistic. Therefore, SM should be re-conceptualized in a broader multidimensional framework that embraces the complexity of pain and respects the patient singularity [
7,
105]. Such a comprehensive and dispositional approach accounts for the dynamic interplay between a myriad of factors, such as the sensory, cognitive, and affective processes, situationally influenced by expectations, mood, desires, culture, and past experiences [
7]. Nonetheless, manual therapy, including SM, is a socio-culturally integral part of the professional identity that respects the patient’s expectation when they seek help from a physiotherapist [
60]. Thus, SM remains an important skill within physiotherapy and does not merit being replaced because it might have been deemed to be outdated [
109].
Many randomized controlled trials lack of pragmatism across the Rating of Included Trials on the Efficacy-Effectiveness Spectrum domains (i.e., they emphasize efficacy); additionally, randomized controlled trials are not designed to determine how moderators influence the treatment effect [
110]. We believe that perhaps it is time to also focus on understanding how SM works and what are the mechanisms behind it. Future studies should attempt to establish links between the associated responses to SM and clinical outcomes, and the covariance of their changes. Therefore, there is a clear need for more focused research to understand what SM actually does, and how we might do it better; additionally, a mechanistic-based approach may provide a more robust approach to the design of clinical trials. Also, qualitative or mixed methods research should attempt to establish the nature of patients’ expectations and physiotherapists’ beliefs.
Our results suggest that anecdotal beliefs from prior educational background, research and practice are reciprocally influenced and lead to the propagation of misconception on the appropriate use of SM in clinical practice. However, specialization and updated programs seem to impact the implementation of current literature into practice. Given the paucity in the current literature, this article serves as an updated framework on the evidence-based use of SM. An infographic was designed for public use with the dual objective of raising awareness among physiotherapists about this subject and providing practical and easy to implement resources for the everyday use of SM in clinical practice (Additional file
3: Appendix 3).
Strengths and limitations of this study
This study is built on existing surveys and adds scope to explore the differences with prior studies on this topic [
27,
28]. One key strength is the rigorous survey’s developmental process in line with previous published surveys [
27,
28,
48‐
50]. However, the content and face validity of the questionnaire could have been improved by statistical testing. Another limitation is that the survey was not translated forward and backward. The achievement of the required sample size confirms the willingness of physiotherapists to participate on this topic. Furthermore, although we did not send personal invitations, the publication of several reminders helped to reach a number of Italian physiotherapists in line with previous surveys [
48‐
50]. Where the purpose of a study is to gain a general sense of a belief or attitude, a lower level of precision may be acceptable, and hence a smaller sample size may be drawn [
47]. Although we obtained the required sample size calculation, this study does have limitations in the generalization of the results. Although the web-link to the survey was only distributed initially by email to the members of the Italian Physiotherapists Association, the participation invitations methods used in this survey are potentially subject to selection bias. As reminders were sent via mailing list and publicly posted on social media it is not possible to know how many people saw the reminders. Therefore, our results could be challenged in their generalizability. For example, this could potentially be an explanation for the relatively "young" sample in this study. Additionally, responder bias is possible because of the very detailed and specific questions in our survey, and the potential influence of the point of view of the survey respondents by the public debate of the topic within the profession should be considered. Although previous surveys endorse our results [
27,
28], generalizability of findings may be limited, because we included only Italian physiotherapists. That is, educational standards between professions providing SM and national standards may differ.