Aim and objectives

The aim and objectives of the feasibility trial are detailed in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Aim and objectives.

https://doi.org/10.1371/journal.pone.0229792.t001

Ethics approval and consent to participate

To ensure that the trial was conducted in an ethical manner within best research practice, ethical approval obtained on the 30^th October 2018 (IRAS project ID: 250734, Protocol number: RG_18–101, REC reference: 18/LO/1793) and HRA approval obtained, with R&D obtained from all sites [43, 44].

Design

It is proposed that a SWcRCT design will be used to evaluate the clinical and cost-effectiveness of physiotherapist prescribing for LBP in the future. As the use of physiotherapist independent prescribing and the utilisation of FCP roles are new innovations in primary care, there are a limited number of APPs working as FCPs that are currently registered independent prescribers. Use of a stepped-wedge design allows APPs to cross from the control group to the experimental group once they are registered independent prescribers and start to utilise physiotherapists independent prescribing in their practice. This transition facilitates a robust and timely evaluation, which is reflective of current clinical practice. A full explanation of the use of the SWcRCT design for a definitive trial is detailed in the published protocol [37].

No existing framework describes best practice for completing feasibility trials in preparation for SWcRCTs [45]. Two-arm feasibility studies aiming to calculate intra-cluster correlation coefficients (ICCs) required for sample size calculations have been shown to exhibit insufficient accuracy [45]. Therefore, a prospective, mixed-methods, single-arm feasibility trial, exclusive of sample size estimation was employed to evaluate the trial objectives on the experimental arm of the future SWcRCT [34, 44, 46].

The mixed-methods approach comprised two component phases [47–49]:

1. Trial Component: a quantitative one-arm feasibility trial
2. Embedded Qualitative Component: qualitative semi-structured interviews and patient focus groups, using thematic analysis

Trial component

A single-arm feasibility trial design was used to evaluate the trial objectives [36, 46]. Patient reported outcome measures (S2 File) were completed digitally via an online survey at initial assessment (baseline), 6 and 12 weeks following recruitment, to enable the evaluation of the data collection tool and the feasibility of follow-up data collection (date range for participant recruitment and final follow-up: Rural town 3^rd December 2018-11^th January 2019, final follow-up 15^th April 2019; Regional city 28^th November- 19^th December 2018, final follow-up 13^th March 2019; Capital city 28^th February-18^th July 2019, final follow-up 10^th October 2019) [44, 50]. Follow up time points were selected based on prognostic literature demonstrating the ‘normal’ resolution time of LBP [51–53]. To allow real-time data capture and storage, the online outcome measures survey was built using REDCap (Research Electronic Data Capture) software hosted at the Centre for Precision Rehabilitation for Spinal Pain (CPR Spine) at the University of Birmingham, UK [54]. The number of participants that declined to participate as they were unable to complete the outcome measures survey online were collated to evaluate the suitability of only using digital data collection in a full trial.

Baseline measurements were collected in the clinical setting. At 6 and 12 weeks a link to the online survey was emailed to participants for completion. A reminder email was sent 24hrs and 48hrs later to facilitate compliance if a participant failed to complete the survey on the required day [44, 55]. To evaluate the feasibility of fitting participants with accelerometers in clinic, ease of data collection and participant compliance with wearing the accelerometer device for a 7 day period, the ten participants recruited at the rural town site had an accelerometer fitted to their left thigh for 7 days following the first consultation [34, 35]. Stamped/addressed envelopes were provided to enable return of the devices after use.

Participants

The STarT Back Tool was used at initial assessment by the APPs to identify patients stratified into the medium risk LBP group [5]. This group of patients have been acknowledged as the prevalent group presenting for management of LBP in primary care; exhibiting both physical and psychosocial prognostic factors and potentially requiring physiotherapist prescribing to optimise their treatment outcomes [5, 56–58]. Patients in this group were eligible for recruitment if they met the inclusion criteria detailed in Table 2. Convenience sampling was employed as feasibility trials demand fluid recruitment and follow up with good participant retention [35, 44, 50, 59]. Convenience sampling was used as this method has the advantages of fluid recruitment. To minimise selection bias, patients fitting the eligibility criteria were recruited consecutively [60]. Patients interested in participating were provided with a participant information sheet (S3 File) explaining the rationale, content and research dissemination plans. Inclusion within the trial was entirely voluntary, with no incentives offered to participants to minimise bias [44, 50]. The APP answered patient queries and contact details for the research team were provided if the APP was unable to answer specific questions. Consent was obtained from willing participants using an online consent form (S4 File). Participants were free to withdraw at any time, without any impact on their care [44, 50].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Participant inclusion criteria [37].

https://doi.org/10.1371/journal.pone.0229792.t002

Interventions

The experimental arm of the definitive planned trial was used to evaluate the feasibility trial objectives [33–36]. An APP completed the initial assessment and management of participants in line with evidence-based practice. If medicines advice or prescription drugs were required/no longer required, these were prescribed/de-prescribed by the APP immediately.

Outcomes

Outcome measures selected for use within the trial were informed by the literature and a team of subject-experts (including physiotherapists, pharmacists, medical practitioners, academics and health-service managers) and deemed most appropriate to evaluate the trial’s objectives whilst attempting to minimise the burden on participants [28]. Detail of the primary and secondary outcome measures are detailed in Table 3. Assessment of the quality of sleep via accelerometer data was detailed in the published protocol [37]. Unfortunately, the devices available for testing the feasibility of fitting the accelerometers in FCP clinics and evaluating participant’s tolerance of wearing the devises, were not validated for measuring quality of sleep. Previous quality research has established that it is feasible to evaluate time sleeping using the data collected by the devises when cross referenced with a participant diary mapping activity, sedentary time and time asleep [32, 65]. It was not deemed necessary to re-evaluate this process. This was the only deviation from the feasibility trial protocol.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Secondary outcome measures and their rationale [37].

https://doi.org/10.1371/journal.pone.0229792.t003

Sample size

Three APPs, across 3 primary care sites representative of English geography, recruited up to n = 10 participants each within a 6 month recruitment period. This enabled evaluation of recruitment rates across clinicians and the feasibility of the trial methods in both metropolitan and rural healthcare services [34, 45, 46]. A sample size of n>20 is regarded as adequate within the literature, when testing feasibility objectives for cRCTs, however a total sample of n = 30 participants was planned to allow for under-recruitment within the specified time period and loss to follow up [34, 35, 45, 46].

Data analysis

Participant flow and loss to follow up was described using a CONSORT diagram to evaluate the feasibility of eligibility criteria and acceptability of recruitment and follow up rates [38]. Data from fully completed outcome questionnaires were included in the data analysis. Data were tabulated, and primary descriptive analysis was completed to test procedure [34, 44, 50]. Effectiveness was not statistically analysed as this was not within the scope of the feasibility trial [44, 50]. As a definitive trial will aim to evaluate the clinical effectiveness of independent prescribing by APPs for patients with LBP in primary care, data distribution of the primary outcome measures across participants, was evaluated at baseline, 6 and 12 weeks, with 12 week data used to measure for a potential floor effect [80].

Embedded qualitative component

Analysis and findings

A grounded theory theoretical framework enabled a thematic analytical approach to analyse and synthesise the qualitative data. This method enables identification of the important thoughts and views of the population being studied, providing explanations alluding to how the concerns may be resolved or processed in preparation for a full trial [44, 87, 88]. Transcripts were coded line-by-line using NVivo 11 software (QSR International, Melbourne, Australia) by one researcher (TN) and verified by a second researcher (AR) [50, 88, 89]. Rigorous comparative analysis was completed to identify similarities and differences within the data, informing the development of descriptive categories which were linked, merged or split to synthesise a conceptual understanding of the data [88, 89]. To avoid single researcher bias, the second researcher (AR) re-interrogated the data to validate or contradict findings. Following this process, to ensure trustworthiness, outcomes were discussed with a panel of experts for confirmation and agreement [87, 88, 90].

Data storage

All data were stored in password protected computer files that could be accessed only by trial investigators at the University of Birmingham. The password-protected files will be retained for 10 years satisfying university code of practice.

Integration: Feasibility, suitability and acceptability

Following quantitative and qualitative analysis, data were assessed against a priori defined success criterion developed by experts and informed by the methodological literature [44, 59, 91]. Success criteria can be found in S7 File. Trial objectives were considered successful if the success criteria were satisfied following the integration of the quantitative and qualitative findings [35, 91].

Patient and public involvement (PPI)

Patients with LBP were part of the research team to ensure the patient perspective was central to planning and decision making. There was a PPI representative on both the trial management group and trial steering group to ensure that patients and the public were involved at all steps of the research process. Patients were involved in the development of the participant information sheet, consent form and questions used in the semi-structured interviews and focus group.

Trial component

Demographics and participant flow.

Demographic and recruitment data are presented in Table 4. n = 29 participants (n = 12 male, n = 17 female) were recruited. The mean recruitment rate was 1.07 participants/week. Two sites recruited the pre-defined n = 10 participants within the 6 month recruitment period (3 and 4.5 weeks). The capital city site recruited n = 9 participants over the 6 month period. Successful loss to follow up was defined a priori as <20%. 48% of participants were lost to follow up at 6 weeks, with 65.5% at 12 weeks (Fig 1). One site had a loss to follow up of 89%, suggestive of site-specific issues. No patients refused to participate owing to the inability to complete the outcomes measure survey online.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. CONSORT flow diagram [38–40].

https://doi.org/10.1371/journal.pone.0229792.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Demographic and recruitment data.

https://doi.org/10.1371/journal.pone.0229792.t004

Outcome measures survey

Table 5 presents mean primary and secondary outcome measure data collected from the outcome measure questionnaire with variability reported by the use of standard deviations (SD). Reductions in pain were found for all pain categories as time progressed. Mean scores on the RMDQ reduced from 9.21 (SD 5.58) at base line to 8.07 (SD 5.82) at six weeks, then increased to 9.70 (SD 5.33) at 12 weeks. Between baseline and 12week, improvements were seen across all components of the secondary outcome measures other than anxiety and depression (EQ-5D 5L) which increased with time. No participants scored the distinct lower limit in any of the outcome measures; therefore, no floor effects were found. As primary and secondary outcomes improved, absence from work and prescription utilisation reduced.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Primary and secondary outcome measures data.

https://doi.org/10.1371/journal.pone.0229792.t005

Accelerometery

Ten participants (n = 2 male, n = 8 female) wore an ActivPal accelerometer 24 hours a day for seven days. Data collected by the accelerometers are displayed in Table 6. There were no missing data. Participants spent an average of 18.57hrs (SD = 1.54) sitting per day, 4.14hrs (SD = 1.17) standing and 1.3hrs (SD = 0.39) walking. Participants completed 5884.66 steps per day (SD = 2255.11), with a mean activity score of 32. 94MET.h (SD = 1.03).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Accelerometer data across 7days.

https://doi.org/10.1371/journal.pone.0229792.t006

Embedded qualitative component

Interviews: Advanced Physiotherapy Practitioners (APPs).

Demographics. Demographic details of the APPs can be found in Table 7. APPs all had a post-graduate qualification in the musculoskeletal specialty. The number of years qualified as a physiotherapist ranged from 15–28 (mean = 21 years).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Demographic data of focus group participants.

https://doi.org/10.1371/journal.pone.0229792.t007

The APPs’ (n = 3) views, perceptions and experiences related specifically to the trial objectives were analysed and synthesised into three themes and associated subthemes:

1. Trial design, conduct and processes
   1. Eligibility criteria, recruitment strategy and follow up procedures
   2. Capacity (time and effort required)
   3. Training requirements
2. Data Collection, outcomes and measures
   1. Data collection tool and accelerometery
   2. Equipment, services and infrastructure
3. Adequacy of the feasibility trial

1. Trial design, conduct and processes

1. Eligibility criteria, recruitment strategy and follow up procedures
   APPs felt that a full trial should enable the evaluation of all stratifications of NSLBP +/- leg pain. They advocated for a full trial that include participants across all LBP stratification groups defined by the STarT back tool. This would allow for both intra- and inter-group comparisons evaluating the effectiveness of independent prescribing by APPs for patients with LBP in primary care across the spectrum. The APPs advised that variation would exist regarding the utilisation of skills within the scope of independent prescribing across each of the STarT back stratification groups due to patient-specific factors such as the length of time the participant has suffered with LBP, associated psychosocial factors, previous treatments and experiences and related medical history. Therefore, it was felt that the specific prescribing skills used, such as prescribing medicines, de-prescribing, advising about the use of over-the-counter medication and medicines management, should be captured throughout the individual patient journey. All APPs expressed that the trial methods tested by the feasibility trial would enable appropriate comparison of STarT back stratification groups, however they highlighted that the complex nature of persistent LBP would require follow-up for longer than the 12-week period tested, demonstrating the need to update the primary end point for a full trial.
   All the APPs felt that the recruitment strategy was suitable, although recognised that recruitment rates could potentially differ between APPs in a full trial due to varying experience of clinicians working in FCP roles and proficiency and/or confidence in prescribing medicines. Two of the clinical sites (regional city and rural town) recruited n = 10 participants over 3- and 4.5-week periods respectively. The third site (capital city) used the full six-month recruitment period to recruit n = 9 participants. The APP from the capital city site reported that patients were frequently excluded from participation due to language barriers. Additionally, English-speaking young professionals in the area often declined recruitment, declaring that they were too busy to commit to complete the 6 and 12-week follow-up surveys. Recruitment was curtailed across all sites by patients consistently stating that they did not want to take analgesia for their LBP, preferring to engage in other conservative management strategies such as exercise and manual therapies. This was cited as the key benefit of consulting a physiotherapist for their LBP rather than their GP.
   To optimise recruitment rates in a definitive trial, it was suggested that posters and social media posts might encourage patient interest. Administrative staff could be utilised to highlight potential participants when booking appointments and provide patients with participant information leaflets prior to seeing the APP. APPs across both city-based sites recognised the risk associated with participants using data collection equipment unsupervised, fearing that there was potential for theft or damage. Patient’s literacy was highlighted as a potential barrier to independent completion of the online outcome measures surveys, and thus a barrier to recruitment of a sample representative of the LBP population. All APPs highlighted that due to restricted consultation time in FCP practice, the use of research assistants to recruit and consent participants identified by the APP during consultations, could further optimise recruitment rates. APPs postulated that research assistants at each site, recruiting, consenting and aiding patients with completing the online outcome measures survey, would not only simplify recruitment and follow up procedures but would also minimise risk, especially at locations where multiple APPs are recruiting to a full trial. The APPs advised that to improve compliance with follow-up procedures, participants should be asked to consent to reminder telephone calls from the research team, with the choice of completing the follow-up outcome measures survey online, on paper, independently or face-to-face with a researcher or research assistant. It was proposed that these changes would increase the likelihood that patients would consent to participating in the trial and minimise drop-out due to non-compliance.
2. Capacity (time and effort required)
   Twenty-minute FCP consultations were scheduled at each site. The APPs stated that the recruitment and consent of each participant and completion of the initial outcome measures survey took approximately 10–15 minutes. If the application of an accelerometer was required, this took an additional 5–10 minutes dependant on the “tech savvy nature” of the individual participant. Additional time was also required to answer participants’ questions, confirm data entry and upload data to the REDCap server. All APPs described the time pressures as “stressful”, causing their clinics to run late. Further, all APPs reported “rushing” clinical assessments if they were aware that they needed to recruit a patient to the trial. It was recommended that if APPs are to be used to recruit, consent, aid in the completion of the initial outcome measures survey and apply an accelerometer in a full trial, 30-40-minute consultations would be necessary. Unfortunately, all of the APPs agreed that healthcare service commissioners would not agree to extended appointment times to allow for research activity. Therefore, a full trial would require a research assistant at each site to ensure appropriate clinical and administrative capacity.
3. Training requirements
   All APPs reported that they felt prepared and confident to lead the recruitment of participants and initial data collection processes. Each APP reported acting as principal investigators for previous clinical trials and observed that less experienced clinicians may require additional training in ethical recruitment and consent procedures, documentation and data management. All recruiting clinicians would require training to effectively apply accelerometers to participants and to educate participants on the re-application should dressings become unsealed or cause a skin reaction. For future trials it was recommended that peer support and clinical mentorship would be essential for new prescribers and less experienced APPs. This would ensure best practice and assist in managing variation in clinicians’ confidence to prescribe.

2. Data collection, outcomes and measures

1. Data collection tool and accelerometery
   All of the APPs agreed that the outcome measures within the survey were suitable and encompassed the majority of the multi-faceted dimensions of LBP. Two of the APPs recommended that tools evaluating sleep, lifestyle and confidence with global physical activity could further improve the richness of data collected in a full trial. Although the APPs recognised that the outcome measures within the survey were validated and tested for reliability and reproducibility, it was reported that some of the participants required help to complete all questions in the survey. It was suggested that this variation in understanding was due to differing participant literacy levels.
   The survey logic was reported as user-friendly other than an issue with one question (“Number of days taken off work?”) as this question was not applicable to participants who were part-time, retired or registered as disabled. The APPs all commented that the difficulty in the ability to capture data about the wider use of the independent prescribing qualification was a weakness. The use of prescribing, de-prescribing and medicines advice and management were deemed fundamental to the effective use of NMP within the FCP role. APPs were also interested in capturing data around what, as well as how drugs are currently being prescribed for the management of LBP by APPs in primary care across the country.
   No problems were reported regarding the fitting of accelerometers. Participants were happy to be taught to re-apply accelerometers themselves should the dressings become loose or need replacing. No issues were reported to the APPs regarding self-removal of accelerometers by participants after seven days, and all units were returned in a timely manner without damage. No adverse effects or compliance issues were reported. It was highlighted that all participants lived locally to the clinic and returned the accelerometers in person. If participants lived further away from clinical sites it was deemed acceptable for participants to return accelerometers by post following provision of a padded pre-paid envelope.
2. Equipment, services and infrastructure
   Overall, the APPs agreed that the services and infrastructure tested by the feasibility trial were suitable and would enable the completion of a robust full trial. All of the APPs disclosed technical issues with the tablet computer provided for data collection. One APP stated that the tablet was not sophisticated enough to optimally run the REDCap application. The others described the application as “clunky”, reporting significant issues with the vertical sliding visual analogue scale (VAS) within the EQ-5D 5L tool. The security of the tablet computer was a concern if the outcome measures survey was to be used outside of the consultation room. It was recommended that the tablet computer should be securely attached to a wall in a quiet waiting area to stop theft, however it was accepted that participant privacy and dignity may preclude this solution from being a valid alternative.

3. Adequacy of the feasibility trial

Each of the APPs interviewed stated that the feasibility trial was adequate for assessing the feasibility suitability and acceptability of the trial methods. The clinicians all agreed that iNMP enables better holistic care by FCPs for LBP in primary care, therefore a full trial is essential to evaluate efficacy. The APPs deemed all ethical conduct and trial documentation (including participant information sheet, consent form) acceptable for use in a full trial. The use of digital consent and data collection was seen to be positive, securely storing participant data and enabling the blinding of the analytical research team.

To better inform clinical decision making, the clinicians enquired as to whether collaboration with NHS digital might further streamline data collection. It was felt that connectivity between data collection technology and the participant’s digital clinical notes, might highlight psychosocial factors not identified by clinicians in a time-pressured consultation, further improving management of the ‘whole patient’.

Table 8 provides illustrative quotations to demonstrate each theme.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Interviews, comments that reported or discussed each theme and illustrative quotations from APPs (quotations have been copied verbatim).

https://doi.org/10.1371/journal.pone.0229792.t008

Focus group: Patients

Demographics.

Focus group demographic data is presented in Table 9. Six participants from the trial component of the feasibility trial consented to participate in the focus group. Purposive sampling enabled a representative spread of ages. 66% (n = 4) of the participants were female, 66% (n = 4) of participants completed the feasibility trial, with 33% (n = 2) lost to follow up at 12 weeks.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 9. Demographic data of focus group participants.

https://doi.org/10.1371/journal.pone.0229792.t009

Patients’ views, perceptions and experiences related specifically to the trial objectives were analysed and synthesised. Data were synthesised into three themes:

1. The use of physiotherapist independent prescribing in FCP clinics
2. Trial conduct and processes
   1. Recruitment processes
   2. Follow up processes
3. Data Collection
   1. Accelerometery
   2. Outcome measure survey

1. The use of physiotherapist independent prescribing by APPs

The participants expressed that they were happy with the introduction of physiotherapist independent prescribing and felt confident in the APPs’ skills. All participants agreed that physiotherapists are experts in the management of LBP. APPs were felt to provide a more detailed assessment than GPs, listening to patients’ problems prior to developing a holistic treatment plan alongside the individual patient. All participants agreed that they would prefer to utilise non-pharmacotherapy methods to manage their LBP. If analgesia was required by a patient to facilitate management and rehabilitation, the APP was able to advise why the medication was needed and how best/ when to take the medication within the context of their social and family life, work commitments and associated treatments, including exercise therapy and physical activity. It was recommended by participants that more clinicians in similar roles should be employed across the whole spinal pain pathway, particularly within urgent care centres and emergency departments.

2. Trial conduct and processes

1. Recruitment processes
   The participants reported that they were happy with the ethical conduct throughout the feasibility trial. They deemed the participant information sheet to be satisfactory and suitable for use in a full trial. All participants were happy with the recruitment process and could not identify any adverse effects or risks to being involved in the trial. It was proposed that the use of posters, social media and advertising on waiting room television screens might encourage participation in a full trial. Further, the focus group recommended marketing aimed to involve all NHS patients in research activity. It was thought that this broader marketing strategy would reduce fear of participation and increase public awareness about the social responsibility for participation in health research.
2. Follow-up procedures
   All participants agreed that the follow-up procedure used in the feasibility trial was acceptable however, participants in a full trial would benefit from choice of communication options. Participants agreed that personal preference would vary between patients, some preferring contact via telephone or post rather than email. They requested that each clinical site should have a liaison point for face-to-face discussion if required, with the option to complete the follow-up outcome measure surveys digitally or on paper, over the telephone, via video call or face-to-face with a member of the research team. These options should be offered to the participants during the consent procedure with the participants being able to change their preference during participation in the trial if required. It was posed that the probable reason for non-response to the six week or 12-week follow up survey was that the participants were no longer suffering from LBP. It was felt that the clinicians recruiting participants to a full trial should be explicit about the necessity to complete all follow-up surveys whether LBP had resolved or remained present. It was felt that reminders from a research assistant would further assist compliance.

3. Data collection

1. Accelerometery
   Participants agreed that all aspects of the feasibility trial testing the use of accelerometery suitable and acceptable. They reported that the APP’s explanation regarding the application, reapplication and rationale for use of the accelerometer was clear and understandable. Participants felt confident to reapply the accelerometer with a fresh dressing if required. However, this was not necessary across the trial participants. Three participants reported worrying about re-positioning the devices if re-application was necessary, however were happy that they could seek help from their APP if required. It was recommended that a further patient information sheet should be developed demonstrating the use, application and removal of the accelerometer to prompt a participant’s memories.
   Participants reported that the devices were easily fitted, and none experienced any adverse effects. One participant reported slight skin irritation in the final hours wearing the device but did not feel this was bad enough to warrant a change of dressing. Another participant stated that although wary of her sensitive skin during application, she had no reaction to the device-cover or adhesive dressing. Participants all concurred that the use of accelerometers attached to the skin enabled them to forget that they were wearing the device and continue with normal activity. It was felt that the device did not prompt additional activity after the first 24 hours and were no problem during land- or water-based exercise. Participants reported that attaching the accelerometer to the skin was preferable to wearing a device around the wrist, ankle or on their clothing, as these types of devices might prompt additional physical activity that they would not have otherwise undertaken. It was felt that removable accelerometers might provoke feelings of stress due to a sense of constant examination and worry that results would be skewed if the device was removed and not replaced immediately.
2. Outcome measure survey
   All participants agreed that the outcome measures survey was suitable for assessing the progression of their LBP. Completing the initial survey on a tablet was acceptable as long as help was available from a researcher or research assistant if required. The participants describe the tablet as “clunky”, explaining the problem with using the vertical sliding scale within the EQ-5D 5L tool.

Some participants reported difficulty in understanding the wording contained within the Tampa Scale for Kinesiophobia, other than this all other questions within the outcome measures survey where deemed clear and understandable. Participants debated the use of a 10-point or 100-point NRS, however no consensus in preference was reached, concluding that both numerical scales are acceptable. They warned that participant’s answers may vary dependent upon the time at which the survey is completed relative to a patient’s diurnal pain pattern and the timing of analgesia. However, participants also acknowledging that dictating a specific time for survey completion would not be feasible due to variation in participant’s daily lives. Overall, the participants agreed that the survey evaluated their LBP journey well but recommended the formal assessment of sleep within a full trail.

Table 10. Provides illustrative quotations to demonstrate each theme.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 10. Focus group, comments that reported or discussed each theme and illustrative quotations from patients (quotations have been copied verbatim).

https://doi.org/10.1371/journal.pone.0229792.t010

Integration: Feasibility, suitability and acceptability

For the ‘general’ trial objectives 90% of the success criteria were met. Both the general and specific objectives demonstrated good overall feasibility, suitability and acceptability. Table 11 displays evidence demonstrating the extent to which success criteria were met and potential improvements to trial design.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 11. Success criteria.

https://doi.org/10.1371/journal.pone.0229792.t011

Principal findings

This feasibility trial evaluated the feasibility, suitability and acceptability of assessing the effectiveness of independent prescribing by APPs for patients with LBP in primary care, to inform the design of a future definitive SWcRCT. Over a recruitment period of 6 months, 29 participants were recruited across three clinical sites. The average age range of participants was 40–49 years, reflective of international demographic data for LBP [92–94]. Trial objectives were evaluated against predefined success criteria. 90% of the success criteria were met. Specific objective benchmarks evaluating adequate time to complete ‘trial-related tasks’ and recruitment and retention targets were not met. 48% of participants were lost to follow up by 6 weeks with 65.5% lost to follow up by 12 weeks. Both the planned primary and secondary outcome measures were feasible and acceptable.

Acceptability of interventions

90% of the success criteria were met indicating that the methods tested are feasible, suitable and acceptable for use in a definitive trial. The data further strengthens trends found in the literature, demonstrating that healthcare service users are accepting and satisfied with NMP and have confidence in clinicians’ NMP skills and competence [95, 96]. Specifically, participants welcomed the APP’s ability to include prescribing as one part of a comprehensive and holistic management plan. The clinicians all agreed that NMP enables better holistic care by FCPs for LBP in primary care.

Eligibility criteria

The eligibility criteria were designed to enable the recruitment of patients experiencing medium risk LBP, stratified by the STarT back tool. The majority of patients stratified to the medium risk group are acute or subacute in nature, potentially benefitting from a multi-modal management approach including the use of analgesia [5, 6]. APPs participating in the feasibility trial agreed that the eligibility criteria were suitable to allow the evaluation of the trial objectives. They echoed the literature in recognising that the condition is the predominant musculoskeletal problem presenting at primary care clinics potentially requiring analgesia as part of its management [5, 56–58]. Synthesis of findings support amendment of the eligibility criteria for a full trial to include LBP patients across all three STarT back stratification groups.

Epidemiological literature highlights that for first episodes of LBP, pain is seen to improve rapidly in the first four to six weeks and is commonly fully resolved by 12 weeks [94, 97]. This is not the case for the majority of patients with recurrent episodes or persistent LBP. In these patient groups, pain is often accompanied by more prominent psychosocial drivers, with patients commonly developing issues which require a clear long-term individualised psychosocial management plans [94, 97]. If a definitive trial is to include all STarT Back stratification groups, additional longitudinal follow up procedures should be incorporated, rescheduling the trial primary endpoint to 1 year to allow for evaluation of patients in the long term.

The eligibility criteria specified that only patients with non-specific back pain +/- leg pain requiring medication advice and drug prescription qualified for inclusion. Findings highlighted that the scope of iNMP includes not only the prescription of medicines but de-prescribing and medicines advice and management. Clinicians advised that these key skills are the prescribing skills most frequently optimised across the spectrum of LBP. The NHS spent £17.4 billion on prescription medications in 2016/17, with prescribing of analgesia for MSK pain significantly increased compared to the previous decade [98]. The APPs reported that they are frequently required to de-prescribe inappropriate and/or potentially harmful analgesia provided by other clinicians, or to optimise the use of drugs already prescribed to enhance rehabilitation potential. These observational findings emphasise the inappropriate overuse of paracetamol, non-steroidal anti-inflammatory drugs (NSAIDs), opioids and gabapentinoid medications for the treatment of pain reported in the literature, despite published prescribing guidelines [6, 99]. In the UK, 24 million prescriptions for opioids were issued in 2017 [100], with gabapentinoid prescribing tripling over the last decade. Many of these drugs were prescribed for persistent LBP +/- leg pain [101]. Although it is hoped that these drugs are prescribed appropriately within governance frameworks, it is postulated that repeat prescriptions alongside insufficient clinical follow up, propagate prolonged use of these potentially dangerous drugs [102]. There is a current deficit in research evaluating how physiotherapist independent prescribing is used to manage LBP. It is imperative that a full trial collects and evaluates data inclusive of the whole scope of physiotherapist independent prescribing (including what is prescribed, de-prescribed or advised), with eligibility criteria enabling the inclusion of all patients with non-specific back pain +/- leg pain.

Recruitment

Recruitment rates were found to vary between the sites where identical weekly appointment slots were available. The rural and regional city sites took approximately one month to recruit 10 participants with the capital city site recruiting nine participants over the full six-month recruitment period. Interestingly, the key reason identified for the slower recruitment rate was that patients fulfilling the eligibility criteria at this location did not want to take medication for their pain. Instead, participants reported that access to a physiotherapist for assessment and management permitted an alternative to the pharmacotherapy provided by the GP. This reflects the literature evaluating the use of direct access to physiotherapy in primary care. Physiotherapeutic holistic assessment for MSK conditions and joint decision-making regarding the appropriate management plan for the individual, has been shown to provide greater levels of patient satisfaction when patients are able to seek care directly from a physiotherapist without prior mandatory medical-input [103, 104]. Completion of an adequately powered trial would be feasible using recruitment rates based on the rural and regional sites not that obtained in the capital city. However, it is posited that with the expansion of the proposed eligibility criteria to include all patients with non-specific LBP +/- leg pain, the full scope of physiotherapist prescribing and the adoption of additional recruitment capacity via research assistants and administrative staff, that recruitment rates and retention at all sites would be acceptable.

Follow up procedures and retention

Poor clinician time capacity is a recognised barrier to conducting clinical trials [44, 105, 106]. Both patients and APPs advocated for the use of research assistants to aid with trial recruitment, consent and follow up administration. To improve retention, participants recommended reminder telephone calls and one-to-one appointments where participant literacy levels limited completion of follow up surveys. Further, it was proposed that research assistants would improve retention by acting as a consistent point of communication, encouraging smooth participant flow through the trial.

Previous literature has linked the use of a combination of recruitment and follow up strategies, with improved retention rates [105, 107]. This improvement is attributed to sustained, frequent contact with participants as they move through a trial. Adequate statistical power and good external validity rely on sufficient participant numbers [44, 105, 106]. As loss to follow up in this feasibility trial was higher than the 20% deemed acceptable within research methods literature [105, 107], it is essential that the design of a full trial engages several strategies to improve retention. The literature proposes that a minimum of three communication channels should be provided by each participant, including contact through friends and relatives, with regular updates, ‘check in’ communication via text, email and telephone and face-to-face appointments if preferred. ‘Branding’ the trial with a recognisable name and logo embossed on all trial documents and correspondence may also improve retention owing to inferred credibility, enabling participants to build a bond with the research [105, 107].

Outcome measures and data collection methods

The literature reports that the use of a core outcome set assessing pain intensity, health related quality of life and physical function is required for the assessment of non-specific LBP [28]. As optimal tools are not defined in the literature, the primary and secondary outcome measures were selected and agreed upon by a group of clinical and academic experts. The appropriateness of the selected outcome measures for use in a full trial were evaluated. All APPs and patients agreed that the outcome measures were suitable and acceptable. However, the patients recommended that the option of completing the survey on paper would be beneficial for those with limited access to email or poor IT skills. Data demonstrated graduated improvements in pain, function, disability and activity over the 12-week assessment period, mirroring the trends for medium and low risk LBP reported in the literature [92, 93, 97]. No floor effects were detected across the outcome measures used. 100% data completeness was achieved by using an online survey. Data collection via an online survey was deemed acceptable and feasible by both the clinicians and patients, supporting the literature that demonstrates better and quicker response times with fewer missing responses across both open and closed survey questions [108, 109].

This feasibility trial aimed to evaluate participant compliance through assessments of wearing an accelerometer alongside the ease of fitting the devices and data collection. Treatment effect was not assessed. Participants fitted with accelerometers achieved 100% compliance and 100% data completeness, demonstrating feasibility of use in a full trial. All participant returned the accelerometer in person, therefore the feasibility of returning the device by post was not formally evaluated. Participants and APPs reported that the devices would be useful to include in a full trial. They were fitted easily and owing to their positioning participants did not feel that the devices prompted them to increase activity levels after the first 24hrs. This is consistent with the accelerometer literature which demonstrates that removable accelerometers worn on the wrist, ankle or clothing may prompt increased physical activity and might lead to poor data completeness owing to participant removing the devices and forgetting to replace them [110–112].

The published protocol for the feasibility trial detailed the assessment of sleep via accelerometer data. Unfortunately, this was not evaluated owing to restriction in the technology available for testing [113]. This deficit was highlighted by the APPs, who recommended the addition of a questionnaire-based outcome measure assessing participants’ sleep. 50–60% of people experiencing either acute or persistent low back pain experience high levels of sleep disturbance [114]. Poor sleep over long periods of time may lead to depression, obesity, diabetes and cardiovascular disease [114, 115]. Patients with LBP suffering with sleep disturbance have been reported as twice as likely to be hospitalised owing to their pain [116]. The literature demonstrates that improved sleep modulates pain intensity [117], with poor quality sleep associated with increased pain intensity, fatigue, decreased function and psychological stress. Although perceived sleep quality has been shown to be different to the objective reality assessed via polysomnography or actigraphy, subjective assessment via sleep questionnaires and diaries have been shown to be valuable where objective evaluation is not possible [118]. Based on this rationale, it would be suitable to add a validated and reliable sleep questionnaire into the outcome measures survey for use in a full trial.

Findings from this feasibility trial indicate that a definitive SWcRCT is feasible following some minor modifications. The SWcRCT should include all patients presenting with non-specific LBP +/- leg pain and capture data representative of the full scope of physiotherapist independent prescribing. To navigate limited clinician capacity and time restrictions dictated by job plans and service specifications, researchers should consider the use of research assistants to recruit, consent, aid in data collection and complete follow-up and administrative tasks. Prior to the completion of a definitive full SWcRCT, recruitment and follow up procedures should be modified in accordance with the feasibility trial data. The online outcome measures survey should be revised to include a validated sleep evaluation tool, and the survey logic updated. Revised procedures and both online and paper versions of the survey should be piloted across all LBP stratification groups to evaluate successful modification before use in a definitive full SWcRCT.

Strengths and limitations

This feasibility trial used rigorous systematic methods including analysis and synthesis strengthened by an imbedded qualitative component and the engagement of expert trial management and steering groups including clinicians, healthcare managers, academics and patient and public representation. This combination ensured specialist knowledge of physiotherapist independent prescribing and LBP alongside specific primary care perspectives, facilitating a rigorous analytical process. There were no adverse effects to the treatments or methods evaluated. Individuals recruited to the qualitative component of the trial were observed to be comfortable throughout the process, expressing their thoughts and opinions openly. This feasibility trial is limited by the small samples used in both the trial and qualitative components; however, samples did satisfy the theoretical representation of the population essential to evaluate the trial objectives. No guidelines exist defining best practice for feasibility trials evaluating trial methods prior to SWcRCT. Although this may limit the trial, the authors have utilised transparent, integrated best practice from aligned guidelines, whilst ensuring robust consultation with subject and methodological experts and representatives from the public, throughout trial design.