Adherence to key domains in low back pain guidelines: A cross‐sectional study of Danish physiotherapists

Abstract Objective The aim of the study was threefold: Firstly, to investigate the adherence to clinical practice guidelines for low back pain (LBP) among Danish physiotherapists with regard to three key domains: (a) activity, (b) work and (c) psychosocial risk factors. Secondly, to investigate whether adherence differed between physiotherapists working in private clinics (private physiotherapists) and physiotherapists working at public healthcare centres (public physiotherapists). Thirdly, to describe the physiotherapists' treatment modalities for patients with LBP. Methods A cross‐sectional online survey was conducted with 817 physiotherapists working in the Central Denmark Region. Adherence to the guideline domains was assessed using two vignettes. The difference in adherence between the groups was assessed using the Chi‐squared test. Treatment modalities were reported using descriptive statistics. Results A total of 234 physiotherapists responded, hereof 163 private physiotherapists and 71 public physiotherapists (response rate 29%). The proportions of physiotherapists managing the patients strictly in line with the guideline domains were 32% (activity), 16% (work) and 82% (psychosocial risk factors) for Vignette 1 and 6% (activity), 53% (work) and 60% (psychosocial risk factors) for Vignette 2. Public physiotherapists were more likely to manage patients strictly in line with guidelines for assessing the psychosocial risk factors compared to private physiotherapist (Vignette 1: 92% vs. 77% p = .030; Vignette 2:70% vs. 55% p = .035). Regarding the other two domains, there was no significant difference between the two groups in terms of adherence (p > .05). Concerning treatment modalities, the majority of physiotherapists instructed the patients in adopting an exercise program or informed the patients about the benign nature and prognosis of LBP. Conclusion Overall, the participating Danish physiotherapists strictly adhered to only one out of three key domains. This underlines the importance of bringing focus on implementing the current guidelines' recommendations in clinical practice.


| INTRODUCTION
Low back pain (LBP) is of great importance, as it is the world's leading cause of years lived with disability (Hoy et al., 2014;James et al., 2018). In Denmark, patients with LBP comprise 30% of all visits to the physiotherapist (PT) or chiropractor and 20% of all sick leave days is caused by LBP (Danish Health Authority, 2015b).
Although Clinical Practice Guidelines (CPGs) for LBP have existed in Denmark since 2012, patients continue to report incoherent care pathways with wide variation in the information and treatment they receive (Danish Health Authority, 2015a). This poses a problem, as studies have shown that adherence to CPGs can reduce inappropriate variations in the healthcare system and optimize the use of evidencebased management. Moreover, CPG adherences is more costeffective, due to a reduction in the number of consultations and less disability (Fritz, Cleland, & Brennan, 2007;Grimshaw, Eccles, & Russell, 1995;Hanney, Masaracchio, Liu, & Kolber, 2016; G. M. Rutten et al., 2010). Nevertheless, a wide variation in CPG adherence and treatment modalities in clinical practice has been documented repeatedly (de Souza, Ladeira, & Costa, 2017;Hendrick, Mani, Bishop, Milosavljevic, & Schneiders, 2013;Keating et al., 2016;Ladeira, Samuel Cheng, & Hill, 2015).
Since the first CPG for LBP management was described by the Quebeck Task Force in 1987 (Spitzer, Le Blanc, & Dupuis, 1987), a number of CPGs have been published internationally (Koes et al., 2010;Oliveira et al., 2018;Pillastrini et al., 2012) and nationally in Denmark (Danish Health Authority, 2013, 2016b, 2016a. Across these CPGs, there is agreement that evidence-based care for patients with LBP (in the absence of severe spinal pathology or other serious disease) should include information about the benign nature and prognosis of LBP, advice to stay physically active (despite pain), and advice to stay at work or return to work as soon as possible (despite pain) (Danish Health Authority, 2016a, 2016bKoes et al., 2010;Oliveira et al., 2018;Pillastrini et al., 2012). In addition, CPGs agree that certain psychosocial factors entail a risk of developing a chronic condition and therefore should be identified and addressed to succeed in LBP management (Danish Health Authority, 2013;Koes et al., 2010;Lin et al., 2020;Oliveira et al., 2018;Pillastrini et al., 2012).
In a Danish context, management of LBP by PTs has been investigated in only one study, published in 2003 (Hamm et al., 2003). The study found that PTs often used at least one treatment recommendation according to the Health Technology Assessment report 'Low Back Pain' (Danish Centre for Evaluation and Health Technology Assessment., 1999), but in addition, used passive treatments such as massage, heat and cold treatment and electrotherapy, which were recommended against by the report (Hamm et al., 2003). CPGs have been published in the field of LBP management since then, but no recent studies have been conducted investigating the current level of CPG adherence and treatment modalities used by Danish PTs.

| Data collection
The data were collected via an electronic survey sent by email to all private clinics working within health insurance and public healthcare centres in the Central Denmark Region. The survey was designed using Research Electronic Data Capture (REDCap) hosted at the Department of Clinical Medicine at Aarhus University (Harris et al., 2009). The email consisted of an invitation with information about the study and a link to access the electronic survey. The emails were addressed to either the managers of the public healthcare centres or the secretaries or owners of the private clinics, who were then asked to forward the email to all eligible PTs at each location. Electronic reminders and telephone calls were made after 1, 3 and 6 weeks.

| Survey instrument
The questionnaire was designed specifically for the current study and was developed and pilot-tested in the autumn of 2017, in collaboration with researchers and clinicians at the Regional Spine Clinic, located at Silkeborg Regional Hospital, as well as with private and public PTs.
The first section of the survey included a series of questions designed to describe the demographic characteristics of the participating PTs (e.g., age, sex and years of practice experience). The last section of the survey included two vignettes, described below.

| Vignettes
The vignettes are written as hypothetical patient scenarios, which are considered valid instruments for measuring the clinicians' adherence to CPGs (G. M. J. Rutten, Harting, Rutten, Bekkering, & Kremers, 2006), and have previously been used in similar studies (Derghazarian & Simmonds, 2011;Hendrick et al., 2013;Keating et al., 2016;Learman, Ellis, Goode, Showalter, & Cook, 2014). In this study, all PTs were presented with two vignettes (referred to as Vignette 1 and Vignette 2). The descriptions of both vignettes were developed for this study based on real patients and were subsequently validated by five clinicians with extensive expertise in the field. Vignette 1 was designed to reflect a patient with psychosocial risk factors and reduced activity level. It represented a 35-year-old woman with severe LBP during the previous 12-13 weeks. She had been on sick leave since the LBP started. The MRI scan showed age-related degenerative changes. Vignette 2 represented a 23-year-old man, with no sign of psychosocial or work/activity related risk factors. He had LBP with radiculopathy to the right leg for approximately 3 months. The MRI scan showed a L5-S1 disc prolapse with signs of nerve root pressure. Neither vignettes reported red flags (signs of severe spinal pathology or other serious disease). The two vignettes are described in detail in Table 1.
The questions, response options and classifications of the responses regarding work and activity were based on a previous study by Bishop, Foster, Thomas, and Hay (2008). The study by Bishop et al. did not investigate the domain of psychosocial risk factors, but as this is now considered essential in the management of LBP in the recent CPGs (Lin et al., 2020), the psychosocial domain was included in the current study.

| Measuring adherence
The PTs' adherence to the CPGs was assessed on the basis of the advice they would give to the patients in the two vignettes on the three key guideline domains; (a) activity, (b) work and (c) psychosocial risk factors. Thus, for each vignette, there was one question for each guideline domain, with four possible answers for each question. For Description of the two vignettes

Patient 1
A 35-year-old woman is referred from the secondary sector following an episode of severe LBP, which has lasted for 12-13 weeks. She is married and has a 4-year-old child In the past few years, she has not had the energy to be physically active. She has been on sick leave from her job as a healthcare assistant since the episode started This is the third and worst episode of LBP she has experienced. In the two previous episodes, the pain has resolved spontaneously. The pain is currently reduced to approximately 50% of its worst intensity during this episode. The pain does not disturb her sleep. She is currently taking paracetamol and NSAID when needed She is very concerned about the intensity of the pain and she is nervous that her back problems will not resolve this time. The patient feels she still needs to rest her back once in a while. She is afraid of exacerbating the pain again, in case she has to lift something from an awkward position Extension and lateral flexion is moderately reduced, while flexion is limited only minimally. The neurological examination is normal. The patient experiences some sensory disturbances in the lower right leg, but the medical examination shows normal reflexes and no strength reduction. The MRI scan shows age-related degenerative changes

Patient 2
A 23-year-old man is referred from the secondary sector following an episode of LBP with radiculopathy to the right leg. The pain started about 3 months ago without prior trauma. Various physiotherapeutic treatments have been attempted without considerable effect, for example, McKenzie exercises and manipulation The patient has paused football and other sports activities, but has resumed running again. He still cannot manage playing football, which bothers him a lot. He has remained at work, which consists primarily of office work. He now plans to study at the business school Initially the pain slowly worsened and he became increasingly disabled due to pronounced pain on the backside of the right leg down to knee, periodically to heel level. He is currently only taking painkillers prior to physical activities He has normal range of motion in the back although slightly reduced lateral flexion to the right, which causes known pain in the right buttock. Walking on heels and toes as well as one-leg-stand is normal and all reflexes can be triggered. There is normal strength and sensibility in the hip, knee, ankle and toes. MRI scan shows L5-S1 prolapse on the right side with signs of nerve root pressure  Table 2.

| The PTs' treatment modalities
In addition to answering the questions regarding advice on each of the three key domains, the PTs were asked to select which treatment modalities they would initiate in each vignette. A maximum of three treatment modalities out of 12 could be selected (see Appendix 1).
The 12 treatment modalities were determined by the research team and validated by private and public PTs who, in the pilot test, were asked to add more treatment modalities if needed to make the responses as valid as possible to their daily practice. The treatment modalities were not included in the assessment of adherence to the three-guideline domains, but were used for descriptive purposes only. 3 | RESULTS

| Participant flow
A total of 255 out of 817 invited PTs answered the electronic survey.
Twenty-one of the responses were removed, either due to blank responses (PTs who only opened the survey but did not respond, thereby being registered in REDCap as missing), or due to double responses (as it was not a unique link for the online survey, PTs who unintentionally answered the survey twice would appear as two different respondents in REDCap). Furthermore, several PTs did not complete the survey, resulting in a lower N in parts of the analysis. A total of 234 PTs were included in the study (response rate 29%). Participant flow is shown in Figure 1.
The characteristics of the participating PTs can be seen in Table 3.
Overall, most were female (62%), and had more than 10 years of clinical experience (58%). In terms of differences between public and private PTs, there was more female in public PTs (80%) compared to private PTs (55%; p < .001). More public PTs (91%) allocated 60 min or more for the first consultation compared to private PTs (21%; p < .001). Private PTs treated more patients with LBP per week (79% treated six patients or more) compared to public PTs (59% treated six patients or more; p = .009).

| The PTs' adherence to the three key guideline domains
For Vignette 1, the proportion of PTs giving advice that was strictly in line with CPGs for the three domains was 32% (activity), 16% (work) and 82% (psychosocial factors). For Vignette 2, the corresponding results were 6% (activity), 53% (work) and 60% (psychosocial factors; Table 4).
For the two domains 'activity' and 'work', there was no statistically significant difference between the adherence of private and public PTs, respectively. For the domain 'psychosocial risk factors' there was a divergence in adherence between the two groups in both Vignette 1 (92% vs. 77%, p = .03) and Vignette 2 (70% vs. 55%, p = .04) in favour of the public PTs, who adhered more strictly to the guideline domain.

| Post hoc analyses and results
Due to significant difference in terms of sex, the number of patients with LBP per week and the time allocated for first treatment between the private and public PTs, respectively, an exploratory post hoc To what extent do you assess the patient's psychological and social resources during your examination: 1. In the history-taking, I ask about the patient's psychosocial condition/or use a questionnaire to screen for risk factors 2. I try to be aware of it, but I only have time to do a physical examination 3. I primarily relate to the possible tissue damage that underlies the patient's pain 4. I do not think I have the skills to assess the patient's psychosocial factors Strictly in line Partly in line Not in line Not in line analysis was performed using a chi-squared test, to investigate whether guideline adherence was influenced by these factors.
In terms of adherence to the three-guideline domains, no statistically significant difference was found between the subgroups, except for the domain 'psychosocial risk factors'. PTs who had allocated 60 min or more for the first consultation had a higher level of adherence compared with those with less than 60 min allocated (Vignette 1: p = .001; Vignette 2: p < .001; data not shown).

| PTs' treatment modalities
The two most common treatment modalities were to instruct the patients in an exercise program (Vignette 1: 62%, Vignette 2: 70%) and to inform the patients about the benign nature and prognosis of LBP (Vignette 1: 66%, Vignette 2: 54%). The third most common treatment modality differed between the two vignettes. In Vignette 1, the PTs would examine the patient's worried thoughts about their LBP and in Vignette 2 they would teach the patient a better posture and provide ergonomic advice (data not shown).

| DISCUSSION
The current study aimed to investigate the adherence by Danish PTs working in private clinics and public healthcare centres to CPGs for the management of LBP, specifically addressing the following three key guideline domains: (a) activity, (b) work and (c) psychosocial risk F I G U R E 1 Flow chart of the PTs in the study. PT, physiotherapist factors, and whether the adherence differed between private and public PTs. Additionally, the aim was to describe the PTs' treatment modalities for patients with LBP.
We found that the vast majority of the participating PTs partly adhered to two out of three domains and strictly adhered to only one out of the three-guideline domains. More than 95% of the PTs would give advice regarding activity and work that was partly or strictly in line with the CPGs. In contrast, up to 23% did not adhere to the CPGs for assessing psychosocial risk factors in Vignette 2.This domain was also the only domain where there was a significant divergence in adherence between private and public PTs. Significantly more public PTs strictly adhered to CPGs for this domain. Most of the participating PTs also provided treatments recommended in the CPGs (information, exercises, identification of psychosocial issues and correction of working posture).
Assessment of guideline adherence is a complex matter, and although comparison of results across countries reveals great variation in guideline adherence, there are a number of possible reasons for the differences seen. Studies report difference in methods for measuring adherence (vignettes, bills of accounts, treatment options, etc.), difference in the definition of adherence, difference in the patient described in the vignette (symptom complaints, personal characteristics) as well as difference in the key domains measured (activity, work, reference to specialist and X-ray, medicine, bed rest etc.; Casserley-Feeney, Bury, Daly, & Hurley, 2008;Fritz et al., 2007;Hendrick et al., 2013;Keating et al., 2016;Ladeira et al., 2015;Learman et al., 2014).
These variations of assessing guideline adherence may explain the differences in results between studies described below.

| Key guideline domain: Activity
In the current study, none of the vignettes included patients with symptoms suggesting serious spinal pathology. Still, the PTs were cautious about advising the patients to 'perform usual activities'. The recommendation in the CPGs regarding activity may be considered equivocal, because it states that advising patients to 'stay active within pain tolerance' might induce fear-avoidance beliefs, thereby contributing to a reduced level of activity (Danish Health Authority, 2016a). On the other hand, the CPGs emphasize that patients should be encouraged to gradually increase their level of activity. PTs may therefore be confused about which advice they should give their patients (perform usual activities or perform activities within pain tolerance).
The results are comparable to those reported in studies by Hendrick et al. and Keating et al., showing the majority of PTs provided advice that was strictly or partly in line with the CPGs. This is in opposition to Souza et al., who showed that less than 20% of the PTs advised their patients to pursue or maintain an active lifestyle despite similar methods for measuring adherence (de Souza et al., 2017;Hendrick et al., 2013;Keating et al., 2016). Despite the Danish PTs advising patients to remain active, there is still room for improvement.

| Key guideline domain: Work
The PTs' advice regarding work differed between the two vignettes. leave from her work as a healthcare assistant) to 'return to part time or light duties' at work. It could be considered whether this advice should have been categorized as strictly in line as the CPGs recommend that giving advice regarding temporary changes in work functions or working hours is also an option in line with best evidence.
However, due to the fact that the CPGs emphasize that staying at work most often is beneficial for maintaining function, and since the patient did not show any signs of red flags (serious spinal pathology) or radiculopathy, 'return to normal work' was therefore considered to be strictly in line with CPGs. The majority of the PTs would advise the patient in Vignette 2 (not on sick leave from his office work) to stay at work full-time, but 45% still would advise him to reduce his work hours. The results for the work domain in our study are comparable to those from the study by Hendrick et al., who showed that the majority of the PTs in New Zealand gave advice regarding work that was partly in line with the guidelines (Hendrick et al., 2013). However, it is not clear why the PTs in the current study were cautious in their advice regarding work in Vignette 2. The result shows that there still is room for improvement in terms of higher adherence to CPGs by the Danish PTs on the work domain.

| Key guideline domain: Psychosocial risk factor
With regard to the psychosocial domain, the PTs adhered differently in the two vignettes, with 23% of public PTs reporting that they  2017 ;Foster & Delitto, 2011;Synnott et al., 2015). The result confirms the importance of improving the implementation of assessment and management of psychosocial risk factors in PT practice.

| Private versus public PTs
In the current study, the main difference found between private and

| Treatment modalities
The majority of the participating PTs provided treatments which are recommended in the CPGs. Only the third most common treatment modality in Vignette 2 (posture and ergonomic advice) is not mentioned in any of the CPGs. This treatment modality might be explained by PTs having a biomedical focus (Nijs, Roussel, Paul van Wilgen, Köke, & Smeets, 2013 Hamm et al., 2003). A study undertaken by Bernhardsson et al. also found that less than 5% of Swedish PTs reported use of electrotherapy as a treatment modality (Bernhardsson, Öberg, Johansson, Nilsen, & Larsson, 2015). The decrease in the use of electrotherapy may be due to recent findings demonstrating a lack of evidence for its effect.

| Strengths and limitations
To the best of our knowledge, this study is the first Danish cross-  Hendrick et al., 2013;Keating et al., 2016;Ladeira et al., 2015).
Another limitation of this study is the lack of consensus on the definition of adherence as well as lack of standardized instruments for measuring adherence, indicating the complexity of the phenomenon.
In the current study, the classification of responses for the two domains of activity and work were based on a previous study by Bishop et al. (2008), allowing for comparability of the results.
Although vignettes are generally accepted as a valid measure of adherence, one study has shown a poor correlation between the PTs' response to the vignettes and what blinded patients described, to have been advised about activity and work after visiting the PT (Brunner, Probst, Meichtry, Luomajoki, & Dankaerts, 2016). A more valid instrument would be desirable. However, due to lack of time and resources, vignettes were considered the best option.

| CONCLUSION
Overall, the participating Danish PTs partly adhered to two out of three-guideline domains, and would offer treatment modalities in line with CPGs. However, strict adherence was seen in only one out of the three key domains. The majority of the PTs assessed the patients' psychosocial risk factors, but did not consider screening for psychosocial risk factors as a standard part of the assessment. Also, the PTs were cautious to advise the patients to return to usual activities and work. Public PTs were more likely to be strictly in line with the CPGs for assessing the psychosocial risk factors.
this study. We would also like to thank Anne Grøndahl Poulsen for her involvement in the study; we appreciate the professional discussions.

AUTHOR CONTRIBUTIONS
All authors have contributed to the conception, design, analyses and interpretation of the data and been involved in drafting or revising the manuscript.

ETHICAL APPROVAL
The study is registered in ClinicalTrials.gov (Reg. No. NCT03426410).
Under the Danish Act concerning Research Ethics Review of Health Research Projects, health research projects using surveys and questionnaires do not need ethics approval (http://en.nvk.dk/rules-andguidelines/act-on-research-ethics-review-of-health-research-projects).
The data collection was approved by the Danish Data Protection Agency  and consent from the physiotherapists was assumed if they responded to the survey.

IMPLICATIONS FOR PHYSIOTHERAPY PRACTICE
This study shows that there is room for improvement concerning PTs' adherence with CPGs in the field of LBP. It is therefore of great importance to focus on enhancing the implementation of the existing CPG recommendations for managing patients with LBP. Time allocated to the first treatment can be of importance in order to assess and address the patients' psychosocial risk factors. However, this needs to be supported by future research.