Development of a pain self-management intervention framework for people with spinal cord injury

Background Pain is the most common reason for medical visits to primary health care practitioners. Pain self-management interventions are encouraged and there is no known self-management intervention framework available that clinicians and people with spinal cord injury (PWSCI) can use to guide treatment selection. Aim This study aimed to develop a pain self-management intervention framework for PWSCI. Setting Online and facilitated in Gauteng, South Africa. Methods A three-round modified e-Delphi method was used to reach an 80% consensus among a 21-expert panel. Fifty-nine interventions were distributed via REDCap and a final online audio meeting was held to either include or exclude interventions in the final framework. SPSS v27 was used to analyse descriptive data and content analysis was used for qualitative responses. Results The final developed pain self-management framework consists of 56 interventions and includes interventions from multiple health professions to encompass medical, psychological, therapeutic and social interventions. Interventions are also specified for nociceptive and/or neuropathic pain and grouped according to the biopsychosocial model. Conclusion The interprofessional framework may be used as a guideline for PWSCI to alleviate pain, as well as assist health professionals in clinical decision-making, by providing them with the freedom to choose acceptable and adequate interventions that may be appropriate to treat the affected individual’s pain. Contribution Pain management is a basic need at the primary healthcare level and PWSCI need access to the broad range of interventions available to manage their pain. The framework highlights the variety of appropriate interventions to guide both health professionals and PWSCI with pain relief options.


Introduction
A spinal cord injury (SCI) results in partial or complete loss of motor and/or sensory functions below the level of injury and results in common secondary health conditions (SHCs), including pain, pressure ulcers as well as urinary and bowel complications. 1,2Secondary health conditions are debilitating and worsen the experience of disability for people with spinal cord injury (PWSCI) by negatively influencing their dignity and independence. 3The SHCs also cause PWSCI to consult more with health professionals and cause multiple readmissions to the hospital. 4,5Approximately half to two-thirds of PWSCI experience pain, which is an unpleasant sensory and emotional experience that is associated with either actual or potential tissue damage. 6,7Pain often commences within the initial 6 months after SCI, with the possibility of being aggravated over time. 8he experience and perception of pain may be intense and reported as severe to extreme, negatively affecting health satisfaction and interfering with activities of daily living. 9,10The multifactorial pattern of pain ultimately affects how patients react to pain and respond to pain management.In addition, the different underlying mechanisms of pain further challenge pain management. 11For example, pain management can be further complicated by a variety of emotional, behavioural and social factors that can negatively affect the experience of pain. 12The severity of pain is influenced by various factors such as genetics, comorbidities, current psychological state, prior experience of pain and socioeconomic circumstances, 13 and treatment is rarely aimed at all associated factors of pain, resulting in ineffective treatment strategies. 14Despite the challenges, management of pain is essential, and without intervention, PWSCI may experience additional losses in function and community mobility. 15ain management interventions following SCI include task modifications, therapeutic treatments, and pharmacological and surgical options.Therapeutic interventions such as physiotherapy are recommended as first-line treatment 16 with a wide range of therapeutic interventions such as thermotherapy, electrotherapy, massage and exercises. 17espite recommendations, pharmacological treatment of pain remains the mainstream choice of pain management in the spinal cord population.However, PWSCI still report that significant pain relief is difficult to achieve, with pharmacological treatments providing very minimal pain relief. 18Corticosteroid injections and surgical interventions (such as rotator cuff repairs in the shoulder) are only considered as a last resort to manage severe pain.Surgical intervention is often unsuccessful in PWSCI because of the extended 6-week period of rest required after surgery. 14The strict protocols post-surgery hinder independence in wheelchair use 19 and create a high risk of developing potential SHCs such as pressure ulcers when on bed rest.Pain is the most common reason for medical visits to primary health care practitioners, 20 and frequent visits for pain-relieving interventions may not be logistically and financially feasible, especially for affected individuals from low-income households. 21Coupled with findings that self-management tasks and skills incorporated into daily life reduce SHCs, it is unsurprising that self-management programmes are steadily being more common. 22,23Furthermore, literature has highlighted the need for increased self-management efforts for PWSCI to reduce overall SHCs. 24Self-management is defined as the 'individual's ability to manage symptoms, treatment, physical and psychosocial consequence and lifestyle changes inherent in living with a chronic condition (p.178).' 25 Selfmanagement interventions may refer to a single or combination of approaches that PWSCI can learn (or be taught by any health professional) to enable them with the skills to reduce the impact of the pain on their everyday life. 26Behaviour therapy, relaxation techniques, stretching and exercises are among the wide variety of self-management techniques identified to relieve pain. 27Selfmanagement also includes the ability of affected individuals to monitor their pain condition and their behavioural, cognitive and emotional responses. 28Self-management includes problemsolving, decision-making, using resources, communicating with health professionals and taking action when needed. 29eople with SCI must have self-management programmes targeted at all three categories because of the complexity of SCI, which often present with multiple chronic conditions. 30lf-management of chronic pain requires that affected individuals distinguish themselves from the pain they experience so that they do not define themselves by their pain. 31Barriers to self-management include the overwhelming effort to manage pain, unsupportive family members or health professionals, and limited problem-solving abilities. 31,32There is a need for increased self-management efforts by both healthcare providers and PWSCI to reduce SHCs in PWSCI, 33 and this study aimed to develop a guided pain self-management intervention framework from the perspective of HCPs working with PWSCI.

Study design
This study utilised a qualitative approach and a Delphi design to obtain the most reliable consensus from a group of experts by using surveys to receive controlled feedback on their opinions, especially when there is little or no definitive evidence on the subject. 34,35,36A Delphi design involves an iterative process that is useful in gathering subjective information from experts working in the field of interest and there are variations on the classic Delphi method. 37The classic Delphi is modified by presenting the experts with a predetermined list of interventions that may be used to manage the pain instead of the experts determining the list, hence the term 'modified Delphi'.The predetermined list comprises interventions to manage pain and not specifically self-management interventions, as the experts needed to select the interventions that would be appropriate for selfmanagement.The predetermined list of interventions was derived from phase 2 of a mixed-method study by the authors and published literature.Phase 2 explored the experience of pain by PWSCI and the coping strategies they used to manage their pain, following guidance from the first quantitative phase.The first phase found that 85% of PWSCI reported pain with a severity of 6.7/10. 38This study is the third phase of this mixed-method study. 39The approach is referred to as a 'modified e-Delphi' design in this study because all three rounds of the modified Delphi were conducted online. 37The modified e-Delphi study was conducted and informed by a predefined and published protocol 39 and is reported as per the Conducting and Reporting Delphi Studies (CREDES) recommendations. 40

Setting
This study was conducted online: REDCap for the first two rounds and Microsoft Teams for the third and final round.

Study population and sampling strategy
Healthcare professional experts were identified from a variety of different professional backgrounds in the field of SCI and/or pain management in academia and clinical practice.The experts were identified from their global contribution to the following associations: International Spinal Cord Society, Africa Spinal Cord Injury Network, Southern African Spinal Cord Association, and PainSA.Experts in academia needed to be in a possession of a master's degree and experts in clinical practice had to have at least 10 years of experience in their respective fields.Sixty-four experts were purposefully sampled and invited via email to participate in the study.A detailed information sheet explaining the study aims, process and informed consent was emailed to the experts before participating in the study.

Data collection procedure
Figure 1 shows the procedure of this modified e-Delphi study.A pilot study with the same procedure as planned for the main study was conducted, as guided by Clibbens et al. 41 Two experts meeting the inclusion criteria were included in the pilot study (one in the SCI field and the other in the pain management field).The pilot study was conducted over 1 week; feedback was received and collated and modifications were implemented before commencing with the main study.Results from the pilot study were not included in the main results as changes were made, and the two experts who participated in the pilot study were not included in the main study, but remained experts for the round two pilot study.

Round one
REDCap was used to construct and distribute the predetermined interventions.Participants were given 2 weeks to complete the survey and those who did not respond by the end of the 2 weeks were removed from the study.Reminders were sent at the end of the first week and 2 days before the deadline.The participants were given the option to select 'agree', 'disagree' or 'do not know' for each intervention as guided by Eubank et al. 34 Participants were required to give reasons for their selection in the text box provided and also had the option of adding comments and extra interventions that were not on the predetermined list.The third week was allocated for data analysis.The items were analysed descriptively using SPSS v27 to determine the 80% consensus 34 and content analysis was used to analyse the data from the free text boxes.Interventions that did not receive consensus were redistributed in round two without giving participants feedback on how the items were scored to maintain impartiality.

Round two
Feedback from experts in the comment sections recommended that the researchers consider giving more information over and above the descriptions provided for each intervention.Therefore, one moderate-to-high level of evidence article was included per intervention in round two.Medline (PubMed) was used to search for each intervention that was redistributed in round two.In this round, participants had to either 'agree' or 'disagree' only; thus the 'do not know' option was removed to receive firm decisions from the participants.Similar to the previous round, the participants were also required to give reasons for their selection in the text box provided.A pilot study was also conducted for round two by the same experts who were involved in the initial pilot study.Comments and feedback from this pilot study were collated and changes to the round two survey were carried out.The pilot study results were also not included in the main study because some changes were required for the main study and the experts involved in the pilot study did not participate in round two.The survey for round two was also distributed via REDCap and conducted over 2 weeks.Reminders were sent as per round one and non-responders were removed from the study.The same steps as in round one were followed to analyse the data.

Round three
The aim of round three was to discuss the interventions that did not reach 80% agreement to include or exclude from the final framework.Participants who responded in round two were sent a 'Doodle poll' link to select their preferred time slot over 14 days to conduct round three via Microsoft Teams, and were not anonymous.The date and time when most of the experts across different professions were available were chosen as the day of the meeting.Interventions that did not reach the 80% consensus in round two were redistributed in round three and there was expert interaction to exchange important information and reasons for disagreements were clarified.Previous interventions that met the 80% consensus in round two but received objections in round three were discussed and removed following unanimous decisions to do so.Only the interventions that reached 80% consensus at the end of this round were added to the final framework.The consensus was decided by a vote of hands, where the participants typed a 'thumbs up or thumbs down' in the chat and the lead author counted the votes for each intervention.The final list of interventions was grouped for neuropathic and/or nociceptive pain and also according to the biological, psychological and social components of the biopsychosocial model of care.

Data analysis
The SPSS v27 was used to descriptively analyse the participants' demographic information, as well as the items to determine consensus.Data were reported in frequencies, percentages, means and standard deviations.Qualitative data were analysed using content analysis as guided by Erlingsson and Brysiewicz. 42part from maintaining consensus in all the rounds, rigour was also controlled in the last round by giving the experts the opportunity to revise the findings of the two rounds, using consensus, to determine which responses were appropriate.

Ethical considerations
This modified e-Delphi study is the third phase of the mixedmethod study that is registered with the South African National Health Research Database (reference GP201806005) and received ethical approval from the Faculty of Health Sciences Research Ethics Committee of the University of the Pretoria, South Africa (approval number 125/2018).Written informed consent was obtained from all the participants in this study, including consent for the third round to be audio-recorded.Furthermore, participants' names are not used to ensure confidentiality.

Participants' demographic information
Twenty-two participants from South Africa, Sweden and Canada accepted the invitation and 21 participants completed the survey and were included in the study.The mean age of the participants was 42.55 years (s.d.= 10.53) and the mean year practicing in their professions in their highest qualification was 11.09 years (s.d.= 10.47).The majority of the participants had a master's degree (n = 10, 45.5%), practiced in the clinical setting (n = 12, 54.6%) and were mainly physiotherapists (n = 8, 36.4%)as shown in Table 1.Tourniquet application, surgery and recreational cannabis had the majority of disagreements to be included in the final framework; however, the disagreements did not reach 80% consensus (61.9%, 47.6% and 61.9%, respectively).Most of the participants did not know whether to include self-hypnosis (47.6%),Thai massage (52.4%),Maitland and Mulligan joint mobilisations (42.9% and 57.1%, respectively) and electrotherapy to non-innervated areas (38.1%).Although not in the majority, pharmacological interventions saw many participants not knowing whether to include the specific drug groups, namely, anticonvulsants (42.9%), serotonin-selective reuptake inhibitors (SSRI) (42.9%) and serotonin-norepinephrine reuptake inhibitors (SNRI) (47.6%), antidepressants, as well as anxiolytics (42.9%).

Round one
Following the inductive thematic analysis of round 1, three themes were identified.The qualitative results support the quantitative results as it was found that most 'do not know' responses were because of concerns about medications that need to be prescribed appropriately, interventions falling outside of one's scope, and more information necessary before decisions could be made about including or excluding interventions from the framework.The thematic analysis guided the researchers to include the literature in the second round.

Theme one: Medication should be appropriately prescribed
There was a general concern with medication dependency and organ function with prolonged use, and participants reported that there must be the correct use of the medication with consideration of side effects:

Theme two: Beyond the scope of practice
There was a general unfamiliarity with certain interventions that fell outside one's scope of practice:  Ten more pain self-management interventions were added to round two following the content analysis of suggestions from the participants and are shown in Table 3.

Round two
Sixteen .Six of the 10 new interventions that were included in round two reached a consensus.Of the initial interventions, tourniquet application, surgery and recreational cannabis still showed disagreement to be included (75%, 43.8% and 68.8%, respectively) and 10 interventions (40%) from the original 25-item list did not reach consensus.Table 3 also depicts the references of the articles used in this round and the respective level of evidence (LOE) as guided by the Johns Hopkins Nursing Evidence-Based Practice. 76o themes were identified following the inductive thematic analysis of round two.Concerns were raised about side effects and the availability of the equipment that would be necessary for some of the interventions.

Round three
Thirteen interventions that did not reach consensus in round two were redistributed and discussed in round three via an audio conference meeting on Microsoft Teams.Fourteen participants responded to the 'Doodle Poll' and nine participants were available at one time slot and subsequently accepted the meeting invite.The nine participants included four physiotherapists, three occupational therapists and two medical doctors.One physiotherapist did not join the meeting and a final eight experts (all from the medical fratenity) participated in round three (50% response rate from round two).The consensus was achieved with 80% agreement for all 13 interventions that did not meet the consensus in round two to be excluded from the final framework as shown in Table 4.The meeting was conducted in English and was recorded and transcribed.
Table 5 depicts the consensus to specify the type of pain for each item.Following robust discussion, Thai massage and auto-assisted Maitland joint mobilisations, which both received 81.3% agreement in round two to include in the framework, were excluded following a unanimous decision.The reasons cited by the experts included that masseuses providing Thai massage were not medically trained and that Maitland techniques often need a therapist to effectively provide the necessary pain relief.
The general agreement by the participants was that some of the interventions need to be adequately taught to PWSCI by qualified health professionals where applicable.Furthermore, participants emphasised the following interventions: • All prescribed medication: The source of pain must be considered and the prescription must be appropriate for optimum pain relief and avoidance of dependency.• Psychoeducational management: Emphasis was placed more on neuropathic pain than nociceptive pain.• Active stretches of innervated areas: Emphasis was placed more on nociceptive pain.• Transcutaneous electrical nerve stimulation (TENS): To be applied with caution and individuals must be taught the correct parameters before application.
Figure 2 illustrates the final guided pain self-management intervention framework as per the biopsychosocial model.

Discussion
This study derived an expert consensus list of interventions to include in a guided pain self-management intervention framework for PWSCI.The overall response rate was 73.9% across the three rounds, which is slightly lower than the 80% overall response rate from studies reviewed by Gargon et al. 85 Literature has found that a higher number of items tend to result in lower response rates in subsequent rounds and vigorous strategies, such that up to four reminders are necessary for higher response rates. 77,78e  smoking is found to have both causal and reciprocal effects on pain in the general population 80 and worsening pain in PWSCI. 81Increased leisure-time physical activity as well as physical activity targeted at weight loss is associated with decreased pain levels and improved health satisfaction in PWSCI. 82,83,84Lifestyle changes can be achieved by using selfmanagement programmes that focus on behaviour change. 85e majority of the psychological interventions reached ≥ 90% agreement including cognitive behavioural therapy (CBT), mindfulness, psychological control techniques and     psychological educational management, which are some of the most common self-management interventions. 27,32lthough included in the final framework, many of the pharmacological interventions only reached consensus in the second round following the inclusion of literature evidence.This is unsurprising as the panel of experts only consisted of one pharmacist and four medical doctors.Tricyclic antidepressants and NSAIDs were the only groups of medications that were agreed on in the first round.Although this may seem as if the experts did not, at first glance, agree for all groups of medications to be part of self-management, likely, tricyclic antidepressants and NSAIDS are still known by the experts to be first-line pharmacological interventions for neuropathic and nociceptive pain, respectively.Although NSAIDs remain appropriate for nociceptive pain, anticonvulsants such as pregabalin are now the first-line medication for neuropathic pain. 85,86The content analysis identified that medication should be appropriately prescribed and will only be useful when used as prescribed and only under the guidance of a health professional.Inappropriate medication prescription is a public health concern, especially in vulnerable groups such as the elderly and people with disabilities. 87Prescribers should, therefore, take care to not only prescribe appropriately but also address the adverse outcomes of prescribing medication. 88Pain medication misuse is common in PWSCI and the risk is greater for those with increased pain severity. 89Prescribers need to also consider alternative interventions to reduce the potentially catastrophic consequences of pain medication misuse (such as death from overdosing).

Interventions
Although the interventions that were eventually excluded did not reach a consensus to be excluded from round one, there was consistent agreement among experts to exclude tourniquet application, surgery, recreational cannabis and electrotherapy in non-innervated areas.There was a major safety concern when PWSCI need to be performing some interventions that were suitable for self-management without supervision.Indeed, maintaining control over one's care is a facilitator in self-management; 33 however, techniques need to be performed correctly to not only be effective but also to prevent injury.The inclusion of the tourniquet application was a surprise to many experts, as this is a known intervention to stop blood supply and can be dangerous in individuals with already poor blood circulation and impaired sensation.Tourniquets are only applied during surgery of the limbs and are already known for possible nerve, muscle and skin injury. 90One person with SCI used this technique to ease neuropathic pain below the level of injury.A panel discussion by Evans et al. 91 concluded that blood flow restriction training was safe and effective for improved skeletal muscle strength, when performed by a skilled health professional.However, the potential for pain relief needs further research, including its use as a self-management technique because of the potential adverse effects.The need to achieve pain relief, coupled with other factors such as barriers to accessing health facilities, leads PWSCI to attempt anything, including recreational cannabis. 92Although there is some effectiveness shown with the use of cannabis in PWSCI, 92,93  Therefore, appropriate referrals will need to be made for a holistic treatment for PWSCI with pain to improve HCP practice.The onus rests on the HCPs to increase their knowledge of the various interventions available and expertise in managing pain using other techniques within their scope of practice.Although more difficult to establish, an interdisciplinary approach has been shown to increase the effectiveness of pain management programmes 95,96 and could be a consideration for PWSCI.

Strengths and limitations
The limitation of a Delphi study is that the expert opinions are often based on either their experiences or biased.However, the results of Delphi studies form an important foundation for decisions that are relevant for clinical practice, especially because clinical guidelines are often grounded on expert opinions and experiences. 40This study had a high response rate in the first two rounds, losing only 4.5% of the experts in round one and 23.8% in round two.There was a 50% response rate in round three and this attrition rate may be because of the online meeting being held during working hours.All experts were employed and availability would not be possible for some experts despite the wide range of timelines over the 14 days given to them.Furthermore, experts resided in different countries and had up to 9 h of the time difference, further making some experts unavailable for certain time slots.Hasson et al. 97 have reported on attrition because of the length of commitment as the rounds progressed.It was initially planned for 7-14 experts 39 but opted to invite more experts in anticipation of a possible decline in the response rate as the rounds progressed.
Although the study commenced with an expert from each profession involved in SCI care, the participants were not evenly distributed across the different professions, resulting in some professions not being represented in the final round because of attrition or unavailability.
Following comments from participants that more information was needed on certain interventions, articles were included on all the interventions and the participants were encouraged to read up on the interventions themselves.Round two was piloted to ensure credibility; however, the authors take into consideration that the inclusion of the literature may have not prevented bias, because the participants were not expected to make their own research before participating in this study.Despite the inclusion of the literature in round two, 40% of the interventions still did not reach a consensus and needed to be discussed in round three, suggesting that those who did not agree in round one did not change their mind in round two.Following the conclusion of the meeting, the experts involved confirmed that the final document was the true representation of the meeting.
Weaknesses of the developed framework include having no items on education that are specifically tailored to motivate towards self-management.It is important to include formalised patient education programmes that will guide the health behaviour changes so that PWSCI may be able to relieve the pain. 98There were no items about telerehabilitation, and in light of the coronavirus disease 2019 (COVID-19) pandemic, it may be worthwhile to determine which self-management techniques may be conducted using such technology.Telerehabilitation may be telephone-or internet-based and may be beneficial in not only improving communication with health professionals in the core skills of self-management but also ensuring that prescribed interventions are performed accordingly.Telephone-based counselling can be considered for affected individuals who have to isolate or quarantine for prolonged periods, as this technique has been shown to improve pain management, 99 coping skills, and reduce depression in people with disabilities. 27Notwithstanding the aforementioned limitations, the developed guided pain selfmanagement intervention framework has the potential to improve pain management after SCI.

Recommendations
Although the evidence does highlight the involvement of some health professions over others in SCI care, it is recommended that future Delphi studies employ specific sampling strategies to ensure that all professions are adequately recruited to prevent loss of representativeness as the rounds progress.It is also recommended that further research be conducted to test the clinical suitability of the interventions included in this framework.

Conclusion
This study created a list of interventions that were informed by PWSCI and based on the opinions of health professional experts that may be used to manage pain for PWSCI.Having a sense of control over one's care is positively associated with emotional well-being and an improved ability to deal with stress.An individual's belief in their capabilities can positively impact mood, subsequently reducing the experience of their pain.

FIGURE 1 : 1 Feedback
FIGURE 1:The procedure of this modified e-Delphi study.

'
Opioids are addictive and may lead to poor quality of life in the long run.' (Participant 16, 46 years, MOccTher) 'Medication is useful when used appropriately and under the guidance of a clinician or medical practitioner.'(Participant 4, 47 years, MSc Occupational Therapy) NSAIDs, Non-steroidal anti-inflammatory drugs.

Table 2
illustrates the descriptive statistics of the predetermined items, with the first column specifying where the interventions were sourced from.Twenty-one participants completed round one (95.5% response rate) and one participant did not complete the survey.The consensus was achieved with 80% agreement for 34 interventions out of the original 59.Twenty-five interventions, therefore, did not achieve consensus in this round.

TABLE 2 :
Round one descriptive statistics of the items (59 interventions).

TABLE 3 :
Round two descriptive statistics of the items.

TABLE 4 :
Round three descriptive statistics of the items.

Table 5
The biopsychosocial aspects of the developed guided pain selfmanagement intervention framework.
continues in the next column →FIGURE 2: 94nnabis use is still understudied, and inclusion of medicinal cannabis in this framework is performed on the condition that the country of use has legalised it and has been prescribed by a health professional.Experts in this study pointed out that some of the interventions are not widely available in the local context.Although some interventions have proven effective in other countries, it may not always be implementable in another country such as South Africa.94Thismaybe because of the various factors such as differences in the communities, attitudes and beliefs, or financial accessibility.The interventions were therefore included on the premise that they would be available for PWSCI in the local context.By including the 'do not know' option in round one, the authors were able to determine that although healthcare professionals work within a multidisciplinary approach in SCI rehabilitation, there is inadequate information known about interprofessional pain management techniques.This may pose a problem where advocacy is needed for a patient regarding adverse effects picked up by another professional.For example, although physiotherapists in South Africa do not prescribe drugs, knowing the effects of the drug assists in knowing what to expect during therapy or when certain self-techniques are prescribed.A medical practitioner would not know how to adequately advise a patient who is not responding well to nonmedical interventions.
Some of the interventions could be applied by PWSCI at home following appropriate training on how to perform them independently.The final framework consists of 56 interventions that may be appropriate for nociceptive and/ or neuropathic pain as selected by a panel of health professionals, and span across the different domains of the biopsychosocial model.Most of the experts agreed that interventions are only as effective when administered correctly, and thus proper education on how to perform the interventions adequately and safely is imperative.It must be noticed that all the interventions that were agreed on would not necessarily be appropriate for all PWSCI, and the type, area and severity of the pain would need to be thoroughly assessed.Furthermore, the effectiveness of these interventions needs to be established before the implementation of this framework into clinical practice.This is the first framework of self-management interventions developed for the SCI population to the knowledge of the authors.Pain after SCI has been reported to be difficult to manage and the authors envision this framework to play a vital role in assisting health professionals to know the array of interventions appropriate to guide PWSCI towards pain relief.