Physiotherapists' awareness, knowledge and confidence in screening and referral of suspected axial spondyloarthritis: A survey of UK clinical practice

Funding information National Axial Spondyloarthritis Society Educational Bursary Abstract Background: Axial spondyloarthritis (axSpA) is an inflammatory disease associated with significant diagnostic delays and is commonly missed in assessments of persistent back pain. Objective: To explore musculoskeletal physiotherapists' awareness, knowledge and confidence in screening for signs, symptoms and risk factors of suspected axSpA and criteria for rheumatology referral. Design: An online UK survey was undertaken combining back pain vignettes (reflecting axSpA, non‐specific back pain and radicular syndrome) and questioning on features of suspected axSpA. Recruitment utilised online professional forums and social media. Data analysis included descriptive statistics and conceptual content analysis for free text responses. Results: 132 survey responses were analysed. Only 67% (88/132) of respondents identified inflammatory pathologies as a possible cause of persistent back pain. Only 60% (79/132) recognised the axSpA vignette compared to non‐specific low back pain (94%) and radicular syndrome (80%). Most suspecting axSpA would refer for specialist assessment (77/79; 92%). Awareness of national referral guidance was evident in only 50% of ‘clinical reasoning’ and 20% of ‘further subjective screening’ responses. There was misplaced confidence in recognising clinical features of axSpA (≥7/10) compared to knowledge levels shown, including high importance given to inflammatory markers and human leucocyte antigen B27 (median 1⁄4 8/10). Conclusions: Musculoskeletal physiotherapists may not be giving adequate consideration to axSpA in back pain assessments. Awareness of national referral guidance was also limited. Professional education on screening and referral for suspected axSpA is needed to make axSpA screening and referral criteria core knowledge in musculoskeletal clinical practice, supporting earlier diagnosis and better outcomes.


| INTRODUCTION
Spondyloarthritis is an umbrella term for a group of systemic inflammatory disorders which includes axial spondyloarthritis (axSpA). axSpA causes enthesitis involving inflammation of spinal ligamentous and tendinous attachments to bone, which in its advanced stages can lead to joint erosion and fusion of the vertebral and sacroiliac joints (Danve & Deodhar, 2015;Kiltz, Baraliakos, Regel, Bühring, & Brau, 2017;Sieper & Poddubnyy, 2017). A characteristic feature of axSpA is persistent, and insidious back pain with inflammatory features of prolonged morning stiffness and pain which improves with movement but not with rest, termed 'inflammatory back pain' (Kiltz et al., 2017). Other features of axSpA may include peripheral manifestations involving dactylitis, enthesitis (typically at the insertion of the achilles tendon or plantar fascia) and inflammatory arthritis (Sieper & Poddubnyy, 2017). There is also an association with extra-articular inflammatory conditions of uveitis, psoriasis and inflammatory bowel disease (Danve & Deodhar, 2015;Sieper & Poddubnyy, 2017).
Research undertaken in general practice (GP) (Jois, et al., 2008;Tangrungruengkit, et al., 2016) and specialities involved in extraarticular features of spondyloarthritis (Sykes, Hamilton, Jones, & Gaffney, 2018;Villani et al., 2015) have explored potential reasons for diagnostic delay. A lack of awareness of and screening for axSpA is an important factor.
A cluster of features typical of 'inflammatory back pain' have been identified in 89% of axSpA cases, which have formed the basis for suspecting axSpA and the development of referral strategies (Rudwaleit, van der Heijde, et al., 2009;Sieper, Rudwaleit, et al., 2009).
However, people with features of inflammatory back pain alone may not always receive a diagnosis of axSpA (Arnbak et al., 2018;Danve & Deodhar, 2015;Poddubnyy et al., 2011;Rudwaleit, van der Heijde, et al., 2009). Signal changes (modic changes) at the vertebral body bone marrow and neighbouring end plates seen on imaging in degenerative disc disease can present as inflammatory back pain (Arnbak et al., 2018) and sacroiliac changes, such as bone marrow oedema and sclerosis, can also occur in populations without axSpA (Weber et al., 2018

| Research design
A cross-sectional online survey of musculoskeletal physiotherapists working in the UK was undertaken from February to May 2018.

| Recruitment and sample population
Recruitment was directed at physiotherapists with at least one experience in the assessment of persistent back pain. An initial survey question was used to filter out respondents who had never treated a person with persistent back pain.
The survey was promoted through professional networks, musculoskeletal forums and social media (Twitter, LinkedIn, Facebook and Physio Forum). Invitation emails were sent to postgraduate MSc students enrolled at the University of Hertfordshire, physiotherapy members of National Axial Spondyloarthritis Society (NASS) and AStretch, an axSpA specialist interest group. Permissions were sought to promote the survey and respondents were asked to snowball the survey to enhance response rates. Participation was self-selected and anonymous. On log-in, a participant information sheet was provided, and informed consent was assumed through completion of the survey.

| Survey design
A multi-strategy survey design was used which had three sections. A draft survey received feedback from a selection of experienced musculoskeletal physiotherapists, consultant rheumatologists and researchers experienced in survey design before piloting with several clinical physiotherapists using Bristol Online Survey Tool.
Amendments based on feedback were incorporated at both stages. The development of the vignettes and clinical questions was informed by a literature review, clinical practice guidelines and strategies utilised in previous survey research. These included using the same age to avoid age-related factors (Bedson, Jordan, & Croft, 2003) and basing presentation on real client cases (Bishop, Holden, Ogollah, & Foster, 2016). The vignettes were designed to avoid ambiguity between presentations and be typical of the three diagnoses (axSpA, non-specific back pain and radicular syndrome).
Section One contained case presentation vignettes of persistent back pain (>3 months), each followed by a set of text questions. The first vignette was common across all surveys and open-ended questioning on screening for serious pathology and other differential diagnoses. This vignette and questioning were designed to assess whether respondents included an inflammatory cause in responses.
The second vignette was a case presentation of either non-specific back pain or radicular syndrome. The third vignette was one of two axSpA case presentations. These vignettes were all followed by the  Rudwaleit et al., 2009aRudwaleit et al., , 2009bSieper, van der Heijde, et al., 2009), European Spondyloarthropathy Study Group (ESSG; Dougados et al., 1991) and Berlin criteria for inflammatory back pain (Rudwaleit, Feldtkeller, & Sieper, 2006) (see Table 1: Additional criteria from previously published referral criteria). Three sought demographic information.
There was no time limit on how long respondents could take to complete the survey. The participant information sheet advised that it would take approximately twenty minutes with Section One requiring much of that time. This time was determined by feedback on time burden from the respondents in the pilot process.

Data was exported to Microsoft Excel® (Microsoft Corp) from Bristol
Online Survey (Jisc) and analysed using conceptual content analysis and descriptive statistics (Robson & McCartan, 2016;Rossi, Serralvo, & Joao, 2014). The content of the free-text responses was analysed by the main researcher (Eliza Steen ) for a priori features (based on NICE (2017) guidance referral criteria) and emergent features, assigned into categories and subcategories where applicable and then assigned numerical codes (see Table 1).
Descriptive statistics were used to analyse the frequency of the numerical codes within and across responses. The number of codes within responses were used to reflect levels of awareness of the signs, symptoms, and risk factors of axSpA and were graded; full awareness, good awareness, poor awareness, or no awareness (see Table 1). Associations were also analysed between vignette response STEEN ET AL.
-3 results, guideline familiarity, confidence, level of knowledge and various demographic data.

| RESULTS
One hundred and fifty physiotherapists responded to the survey, with 132 usable data sets following data clearing. Data sets were excluded if incomplete, not working in the United Kingdom, or not a qualified physiotherapist. Respondents' demographics are presented in Table 2.

| Section One: vignettes
Vignette 1: Screening of persistent back pain presentations for serious pathology and other differential diagnoses.
Vignette 2 (non-specific back pain or radicular syndrome) and Vignette 3 (axSpA): Recognition of primary diagnosis of persistent back pain case presentations.
Ninety-four percent (n ¼ 50/53) of respondents with an incorrect primary diagnosis for the axSpA vignette misattributed the presentation to non-specific back pain.
There was an association between more accurate answers in the axSpA vignette responses and familiarity with NICE guidance on spondyloarthritis, continuing professional development (CPD) on spondyloarthritis, working for the National Health Service (NHS), receiving GP referrals and higher professional grade. Non-recognition of the axSpA vignette was associated with caseloads of ≤30% back pain patients and ≤3 years musculoskeletal experience (Table 3). Assessment of Spondyloarthritis International Society (ASAS) criteria Sieper, van der Heijde, et al., 2009), ESSG (Dougados et al., 1991) and Berlin criteria for inflammatory back pain (Rudwaleit et al., 2006).

| Confidence in recognising features of suspected axSpA
Correctly identifying the axSpA vignette was associated with higher self-reported confidence (median ¼ 8/10) (using a 1-10-point scale where 1 meant 'not at all confident' and 10 meant 'very confident') in knowledge of clinical features of inflammatory back pain, the extraarticular and peripheral features associated with spondyloarthritis (see Figure 4). However, self-reported confidence was still relatively high in many respondents (59%) who inaccurately diagnosed the axSpA vignette with a median of 7 for knowledge of inflammatory back pain, and a median of 6 for the extra-articular and peripheral feature, although the overall range in self-reported confidence was much wider (see Figure 4).

| Knowledge of features of inflammatory back pain
Only 27% of respondents recognised all features of inflammatory back pain (9/9) based on a combination of ASAS Sieper, van der Heijde, et al., 2009), NICE (2017 and Berlin criteria (Rudwaleit et al., 2006) (see Table 4

T A B L E 3 Association between individual respondent's demographics and their responses to the vignettes
Vignette diagnosis Suspected axial spondyloarthritis (n ¼ 132) The four features of inflammatory back pain highlighted in bold are from the additional referral criteria in the NICE (2017) guidance and were used within the spondyloarthritis vignette questions. STEEN ET AL.

Diagnosis given by respondents Primary diagnosis % (n) Secondary diagnosis % (n) Not recognised % (n)
-9 Higher recognition of the features of inflammatory back pain was associated with familiarity with the NICE (2017) guidance, working in the NHS, prior education on spondyloarthritis and treating GP referred patients.

| DISCUSSION
This study used an online survey using back pain presentation vignettes and a questionnaire to evaluate physiotherapists' awareness, Only 5% of General Practitioners and 9.4% of non-rheumatologists identified all features indicative of inflammatory back pain (Tangrungruengkit et al., 2016). Only 6% of General Practitioners were found to consider all peripheral and extra-articular features of axSpA in their history taking (Jois et al., 2008). A recent UK survey of osteopaths and chiropractors also highlighted a lack of awareness and confidence in aspects of screening for possible axSpA (Yong et al., 2019).
This survey also found that case presentations that were typical of axSpA were commonly misattributed as persistent non-specific back pain. These findings reflect the difficulties that other healthcare professionals have been found to encounter when differentiating the symptoms of inflammatory back pain from non-specific back pain (Jois et al., 2008;Tangrungruengkit et al., 2016). Van Onna, Gorter, van Meerendonk, and van Tubergen (2014) found that 40% of General Practitioners were unfamiliar with inflammatory back pain symptoms and how to differentiate them from symptoms of nonspecific back pain. Furthermore, Seo et al. (2015) found that 59% of axSpA patients had previously been misdiagnosed, of which nonspecific back pain was the diagnosis in 62% of cases.
Misattribution may be partly due to poor awareness of the clinical features of axSpA, along with the common prevalence of nonspecific back pain (90%-95% of back pain presentations) (Danve & Deodhar, 2015;Sieper & Poddubnyy, 2017) and moves to reduce over-investigation and medicalisation of back pain (Foster et al., 2018;NICE, 2016). There has been emphasis on the importance of appropriate screening within clinical history taking that has included inflammatory back pain (Maher et al., 2017;NHS England, 2017;NICE, 2016). The lack of appropriate further subjective screening found in this survey suggests that the questioning required to identify possible axSpA is not core practice in back pain assessments. Awareness of this element of assessments recommended in NICE (2016) guidance on back and radicular pain, UK National Back and Radicular Pain pathway (NHS England, 2017), and NICE (2017) guidance on spondyloarthritis has not adequately filtered into musculoskeletal practice and highlights the importance of an awareness campaign on screening in back pain assessments and how to question and recognise features of suspected axSpA.
The results of this survey suggest there is a lack of awareness of when to refer to rheumatology. Many respondents who cited axSpA as a secondary diagnosis in the suspected axSpA vignette inappropriately chose physiotherapy treatment rather than onward referral in accordance with referral guidance (NICE, 2017). This finding is significant given the diagnostic delays and importance of early intervention (Danve & Deodhar, 2015;Strand & Singh, 2017). An awareness raising campaign on recognition and referral of axSpA is supported by the finding that better recognition and appropriate referral were associated with respondents' familiarity with NICE  (2017), suggests that this confidence is misplaced. There was some awareness shown of previously published referral strategies developed by ASAS Sieper, van der Heijde, et al., 2009), the ESSG (Dougados et al., 1991) and Berlin criteria (Rudwaleit et al., 2006).
However, the generally limited awareness of all these referral strategies indicates their lack of penetration into the physiotherapy profession. This is unsurprising since there has been a paucity of journal articles on axSpA published in core physiotherapy or musculoskeletal health profession literature (McCrum, 2019). Survey analysis found that prolonged morning stiffness as a symptom suspicious of inflammatory disease is strongly embedded in physiotherapy screening practice.
Inflammatory back pain is considered the most recognisable symptom of axSpA in rheumatology literature (Sieper et al., 2009a) and respondents showed most confidence with inflammatory back pain signs and symptoms as opposed to other associated features of spondyloarthritis. This compares with a similar study in GP (Jois et al., 2008). In the current survey, most respondents identified at least These features also lacked mention in screening responses, and likely reflect similar missed screening in clinical practice. Since the presence of these features raises index of suspicion of axSpA, screening in back pain assessments is paramount (Danve & Deodhar, 2015;NICE, 2017). As key musculoskeletal professions in back pain assessment pathways, it is vital that physiotherapists should be skilled in when to suspect axSpA.
Respondents also attributed high importance to pathology investigations in suspecting axSpA, including elevated CRP and ESR and HLA B27 positivity. However, raised inflammatory markers, have low sensitivity and specificity (Almodóvar et al., 2014) and present in only 40%-50% of people with axSpA (Rudwaleit, Landewé, et al., 2009). The high importance given to HLA-B27 positivity may indicate a lack of understanding in the role of risk factors in the diagnosis of axSpA. Although a known risk factor for spondyloarthritis, HLA-B27 positivity has a low specificity (Almodóvar et al., 2014) and is present in the general population, with 8% positivity in Europeans (Sieper & Poddubnyy, 2017). NICE (2017)

| Limitations
Several factors need consideration when interpreting the findings of this study including response bias and the convenience and self-selected sample that is low compared to the numbers of practicing physiotherapists assessing back pain presentations in the United Kingdom. A response rate of 132 usable results is not expected to be representative of all UK musculoskeletal physiotherapists. Respondents also tended to have more specialised musculoskeletal experience which may be explained through the targeted advertising of the survey. Respondents were predominantly female, based in the NHS and England, and with many at a senior level (band 6 as per Agenda for Change) which is representative of the workforce (Chartered Society of Physiotherapists, 2017). Only 11% (14/132) of respondents had a specialist interest in Rheumatology and so responses provided a sample that is reflecting the breadth of expertise within clinical musculoskeletal practice.
Regardless of the limitations of the survey, results strongly indicate that more emphasis must be put on raising awareness of axSpA and its associated features and screening as part of routine clinical practice, thus ensuring timely specialist referral (NICE, 2017).

ACKNOWLEDGEMENTS
The University of Hertfordshire for supporting ES undertaking this research. The National Axial Spondyloarthritis Society (NASS) for funding support and aiding survey dessemintion. All the expert reviewers who provided feedback during the survey construction and all those who participated in the online survey.

CONFLICT OF INTEREST
The authors declare that they have no competing interests.

AUTHOR CONTRIBUTION
ES was responsible for the conception and study design, literature review, data collection and analysis, and interpretation of the data and gaining ethical approval and drafting, submitting, and revising the STEEN ET AL.

ETHICS STATEMENT
Ethical approval was granted from the University of Hertfordshire, Health and Human Sciences Ethics Committee (HSK/PGT/UH/03202).

DATA AVAILABILITY STATEMENT
The datasets used and/or analysed during the current study are available from the corresponding authors on reasonable request.