Introduction

Advancements in spinal cord injury (SCI) rehabilitation, assistive technology, and long-term injury management have led to improvements in life expectancy for people with SCI [1]. However, evidence suggests these additional life years are impacted by complications from premature aging and decreased independence in everyday life [2, 3]. In response to this demographic shift, participation in daily activities and social roles has been recognized as a fundamental rehabilitation goal [4]. As full participation is recognized as a global human right and is associated with improved physical health, psychological health, and life satisfaction [5, 6], participation presents as an ideal intervention target to support people aging with SCI.

Participation has been defined as the nature and extent of involvement in daily activities and social roles [4, 7]. However, the definition’s breadth creates mixed understandings of how to conceptualize and measure participation. For example, participation’s interchangeable use with different terms (e.g., social participation, involvement, community re-integration) creates confusion around the term’s definition [8, 9]. Until recently, disability-focused conceptualizations of participation have primarily focused on the extent of one’s participation, while ignoring the subjective experiences of participating [10]. A conceptualization of aspects accounting for experiences and satisfactions of participation was only recently developed by Martin Ginis et al. [11]. While the importance of participation in daily activities and social roles is evident, theory-based, behaviour change interventions that address participation in people aging with SCI have not been developed.

Behaviour change interventions can be defined as “coordinated sets of activities designed to change specified behaviour patterns” [12]. These interventions have the potential to encompass the plethora of activities and roles that constitute participation [12], as they prompt consideration of many individual, social, and environmental factors that underlie why one cannot participate. Moreover, grounding interventions in theory can improve intervention replicability by providing insights into an intervention’s successes and failures [12, 13]. When applying theory to intervention design, three stages are recommended: Stage 1: understand factors that promote (facilitators) and inhibit (barriers) a behaviour; Stage 2: identify intervention options; and Stage 3: identify implementation options [12]. The Behaviour Change Wheel (BCW) and Theoretical Domains Framework (TDF) can be used to address these stages of intervention design [12, 14]. Supplementary File 1 includes a description of both frameworks. Both the BCW and TDF were developed from syntheses of several behaviour change frameworks and constructs, making them advantageous over using one theory for intervention development [12]. When used together, the BCW and TDF provide a comprehensive assessment of factors that may influence the behaviours people do to participate in daily activities and social roles.

In physical disability contexts, use of the BCW remains limited [15, 16], while the TDF has addressed physical activity and self-management behaviours [17, 18]. However, comprehensive and collective use of both frameworks remains infrequent. Although previous research has identified SCI-specific barriers and facilitators to participation [19], theory was not used, and an aging demographic was not targeted. To address these gaps, this study used the TDF and BCW to (1) theoretically identify barriers and facilitators to participating in daily activities and social roles among people aging with SCI and, (2) systematically co-develop recommendations with SCI community organizations for participation-focused interventions that can support people aging with SCI.

Methods

Study design

This study was conducted using an integrated knowledge translation approach [20], meaning each research stage was conducted in equitable partnership with key stakeholders. Three Canadian SCI organizations: Spinal Cord Injury Alberta, Spinal Cord Injury BC, and Spinal Cord Injury Ontario, with one representative who holds decision-making power at each organization (TC, CM, SC) acted as equitable research partners. A detailed description of the approach is provided in Supplementary File 2. In consultation with the three representatives, single time-point semi-structured interviews were selected as the method for data collection.

Study protocol and recruitment strategy

Ethics approval was obtained through the University of British Columbia Okanagan Behavioural Research Ethics Board. Each organization recruited participants through e-mail and social media. Eligible individuals were required to be a minimum of 45 years old, be a minimum of 10 years post injury, and not be diagnosed with a co-morbidity that impacts cognitive function. Eligibility criteria were informed by both previous literature [21, 22] and needs expressed by the community organization representatives. Interested participants completed a consent form and provided demographic and contact information. To enhance representativeness and likelihood of variation in our sample, maximum variation sampling was used for five variables selected by the organizations: age, gender, level of injury, mode of mobility, and place of residence. Following evidence-based recommendations for theory-based interviews, an a priori minimum of ten participants was selected, while conducting three additional interviews until new ideas no longer emerged [23].

Before beginning each interview, participants gave verbal consent to be audio-recorded. EG conducted the interviews with one participant at a time, either over the phone or in-person. EG knew one participant at the time of the interviews. The interview guide (Supplementary File 3) was co-developed and refined by EG, SC, TC, CM, and HG. To gain interviewing experience, EG conducted three pilot interviews with one research assistant, one SCI community member who met eligibility criteria, and one researcher with extensive experience in qualitative interviews.

Data analysis

Interviews were transcribed verbatim. Figure 1 outlines how the data analyses align with each intervention development stage.

Fig. 1: Visual representation of data analyses in alignment with three phases of knowledge.
figure 1

Creation within the Knowledge to Action Framework (Graham et al. [20]): knowledge inquiry. (Arrow 1), knowledge synthesis (Arrow 2), and knowledge tools/products (Arrow 3).

Full size image

Stage 1: understanding the behaviour

To understand a behaviour, barriers and facilitators to that behaviour must be identified [12]. To address this stage, descriptive statements of participation in daily activities and social roles were deductively extracted from interview transcripts. The Participation in Daily Activities Questionnaire [24] was used to determine if extracted statements described participation in daily activities and social roles. To enhance trustworthiness, seven transcripts were double extracted by EG and KW, and all remaining transcripts were extracted by EG only. Discrepancies during double extraction were discussed and resolved by EG and KW. Barriers and facilitators were categorized as either experienced (a factor that already inhibits/helps participation in a daily activity or social role) or assumed (a factor someone else has mentioned to the participant that inhibits/helps participation in a daily activity or social role). If the same barrier/facilitator was discussed multiple times, the factors were counted separately from one another. EG and KW independently coded all agreed upon barriers and facilitators to the 14 TDF domains (behavioural regulation; beliefs about capabilities; beliefs about consequences; emotion; environmental context and resources (ECR); goals; intentions; knowledge; memory, attention, and decision processes; optimism; reinforcement; skills; social influences; social/professional role and identity) or “n/a” if no domains were applicable. Discrepancies in coding were discussed and resolved by EG and KW.

Stage 2: identifying intervention options

To address Stage 2, the BCW guides the selection of relevant intervention functions and policy categories, respectively. From Stage 1, all identified TDF domains are systematically linked to one of six components from the Capability Opportunity Motivation—Behaviour (COM-B) model. The COM-B model acts as the central tenet of the BCW, allowing for behaviours to be understood as six components: capability (psychological, physical); opportunity (physical, social); and motivation (automatic, reflective). Intervention functions were then identified using a matrix that maps COM-B components to intervention functions. Additional intervention functions were deductively extracted and coded against the BCW’s nine intervention functions (coercion, education, enablement, environmental restructuring, incentivization, modelling, persuasion, restriction, training) [12]. Methods for extraction and coding were identical to Stage 1. To identify policy categories, all identified intervention functions were linked using a matrix that maps intervention functions to seven policy categories (communication/marketing; environmental/social planning; fiscal measures; guidelines; legislation; regulation; service provision) [12].

Stage 3: identifying implementation options

Implementation options were extracted and categorized as messengers (person/organization delivering the intervention) or modes of delivery (MODs). Messengers were grouped into broad themes, and MODs were deductively coded against the four categories of the Mode of Delivery Taxonomy Version 0 (MoDTv0): digital, human, print, somatic [25]. Methods for extraction and coding were identical to Stage 1 and 2.

Reliability

Percent agreement, Cohen’s Kappa and prevalence adjusted bias adjusted Kappa (PABAK) were calculated to examine agreement between EG and KW across all three stages [26, 27]. Inter-coder reliability was assessed using strength of agreement values: 0.41–0.60 (moderate), 0.61–0.80 (substantial), 0.81–1.00 (almost perfect) [28].

Recommendation development

The results were presented to the organization representatives to determine their relevance and usefulness to each organization. Once determined, results were discussed alongside each organization’s local context to co-develop and refine intervention recommendations. The recommendations were then evaluated using feasibility criteria (i.e., APEASE criteria to determine if the recommendations are affordable, practical, effective, acceptable, safe, and equitable) [12].

Results

Participants

Although 22 interviews were conducted, 21 transcripts were analyzed as one audio-recording was inaudible. Interviews lasted between 38 and 116 min, with a mean of 73.91 (19.54) min. Table 1 presents participants’ demographic and injury characteristics.

Table 1 Participant characteristics.
Full size table

Reliability of data extraction and coding

The number of statements extracted across all three stages of intervention design is presented in Supplementary File 4. In Stage 1, there was 64.9% agreement for double extraction of barriers and facilitators across seven transcripts. During TDF double-coding, agreement ranged from substantial to almost perfect: mean κ = 0.69 (0.18); mean PABAK = 0.91 (0.06). In Stage 2, there was 100% agreement across transcripts for double extraction of intervention functions. For double-coding against the BCW’s nine intervention functions, agreement was between substantial and almost perfect: mean κ = 0.61 (0.43); mean PABAK = 0.82 (0.18). In Stage 3, implementation option statements were double extracted with 87.8% agreement. For mode of delivery statements, agreement for double-coding was between substantial and almost perfect: mean κ = 0.77 (0.05); mean PABAK = 0.92 (0.06).

Stage 1: understanding the behaviour

Table 2 outlines the frequency and percent of identified barriers and facilitators by TDF domain. Supplementary File 5 includes additional quotes for each identified TDF domain. The ECR domain accounted for 57.2% of all barrier and facilitator statements. Examples of ECR facilitators (accounting for 60.1% of all facilitator statements) included medications and assistive technology (e.g., accessible vans, in-home lifts), and ECR barriers (49.2% of all barrier statements) were mainly factors of one’s indoor and outdoor environments (e.g., hills, buildings). For example, one participant shared an experience while visiting family:

To go to my mom’s it used to be like I would... you know like you could transfer onto a lower toilet, and you could either transfer onto a shower bench or right into the bath tub or whatever. But nowadays, it’s just so much harder, like I can’t even transfer onto her toilet. (P17, ECR)

Table 2 Identified barriers and facilitators by TDF domain.
Full size table

Almost 80% of assumed and experienced facilitator statements were coded to either ECR or social influences. Statements coded to the social influences domain primarily described how family members, friends, and peers made participating in daily activities and social roles possible or more enjoyable. For example, one participant described how her family makes it possible for her to kayak:

My husband will take me out kayaking and without him there’s no way I’d be kayaking. Between my husband and my son they can help me get into and out of a lake whether we’re going scuba diving, snorkeling, swimming, kayaking, in almost any environment. Because they’re strong, they’re able to physically help me get through any physical challenges and then I’m able to do what I do. (P3, Social Influences)

Second to ECR, 22.5% of experienced barrier statements were coded to the skills domain. Most of these statements described barriers due to a physical change. For example, one participant stated: “Right now I’m on medical leave because I’ve developed so many different health issues... makes it very difficult, if not impossible to go to work.” (P8, Skills)

All statements not aligning with a TDF domain were categorized as experienced facilitators and further coded as “nature of the behaviour”, a domain from a previous TDF version that discusses constructs such as direct experience, past behaviour, and continued engagement in everyday activities [29]. For example, one participant discussed how gardening supports continued and meaningful participation:

(Gardening) is a big thing lately. It’s a thing I took up about three years ago and I’m involved in a community garden. It helps me get a little exercise on the way there, but I also find it very therapeutic in terms of just relaxing. (P14, Nature of the Behaviour)

In discussion with the SCI organization representatives, our team concluded that ECR, skills (body changes), and social influences were the most important and influential domains for participating in daily activities and social roles. These three domains were linked to three COM-B components: physical capability, physical opportunity, and social opportunity to support intervention development.

Stage 2: identifying intervention options

Table 3 outlines the number and percent of statements by intervention function. All identified statements were coded into only four of the nine intervention function categories: education, enablement, environmental restructuring, and modelling. Across the four categories, education and environmental restructuring were the most frequently coded, accounting for 74.36% of all intervention function statements. Within environmental restructuring, 78.5% of intervention function statements described the social environment and 21.4% described the physical environment.

Table 3 Identified intervention function statements by BCW intervention function category.
Full size table

Using COM-B components from Stage 1, the matrix linking COM-B components to intervention functions linked physical capability to training and enablement; physical opportunity to training, restriction, environmental restructuring, and enablement; and social opportunity to enablement, environmental restructuring, modelling, and restriction. Using the matrix linking intervention functions to policy categories, the identified intervention functions (education, enablement, environmental restructuring, modelling, training, restriction) were linked to all seven policy categories.

Stage 3: identifying implementation options

Messengers

Messengers were categorized as individual or organizational. Individual messengers described an intervention being delivered by health care professionals (e.g., family physicians), a peer who is also aging with SCI, or rehabilitation professionals. Many participants felt that interventions delivered by several health care professionals who specialize in different areas would be valuable, as people with SCI can experience different complications from one another. The need for SCI peers as intervention messengers was attributed to having a mutual understanding of aging through lived experience. Participants viewed interactions with rehabilitation professionals during in-patient rehabilitation as valuable opportunities to learn about aging. Participants discussed how aging-focused interventions delivered during in-patient rehabilitation may equip them with knowledge and supports to potentially slow down or prevent certain age-related complications before onset.

Organizational messengers included SCI community organizations, governments, other disability organizations, sports organizations, medical supply companies, and funding agencies. Participants proposed that other disability organizations, sports organizations, and medical supply companies can partner with SCI organizations to improve overall reach. Municipal governments were primarily recognized for helping to improve intervention reach and provincial governments for providing financial support. Refer to Supplementary File 6 for additional participant quotes about intervention messengers.

Modes of delivery

Table 4 presents the number and percent of MODs by MoDTv0 category. Against the MoDTv0, digital (e.g., websites, e-mail, television commercials, social media) was the most commonly coded category. Online videos, although not a MoDTv0 category, were also identified. Human MODs were the second most commonly coded category. When available, participants seemed to prefer in-person and face-to-face delivery, but accepted digital face-to-face delivery (e.g., video calls) when in-person was not possible. Although not a MoDTv0 category, in-person presentations were also identified. Print MODs were the third most commonly identified category. Leaflets and public notices were identified within the MoDtv0, and magazines, newsletters, and advertisements in books were identified outside of the MoDTv0.

Table 4 Identified mode of delivery statements by MoDTv0 category.
Full size table

Recommendation development

Each organization representative felt that the study’s findings were relevant and useful for their organizations. The representatives then spoke to organizational capacity for delivering an intervention shaped by the study’s findings. Throughout the discussion, four intervention recommendations the organizations can use to support their aging membership were co-developed. Each representative ensured that all recommendations met feasibility (i.e. APEASE) criteria for each organization:

  1. (1)

    Educate/Build Capacity Among Organization Staff: Each organization’s representative emphasized the importance of staff and peer mentors being confident and prepared to support their aging clients. This recommendation addresses four intervention functions (education, training, modelling, environmental restructuring) and two policy categories (service provision, communication/marketing).

  2. (2)

    Educate/Build Capacity among Membership: Each organization’s representative wanted aging-specific knowledge and resources to be available and accessible to their membership. This recommendation addresses identical intervention functions and policy categories to Recommendation 1.

  3. (3)

    Advocate to/Educate the Government: Drawing from previous experience with provincial policy makers, each organization’s representative felt scientific evidence informed by lived experience should be shared with provincial government staff. With this awareness and knowledge, provincial government staff may be more inclined to implement policy changes that address the needs of people aging with SCI. This recommendation addresses two intervention functions (education, enablement) and has the potential to address the policy category legislation.

  4. (4)

    Partner with Other Disability Organizations: Each organization’s representative felt that continuing to partner with organizations for other disabilities may improve capacity to create and provide resources to membership and improve advocacy across all organizations involved. This recommendation addresses one intervention function (enablement) and one policy category (environmental/social planning).

Each recommendation could be implemented through multiple individual and organizational messengers, including peers, SCI organizations, provincial governments, and other disability organizations.

Discussion

Findings provide a theoretical understanding of the barriers and facilitators to participation in daily activities and social roles encountered by people aging with SCI. Analyses revealed that participation-focused interventions should aim to target one’s physical capability, physical opportunity, and social opportunity. Targeting these factors requires a multi-faceted approach that includes several intervention functions, policy categories, and MODs. Accordingly, the resulting co-developed, theory-based intervention recommendations target intrapersonal, interpersonal, and environmental barriers at the individual, organizational, and policy levels. To our knowledge, this research is the first to co-develop intervention recommendations for people aging with SCI where a rigorous and systematic behaviour change method for co-developing interventions was used. Findings also provide guidance for implementing interventions into practice.

Stage 1: understanding the behaviour

The identification of skills, ECR, and social influences as the most relevant TDF domains is an important implication that can guide the future development of theory-based, behaviour change interventions. While the influence of environmental and social domains on participation is not new to the SCI literature [30], many behaviour change interventions are grounded in the assumption that only one’s motivation must change to perform a behaviour, ignoring the impact of environmental and social domains [31]. As this study suggests people aging with SCI may already be motivated to participate in daily activities and social roles, future interventions that target participation in people aging with SCI may want to apply theories that address social and environmental domains, rather than using motivation-driven behaviour change theories.

Findings also highlight that both TDF versions may need to be considered when designing interventions that target participation in people aging with SCI. Experienced facilitator statements were coded as nature of the behaviour, a domain not included in the most recent TDF version [29]. As nature of the behaviour has previously been identified as relevant to physical activity and disability [17], using both TDF versions to understand any behaviour within a SCI context may be valuable.

Stage 2: intervention options

With the identification of multiple intervention functions and policy categories, a “one-size fits all” intervention may not be suitable in supporting participation in people aging with SCI. The need for multiple, multi-faceted interventions that target several intervention functions and policy categories aligns with previous efforts undertaken to enhance physical activity in people with SCI [32].

Stage 3: identify implementation options

Several messengers (health care professionals, peers, rehabilitation professionals, family/friends, SCI organizations) identified in this study align with messengers previously identified in research looking at the delivery of physical activity information to people with SCI [33, 34]. As Canadian provincial governments were identified as key organizational messengers, SCI organizations have an evidence-based rationale to advocate for the involvement of policy makers in intervention development. Findings also provide SCI organizations with the confidence to continue with the current practice of working alongside other disability organizations.

The need for multiple MODs was consistent with previous research [33, 34], re-iterating the importance of individualized and tailored interventions. Although MODs that aligned with previous research were identified (e.g., internet, brochures), a notable difference was a change in technology. This discrepancy suggests that research investigating MODs should be conducted more frequently to account for rapid and frequent changes in technology.

Recommendation development & future directions

The recommendations from this study are a promising first step for intervention co-development between researchers and community organizations aimed at enhancing participation in people aging with SCI. A next step is to translate the intervention recommendations into tangible knowledge tools/products [20]. Upon development, future research can assess the developed tools/products for efficacy, effectiveness, and fidelity. This study also provides an important future direction for programme development within SCI organizations as findings from this study can be used to inform new practices, enhancing an organization’s ability to support their aging membership.

Limitations

Despite best efforts with recruitment, our sample did not include women over 65 or individuals over 65 with cervical- or lumbar-level injuries. Second, as the study was specific to a Canadian context, more research is needed to understand barriers and facilitators to participation among people aging with SCI in other countries and contexts. Third, frequency counts were used to address TDF domains, intervention functions, and implementation options. While we aimed to interpret frequencies alongside broader interview data, interpretation beyond frequency counts may assist in finalized selections of relevant TDF domains, intervention functions, and implementation options. As the TDF is a framework and not a theory, it does not explain how domains link to one another. We aimed to mitigate this limitation by collectively using the TDF alongside the COM-B model. Finally, while the MoDTv0’s purpose is to allow for consistent communication of MODs, its application remains limited. The MoDTv0 may potentially be useful for future research, however, it cannot promote consistent communication of MODs if it is not widely available.

By using an integrated knowledge translation and behaviour change approach, this study provides an understanding of the barriers and facilitators to participation and theory-based recommendations for interventions aiming to enhance participation in people aging with SCI. In summary, multiple intervention options targeting physical capability, physical opportunity, and social opportunity should be delivered through multiple messengers and MODs to support people aging with SCI. Continued efforts are needed to develop the recommendations into real-world interventions.