Introduction
Advances in stroke management, and an increasing incidence of stroke earlier in life, has resulted in a growing number of people living with the effects of stroke long-term (Béjot et al., Reference Béjot, Daubail, Jacquin, Durier, Guy-Victor, Rouaud and Giroud2014; Lakshminarayan et al., Reference Lakshminarayan, Berger, Fuller, Jacobs, Anderson, Steffen and Luepker2014). For most people, spontaneous and experience-dependent recovery is facilitated by acute and post-acute rehabilitation and re-engagement in home and community activities (Brock et al., Reference Brock, Black, Cotton, Kennedy, Wilson and Sutton2009; Carod-Artal, Coral, Trizotto, & Moreira, Reference Carod-Artal, González-Gutiérrez, Egido Herrero, Horan and Varela de Seijas2002).
The International Classification of Functioning, Disability and Health (ICF; World Health Organization [WHO], 2001), proposes that function and disability are outcomes of interactions between health conditions and contextual factors. Although personal factors are highlighted as influential in the ICF; little consideration has been given to the relationship between these factors and post-stroke participation. This study examines the associations between personal factors, specifically demographics, self- and threat appraisals, and personality variables, and participation outcomes.
Participation post-stroke
Participation in everyday life is frequently disrupted, restricted and more challenging after stroke (Eriksson, Baum, Wolf, & Tabor Connor, Reference Eriksson, Baum, Wolf and Tabor Connor2013; Fallahpour, Jonsson, Jognataei, Nasrabadi Alireza, & Tham, Reference Fallahpour, Jonsson, Jognataei, Nasrabadi Alireza and Tham2013; Hartman-Maeir, Soroker, Ring, Avni, & Katz, Reference Hartman-Maeir, Soroker, Ring, Avni and Katz2007; Rochette, Desrosiers, Bravo, St-Cyr-Tribble, & Bourget, Reference Rochette, Desrosiers, Bravo, St-Cyr-Tribble and Bourget2007). Despite people returning, in part, to pre-stroke activities, physically demanding and community-based activities are retained less than home-based and sedentary activities (Tse, Lentin, Douglas, & Carey, Reference Tse, Lentin, Douglas and Carey2020). Participation restrictions have been identified to persist up to five-years post-stroke, particularly in the areas of autonomy outdoors, social life and relationships and work/education (Palstam, Sjödin, & Stibrant Sunnerhagen, Reference Palstam, Sjödin and Stibrant Sunnerhagen2019). Participation is considered amendable to change even in the absence of further body function recovery and is an intervention priority for people with long-term health conditions and disability (Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019; Whiteneck, Reference Whiteneck, Field, Jette and Martin2006). As an area of concern (Bergström, Guidetti, Tham, & Eriksson, Reference Bergström, Guidetti, Tham and Eriksson2017) and an unmet need post-stroke (Andrew et al., Reference Andrew, Kilkenny, Naylor, Purvis, Lalor, Moloczij and Cadilhac2014), participation warrants further empirical attention.
A stroke-specific scoping review identified participation as a troublesome construct due to diverse definitions and multiple measurement tools (Engel-Yeger, Tse, Josman, Baum, & Carey, Reference Engel-Yeger, Tse, Josman, Baum and Carey2018). Tse, Douglas, Lentin, and Carey (Reference Tse, Douglas, Lentin and Carey2013) identified 36 post-stroke participation measures, noting that the five most frequently used measures regularly covered community, social and civic life, domestic life and mobility domains, while learning and applying knowledge, general tasks and demands and communication were covered less often. A single measure of participation is considered unlikely to capture the breadth of this construct.
Despite the variability in conceptualisation and measurement, participation outcomes have consistently been associated with body functions/impairments, activity limitations, health conditions, level of independence and sociodemographic factors in adults’ post-stroke (Ahn & Hwang, Reference Ahn and Hwang2018; Carey, Matyas, & Baum, Reference Carey, Matyas and Baum2018; Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019). In addition to objective indices of functioning, the client’s perceived recovery has been identified as a significant predictor of participation and an important clinical consideration when determining barriers to post-stroke participation (Baseman, Fisher, Ward, & Bhattacharya, Reference Baseman, Fisher, Ward and Bhattacharya2010; Wolf & Koster, Reference Wolf and Koster2013).
Personal factors and participation post-stroke
Personal factors are ‘… the particular background of an individual’s life and living, and comprise features of the individual that are not part of a health condition or health states’ (WHO, 2001, p. 17). Personal factors can influence participation post-stroke (Norlander, Iwarsson, Jönsson, Lindgren, & Måsson Lexell, Reference Norlander, Iwarsson, Jönsson, Lindgren and Måsson Lexell2018) and may play a role in strengthening our understanding of functioning, disability and health (Geyh et al., Reference Geyh, Peter, Muller, Bickenbach, Kostanjsek, Ustun and Cieza2011). However, unlike other ICF components, there is no system for categorising personal factors (Geyh et al., Reference Geyh, Peter, Muller, Bickenbach, Kostanjsek, Ustun and Cieza2011; Hoyle, Gustafsson, Meredith, & Ownsworth, Reference Hoyle, Gustafsson, Meredith and Ownsworth2012). A systematic review identified 238 potential personal factors and concluded further standardisation of personal factors within the ICF was needed (Geyh et al., Reference Geyh, Peter, Muller, Bickenbach, Kostanjsek, Ustun and Cieza2011). Grotkamp and colleagues (Reference Grotkamp, Cibis, Nüchtern, von Mittelstaedt and Seger2012) subsequently proposed 72 personal factor categories organised into 6 chapters, but this has not been officially adopted.
Age and gender are personal factors often investigated in relation to post-stroke participation. A meta-analysis of factors associated with post-stroke participation identified that older age was most often associated with poorer participation (Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019). Gender accounted for greater variance in outcomes; however, the evidence was mixed, with women found to have both poorer and better participation than men across different studies (Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019). Other studies included in the meta-analysis also revealed no association between gender and participation (Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019).
The numerous and diverse nature of personal factors precludes consideration of all factors within a single study. Review of the empirical and theoretical literature informed selection of four additional personal factors for this study that have been previously overlooked in stroke research but are likely relevant to post-stroke participation; namely, self-concept, optimism, threat appraisal and adult attachment style. These factors depict how individuals relate to themselves (self-concept, optimism and attachment) and their social environment (threat appraisal and attachment). Two are cognitive appraisals that are potentially amendable to change (self-concept and threat appraisal) and are possibly responsive to therapeutic influence (modifiable), whereas two (optimism and attachment) reflect more enduring personality-related constructs (Badley, Reference Badley2006; Grotkamp, Cibis, Nüchtern, von Mittelstaedt, & Seger, Reference Grotkamp, Cibis, Nüchtern, von Mittelstaedt and Seger2012; Howe, Reference Howe2008). Many influences on participation previously examined are non-modifiable and not susceptible to intervention to improve participation outcomes. Selected factors were chosen as they are representative of individual features that do not relate to the health condition and investigation of their associations with participation outcomes may contribute to an improved understanding of life after stroke and guide the focus of novel interventions.
Self-concept
Studies have revealed that people tend to view themselves more negatively post-stroke. Ellis-Hill and Horn (Reference Ellis-Hill and Horn2000) identified that people typically saw themselves as less active, capable, in control, independent, interested and satisfied following stroke. Holding more negative views of self has also been associated with higher levels of depression (Lapadatu & Morris, Reference Lapadatu and Morris2019) which has established relationships with functional outcomes (Pohjasvaara, Vataja, Leppävuori, Kaste, & Erkinjuntti, Reference Pohjasvaara, Vataja, Leppävuori, Kaste and Erkinjuntti2001) and social participation (D’Alisa, Baudo, Mauro, & Miscio, Reference D’Alisa, Baudo, Mauro and Miscio2005; Schmid et al., Reference Schmid, Damush, Tu, Bakas, Kroenke, Hendrie and Williams2012). Negative changes in self-concept (i.e., negative ‘self-discrepancy’) have been reported to be indirectly associated with occupational gaps via an association with greater anxiety following traumatic brain injury (TBI) (Beadle, Ownsworth, Fleming, & Shum, Reference Beadle, Ownsworth, Fleming and Shum2020a). Beadle and colleagues (Reference Beadle, Ownsworth, Fleming and Shum2020a) postulated that anxiety related to lack of occupational engagement may contribute to more negative appraisals of the post- compared to pre-injury self. Self-reported or perceived level of recovery is another self-related construct for understanding how people view themselves, which may influence or be influenced by participation restrictions post-stroke (Baseman et al., Reference Baseman, Fisher, Ward and Bhattacharya2010; Wolf & Koster, Reference Wolf and Koster2013).
Dispositional optimism
Dispositional optimism relates to an individuals’ life outlook, future expectations and subsequent approach to the world (Scheier & Carver, Reference Scheier and Carver1985). Lower optimism has been linked with higher levels of depressive symptoms (Chung, Bakas, Plue, & Williams, Reference Chung, Bakas, Plue and Williams2016) and revealed as a mediating factor between mental health and perception of overall physical health post-stroke (Shifren & Anzaldi, Reference Shifren and Anzaldi2018). Optimism has also been related to better psychological functioning which can influence cognitive and functional outcomes post-TBI (Ramanathan, Wardecker, Slocomb, & Hillary, Reference Ramanathan, Wardecker, Slocomb and Hillary2011). In the only known study to focus on optimism and participation post-stroke, Schmid et al. (Reference Schmid, Damush, Tu, Bakas, Kroenke, Hendrie and Williams2012) identified positive associations between optimism and social role functioning, which is an important aspect of participation.
Threat appraisal
Threat appraisal, or the perception of an event or situation as potentially harmful, provides a means for understanding the relationship between anxiety and activity avoidance (Riley, Brennan, & Powell, Reference Riley, Brennan and Powell2004). Appraisal of task failure as a threat to one’s sense of self can result in anxiety and avoidance of situations that may end in failure (Goldstein, Reference Goldstein1952; Riley et al., Reference Riley, Brennan and Powell2004). Heightened threat appraisals after brain injury have been linked to greater use of maladaptive coping strategies, such as task avoidance or withdrawal, which restricts participation in valued pre-injury activities (Gracey, Evans, & Malley, Reference Gracey, Evans and Malley2009). People with acquired brain injury experience threat associated with activities related to ‘personal safety’, ‘dealing with people’, ‘doing things’ and ‘awkward situations’ (Riley et al., Reference Riley, Brennan and Powell2004). Individuals with more negative self-discrepancy following TBI reported higher threat appraisals (r = −0.50, p < 0.001) (Beadle, Ownsworth, Fleming, & Shum, Reference Beadle, Ownsworth, Fleming and Shum2020b). There is limited information about post-stroke threat appraisals. One study demonstrated that although perceptions of threat related to stroke decreased in the first 6 months after stroke, a greater initial perceived threat was associated with lower participation at 6 months (Rochette, Bravo, Desrosiers, St-Cyr Tribble, & Bourget, Reference Rochette, Bravo, Desrosiers, St-Cyr Tribble and Bourget2007). These authors suggested that cognitive appraisals may be shaped or influenced by experiences, making them potentially responsive to therapeutic intervention. Thus, threat appraisals post-stroke may be susceptible to modification, possibly resulting in improved participation (Hoyle et al., Reference Hoyle, Gustafsson, Meredith and Ownsworth2012). While Rochette, Bravo et al.’s (Reference Rochette, Bravo, Desrosiers, St-Cyr Tribble and Bourget2007) study included generalised post-stroke appraisals measured by the Stress Appraisal Measure (Peacock & Wong, Reference Peacock and Wong1990), the present study focuses on threat appraisals related to everyday activities and situations measured by the Appraisal of Threat and Avoidance Questionnaire (ATAQ; Riley et al., Reference Riley, Brennan and Powell2004).
Adult attachment style
Adult attachment styles, are individualised schemas/internal working models, developed from infancy, that influence how an individual perceives themselves, others and the world (Bowlby, Reference Bowlby1988, Reference Bowlby1969; Meredith, Strong, & Feeney, Reference Meredith, Strong and Feeney2005; Mikulincer, Reference Mikulincer1995). While these styles have not been widely researched in association with post-stroke participation, attachment theory has been increasingly applied within the medical field (Hunter & Maunder, Reference Hunter and Maunder2001; Maunder & Hunter, Reference Maunder and Hunter2001; Strauss & Brenk-Franz, Reference Strauss, Brenk-Franz, Hunter and Maunder2016). Relationships have been demonstrated between adult attachment and, coping and adjustment in chronic conditions (Schmidt, Nachtigall, Wuethrich-Martone, & Strauss, Reference Schmidt, Nachtigall, Wuethrich-Martone and Strauss2002), including cancer (Cicero, Lo Coco, Gullo, & Lo Verso, Reference Cicero, Lo Coco, Gullo and Lo Verso2009), chronic pain (Meredith, Ownsworth, & Strong, Reference Meredith, Ownsworth and Strong2008), diabetes (Ciechanowski et al., Reference Ciechanowski, Russo, Katon, Von Korff, Ludman, Lin and Bush2004) and HIV/AIDS (Turner-Cobb et al., Reference Turner-Cobb, Gore-Felton, Marouf, Koopman, Kim, Israelski and Spiegel2002). Fearful and dismissing attachment styles have been linked with threat appraisal (Meredith et al., Reference Meredith, Strong and Feeney2005) which, as previously discussed, may be associated with participation restrictions. Attachment warrants further investigation in relation to post-stroke participation because of the potential links with coping, adjustment and threat appraisal. Although unlikely amenable to change as long-standing personality constructs, attachment style and optimism, are considerations that may influence clinical reasoning and support tailoring interventions specific to individual needs.
Present study
The aim was to examine associations between demographic, self- and threat appraisal and personality variables and, post-stroke participation. It was hypothesised that higher levels of participation may be associated with younger age, higher levels of optimism, less difference between pre- and post-stroke self-concept (i.e., lower self-discrepancy), lower levels of threat appraisal and a more secure attachment style. Due to the inconsistent findings in the literature, a hypothesis for gender was not postulated. Finally, level of perceived recovery has been identified within the literature as a significant predictor of participation (Baseman et al., Reference Baseman, Fisher, Ward and Bhattacharya2010; Wolf & Koster, Reference Wolf and Koster2013) and was recorded to explore associations with post-stroke participation further.
Methods
Design
This exploratory cross-sectional study represents the quantitative portion of a larger explanatory sequential mixed methods study (Creswell & Plano Clark, Reference Creswell and Plano Clark2011). Ethical approval was granted by The University of Queensland’s Behavioural and Social Sciences Ethical Review Committee (Project no. 2012001037/Project identifier 2012/HE001037).
Participants
Participants were aged 18 or above, had experienced stroke (one or multiple events) and had returned to living in the community within Australia (either independently or assisted). Potential participants were excluded if inclusion criteria were not met, but were not excluded based on communication difficulties, as people with these challenges post-stroke are often under-represented in stroke-related research (Brady, Fredrick, & Williams, Reference Brady, Fredrick and Williams2012). Recruitment used convenience and snowball sampling methods from November 2014 to December 2017 inclusive, through a variety of means, including print, social media and community-based presentations. Informed consent was obtained and all participants were asked to indicate their willingness to participate in further research opportunities. The structured purpose-designed survey was completed and returned via an online website, email or in hardcopy via mail, based on participant preference. Participants could complete the survey independently or have someone complete it on their behalf.
Measures
The survey included demographic (e.g., age and gender) and general health questions and a variety of questionnaires regarding participation and the selected personal factors. Two self-report measures of participation were chosen to gain a more comprehensive insight into the construct of participation, both in terms of content examined and ways in which the construct is measured.
Participation measures
Community Integration Questionnaire
The Community Integration Questionnaire (CIQ), the first measure of participation, is a 15-item instrument that examines participation in home (five items), social (six items) and productive (four items) activities (Dalemans, de Witte, Beurskens, van den Heuvel, & Wade, Reference Dalemans, de Witte, Beurskens, van den Heuvel and Wade2010; Willer, Ottenbacher, & Coad, Reference Willer, Ottenbacher and Coad1994; Willer, Rosenthal, Kreutzer, Gordon, & Rempel, Reference Willer, Rosenthal, Kreutzer, Gordon and Rempel1993). Twelve items are scored on a 3-point scale which identifies who completes the task (0 = someone else, 1 = yourself and someone else, and 2 = yourself alone) or how often the task is completed (0 = Never, 1 = 1–4 times, 2 = 5 or more times). The remaining three items, relating to work, school and volunteering, are examined in combination and assigned a score between 0 and 5. The overall CIQ score (range: 0–29) is a summation of the scores with a higher score indicating a higher level of integration. The measure can be completed by a proxy (Kuipers, Kendall, Fleming, & Tate, Reference Kuipers, Kendall, Fleming and Tate2004; Salter, Foley, Jutai, Bayley, & Teasell, Reference Salter, Foley, Jutai, Bayley and Teasell2008; Willer et al., Reference Willer, Ottenbacher and Coad1994). The reliability and validity of the CIQ has been established with adults with physical disabilities and TBI (Corrigan & Deming, Reference Corrigan and Deming1995; Hirsh, Braden, Craggs, & Jensen, Reference Hirsh, Braden, Craggs and Jensen2011; Willer et al., Reference Willer, Ottenbacher and Coad1994, Reference Willer, Rosenthal, Kreutzer, Gordon and Rempel1993). Internal consistency for the total CIQ score was adequate (α = 0.71) for the present sample.
Stroke Impact Scale Version 3.0
The Stroke Impact Scale Version 3.0 (SISv3) examines the impact of stroke across eight domains (Duncan, Bode, Lai, & Perera, Reference Duncan, Bode, Lai and Perera2003; Lin et al., Reference Lin, Fu, Wu, Hsieh, Chen and Lee2010). A final item assesses the person’s global perception of recovery (0 = no recovery to 100 = full recovery). The original Stroke Impact Scale demonstrated fair to good convergent validity (Doyle et al., Reference Doyle, McNeil, Bost, Ross, Wambaugh, Hula and Mikolic2007; Kwon et al., Reference Kwon, Duncan, Studenski, Perera, Lai and Reker2006) and robust psychometric properties for the physical and participation domains (Duncan et al., Reference Duncan, Bode, Lai and Perera2003). Support for proxy completion has been reported (Carod-Artal et al., Reference Carod-Artal, Coral, Trizotto and Moreira2009; Duncan et al., Reference Duncan, Lai, Tyler, Perera, Reker and Studenski2002). The participation domain was the second measure of participation in the present study. Participants rated how much during the past four weeks they have been limited in work, social activities, quiet recreation, active recreation, role as a family member/friend, spiritual or religious activities, ability to control life as they wish and ability to help others. A summative score was produced for this domain and transformed into a scale score (range: 0–100) by using the formula: [(actual raw score – lowest possible raw score)/possible raw score] × 100. Internal consistency for the present sample in this domain was good (α = 0.89). The global perception of recovery item (range: 0–100) was also recorded to enable consideration of the associations between perceived recovery and post-stroke participation outcomes.
Personal factor measures
Head Injury Semantic Differential III
The Head Injury Semantic Differential Scale III (HISD-III) measures self-concept and self-discrepancy. Participants rate themselves on 18 bipolar adjective/attribute pairs (e.g., Bored/Interested, Unhappy/Happy) on a 7-point scale (1 = negative pole and 7 = positive pole), according to past-self (i.e., six months prior to injury), present-self (i.e., over the past two weeks) and future-self (i.e., expectation for one year into the future) (Ownsworth, Reference Ownsworth2014; Tyerman, Reference Tyerman1997). Total scores can be calculated for each self (range: 18–126), allowing for comparison between perceptions of past-, present- and future-selves (Ownsworth, Reference Ownsworth2014; Tyerman, Reference Tyerman1997, Reference Tyerman, Tyerman and King2008), and perceived change in self-concept (Reddy, Ownsworth, King, & Shields, Reference Reddy, Ownsworth, King and Shields2015). For this study, only past- and present-self scales were administered to align with pre- and post-stroke selves. Self-discrepancy was calculated by subtracting the pre-stroke from the post-stroke score. A larger negative discrepancy score represented a more positively viewed past-self than present-self. The HISD-III and its prior iterations have demonstrated strong reliability for past- and present-selves and have been used with people with TBI and stroke (Beadle et al., Reference Beadle, Ownsworth, Fleming and Shum2020a, Reference Beadle, Ownsworth, Fleming and Shum2020b; Carroll & Coetzer, Reference Carroll and Coetzer2011; Lapadatu & Morris, Reference Lapadatu and Morris2019; Reddy et al., Reference Reddy, Ownsworth, King and Shields2015). Internal consistency in the present sample was good for past-self (α = 0.87) and excellent for present-self (α = 0.96).
Relationship Questionnaire
The Relationship Questionnaire (RQ) is a self-report measure of adult attachment style comprising two parts (Bartholomew & Horowitz, Reference Bartholomew and Horowitz1991). First, participants identified which one of the four attachment styles best describes them. Second, participants rated how well the description of each style corresponds to their general relationship style on a 7-point Likert Scale (1 = not at all like me to 7 = very much like me). While only displaying moderate test–retest reliability (Herzberg et al., Reference Herzberg, Hammen, Burge, Daley, Davila and Lindberg1999; Scharfe & Bartholomew, Reference Scharfe and Bartholomew1994), the construct validity of the four-category model on which the RQ is based has been demonstrated (Bartholomew & Horowitz, Reference Bartholomew and Horowitz1991; Griffin & Bartholomew, Reference Griffin and Bartholomew1994) and moderate to high correlations have been found between the RQ and the Relationship Styles Questionnaire (Bäckström & Holmes, Reference Bäckström and Holmes2001; Griffin & Bartholomew, Reference Griffin and Bartholomew1994).
Appraisal of Threat and Avoidance Questionnaire
The ATAQ is a 41-item questionnaire that documents threat appraisals and associated activity avoidance in four categories: dealing with people, doing things, personal safety and awkward situations (Riley et al., Reference Riley, Brennan and Powell2004). Items are presented as statements, and participants are asked to reflect on the previous month and respond with yes or no. If yes, participants are asked to identify, whether they had avoided meeting people/going out/doing things due to the original appraisal (Riley et al., Reference Riley, Brennan and Powell2004). Threat appraisal and avoidance scores are calculated overall and for all four categories (Riley et al., Reference Riley, Brennan and Powell2004); only total threat appraisal scores (range: 0–41) were used in this study. Satisfactory internal consistency has been established for overall threat appraisal and avoidance, and for the first three categories (Riley et al., Reference Riley, Brennan and Powell2004). Internal consistency for the threat appraisal index in the present sample was excellent (α = 0.92).
Life Orientation Test – Revised
The Life Orientation Test – Revised (LOT-R) is a measure of participants’ generalised sense of optimism. It is comprised of three ‘optimism’ items, three ‘pessimism’ items and four ‘filler’ items. Each item is rated on a 5-point Likert scale (0 = strongly disagree to 4 = strongly agree) (Scheier et al., Reference Scheier, Carver and Bridges1994). Totalled scores range from pessimistic (lowest possible = 0) to optimistic (highest possible = 24) (i.e., range: 0–24) (Ramanathan et al., Reference Ramanathan, Wardecker, Slocomb and Hillary2011). The LOT-R has displayed internal consistency at an acceptable level and moderate correlations with related scales of neuroticism, self-esteem, self-mastery and trait anxiety (Scheier et al., Reference Scheier, Carver and Bridges1994). Internal consistency for the present sample was good (α = 0.80).
Data analysis
Each returned survey was inspected for missing or unclear responses and, where possible, clarification regarding responses were sought. Data were analysed using the IBM Statistical Package for Social Sciences (SPSS) Version 27 and as an exploratory, hypothesis-generating study, a power calculation for desired sample size was not conducted. Descriptive analyses and screening explored whether dependent variables (CIQ and SISv3 participation) met assumptions for parametric testing (i.e., Shapiro–Wilk test). Associations between independent variables (age, gender, self-concept/self-discrepancy, threat appraisal, optimism, adult attachment style and perceived recovery) and participation variables were determined using univariate linear regression and correlation (see supplementary material) analyses. Scatter plots with lines of best fit examined the directionality of the associations indicated by the univariate regression analyses. Time since stroke, a potential confounding variable, was investigated using correlations to determine whether it should be considered in further analyses. Two models, one for CIQ participation and one for SISv3 participation, were explored using multiple linear regression analyses. All independent variables that demonstrated significant associations (p < 0.05) with each of the dependent variables in the univariate regressions were considered for entry into the relevant models. Variables included in the final models were chosen based on relevance to the aim of the study and the total variability they described in the univariate analyses. Both multiple linear regression analyses were examined for outliers. Outliers (studentised residuals > 2.6) were removed one at a time (i.e., largest studentised residuals first) and the multiple regression models were re-run. Two outliers were removed from the CIQ participation model and one outlier was removed from the SISv3 participation model. The final models were checked for multicollinearity, homoscedasticity and normal distribution of residuals before finalisation.
Results
Descriptive and sample characteristics
Seventy-two people expressed interest and ten were excluded due to no diagnosis of stroke (n = 1), not living in the community (n = 4), completing less than 70% of the survey (n = 4) or completing the survey twice (n = 1). Sixty-two participants were subsequently included in data analysis, and demographics and stroke and health characteristics are shown in Tables 1 and 2, respectively.
Table 1 Demographics of Participants
Note.
† measured in months; Freq., Frequency; n, number of participants; SD, standard deviation; %, percentage.
Table 2 Stroke and Health Characteristics of Participants
Note. Freq., frequency; n, number of participants; %, percentage.
Three participants (5%) reported having expressive aphasia, and two (3%) reported having both expressive and receptive aphasia. Nine (14.5%) indicated that another person completed the survey on their behalf. Of those nine, seven (78%) indicated that they had neither expressive and/or receptive aphasia, one (11%) identified they had expressive aphasia and one (11%) identified they had both expressive and receptive aphasia.
CIQ participation data were normally distributed (p = 0.33) while SISv3 participation data were not (p = 0.007) pre-multiple regression. Regression diagnostics showed that residuals for both CIQ and SISv3 participation data were normally distributed (p = 0.81, p = 0.44 respectively). Thus, parametric tests were utilised for all regression analyses. Time since stroke did not show significant correlations with either CIQ participation (r(49) = 0.08, p = 0.59) or SISv3 participation (rs(58) = 0.14, p = 0.29) and was not included in further analyses.
Univariate regression analyses
CIQ participation univariate regression analyses
Self-discrepancy and perceived recovery and were significantly positively related with CIQ participation (Table 3). Age was significantly negatively related with CIQ participation. A significant association was seen between CIQ participation and gender, such that women were more likely to report higher participation than men (female M = 16.36; male M = 13.36). No other variables were significantly associated with CIQ participation.
Table 3 Univariate Analyses Between Participation (CIQ) and Independent Variables
Note. B, unstandardised beta; CIQ, Community Integration Questionnaire; df, degrees of freedom; F, F value; p, Significance; R 2 , R squared; t, t value; 95% LB = Lower Bound of 95% Confidence Interval for B; 95% UB = Upper Bound of 95% Confidence Interval for B.
SISv3 participation univariate regression analyses
Secure attachment, self-discrepancy and perceived recovery demonstrated significant positive associations with SISv3 participation (Table 4). Threat appraisal, and preoccupied attachment were significantly negatively related with SISv3 participation.
Table 4 Univariate Analyses Between Participation (SISv3) and Independent Variables
Note. B, unstandardised beta; df, degrees of freedom; F, F value; p, significance; R 2 , R squared; SISv3, stroke impact scale version 3.0; t, t value; 95% LB = Lower Bound of 95% Confidence Interval for B; 95% UB = Upper Bound of 95% Confidence Interval for B.
Multiple regression analyses
Results of multiple regression analyses (Table 5) identified relevant personal factors and the associated percentage of explained variability for CIQ and SISv3 participation. In the multiple variable modelling, age (p = 0.02), gender (p = 0.008), self-discrepancy (p = 0.02) and perceived recovery (p < 0.001) were significantly associated with CIQ participation. Only age had an inverse relationship indicating that increased age was associated with decreased CIQ participation, while increases in perceived recovery were associated with increased CIQ participation. If pre-stroke self-concept was greater than post-stroke self-concept, greater discrepancy between the two was associated with decreased participation. If post-stroke self-concept was greater than pre-stroke self-concept, greater discrepancy was associated with increased participation. Gender was positively associated with CIQ participation, indicating that women were more likely to report higher participation than men (female M = 17.22; male M = 12.15). The multiple variable modelling for SISv3 participation demonstrated similar associations between SISv3 participation and self-discrepancy (p = 0.002) and perceived recovery (p < 0.001).
Table 5 Summary of Final Multiple Regression Analyses for Variables Associated with Participation as Measured by the CIQ (n = 36) and SISv3 (n = 41)
Note. B, unstandardised beta; CIQ, Community Integration Questionnaire; F, F value; SE B, standard error for the unstandardised beta; SISv3, stroke impact scale version 3.0; β, standardised beta; ΔR 2 , adjusted R squared value.
* p < 0.05.
** p < 0.01.
*** p < 0.001.
Correlations between participation measures
In response to the different findings revealed for the two participation measures, a Spearman’s rank-order correlation was conducted to investigate the relationship between the two participation measures. The result, rs = 0.34, p = 0.02, demonstrated a significant but weak correlation between the measures.
Discussion
The study aimed to identify demographic, self- and threat appraisal and personality variables that were strongly associated with participation for people living in the community post-stroke. Findings at a univariate level demonstrated that a range of personal factors, including age, gender, self-concept, threat appraisal, secure attachment and pre-occupied attachment were variably associated with CIQ and SISv3 participation. However, at a multivariate level, self-concept, age and gender were the only personal factors related to participation outcomes. Results highlight the potential need for greater consideration of the influence of personal factors on participation outcomes.
Consistent with the hypothesis, self-discrepancy demonstrated a strong positive association with both CIQ and SISv3 participation. A low self-discrepancy score was associated variably with participation dependent on which self-concept was scored higher. When pre-stroke self-concept was greater than post-stroke self-concept, a lower self-discrepancy score was associated with increased participation. When post-stroke self-concept was greater than pre-stroke self-concept a lower self-discrepancy was associated with decreased participation. Whilst there is no established causal relationship, findings indicate that people who perceive their pre-stroke self to be better than their post-stroke self and have higher self-discrepancy may be at higher risk of poor participation outcomes. Equally, people with low participation may be susceptible to a greater change in self-concept. While previous literature has indicated that people post-stroke tend to experience challenges in participation (Eriksson et al., Reference Eriksson, Baum, Wolf and Tabor Connor2013; Palstam et al., Reference Palstam, Sjödin and Stibrant Sunnerhagen2019; Wolf & Koster, Reference Wolf and Koster2013) and hold more negative views of themselves (Ellis-Hill & Horn, Reference Ellis-Hill and Horn2000; Lapadatu & Morris, Reference Lapadatu and Morris2019) this study is the first known to demonstrate the association between these two constructs post-stroke.
Associations were expected between younger age and increased participation, while no hypothesis for gender was postulated. Results revealed that younger age and female gender were associated with increased participation as measured by the CIQ; however, no association was found with the SISv3. Variation in associations with gender were consistent with previous literature and warrant additional attention in stroke research. The lack of association between age and SISv3 participation was unexpected given that older age has consistently been associated with poorer participation outcomes in earlier research (Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019). Differences in the focus and structure of the two participation measures, discussed below, could have contributed to differences seen in findings. As a range of participation measures exist, further investigation into the construct, its operationalisation in measures and associations with personal factors would be valuable.
The hypotheses that increased optimism, decreased threat appraisal and secure attachment would be associated with increased participation were partially supported. In contrast with previous studies which have found associations between optimism and better psychological functioning (Ramanathan et al., Reference Ramanathan, Wardecker, Slocomb and Hillary2011) and social role functioning (Schmid et al., Reference Schmid, Damush, Tu, Bakas, Kroenke, Hendrie and Williams2012), optimism was not directly associated with either CIQ or SISv3 participation at a univariate level. Variations in results may relate to different participant samples or differences in outcome variables measured. While a direct association was not evident, optimism may indirectly impact on participation through mediating factors, and this would benefit from further investigation.
Attachment variables were not directly associated with CIQ participation. In contrast, secure attachment was positively associated, and preoccupied attachment was negatively associated, with SISv3 participation at a univariate level. Associations with secure attachment were anticipated based on previous research (Meredith et al., Reference Meredith, Strong and Feeney2005), however, identified relationships between threat appraisal and insecure attachment had been seen for fearful and dismissing attachment rather than preoccupied attachment. Differences in results may relate to differences in participant samples or different dependent variables of interest.
As expected, people who reported higher levels of threat appraisal were more likely to report lower levels of participation as measured by the SISv3; however, this association did not persist in multiple variable analyses. It is possible that, given the theoretical association between threat appraisal and self-concept (Goldstein, Reference Goldstein1952; Riley et al., Reference Riley, Brennan and Powell2004), self-concept could have accounted for the variance in participation attributed to threat appraisal.
Perceived recovery, a possible alternative approach for examining view of self, was positively associated with both CIQ and SISv3 participation. Recovery from stroke incorporates inter-linked physical, psychological and social features and is defined at an individual level by people with stroke, rather than determined by population norms or other external standards (Dowswell et al., Reference Dowswell, Lawler, Dowswell, Young, Forster and Hearn2000). A person’s perception of their recovery, if poor, is a possible barrier to participation after mild stroke (Wolf & Koster, Reference Wolf and Koster2013) and thus may be a barrier to participation for the wider stroke population. Eriksson and colleagues (Reference Eriksson, Baum, Wolf and Tabor Connor2013) suggested that perceived recovery might be related to participation in two ways. Firstly, people post-stroke might perceive recovery as necessary before participating in activities. Conversely, as people become aware of increased engagement in pre-stroke activities, they may perceive a higher level of recovery. Regardless of directionality, the association between recovery and participation warrants further investigation. Future work may also examine the difference between ‘perceived recovery’ of the individual and ‘measured recovery’ determined by clinical assessment in relation to post-stroke participation outcomes.
Previous literature has identified the challenges in conceptualising and operationalising participation (Dijkers, Reference Dijkers2010; Engel-Yeger et al., Reference Engel-Yeger, Tse, Josman, Baum and Carey2018; Whiteneck, Reference Whiteneck2010). While both participation measures were associated with self-concept and perceived recovery, variations in associations were also evident suggesting that there were differences in the operationalisation of participation across the two measures (Dijkers, Reference Dijkers2010; Engel-Yeger et al., Reference Engel-Yeger, Tse, Josman, Baum and Carey2018; Whiteneck, Reference Whiteneck2010). The CIQ measures how people participate in home and community activities and how often particular activities are completed. In contrast, the SISv3 measures how often a person has been limited in participating in specific activities. Despite some similarity between activities examined there are also notable differences. The CIQ has a greater focus on home-related activities and includes education which is not addressed in the SISv3. The SISv3, conversely, includes questions about spirituality/religion and ‘ability to control your life as you wish’, which are not present in the CIQ. These observed differences, supported by the relatively weak correlation between the measures, suggest that the consistent associations between both measures and self-concept and perceived recovery are particularly important.
Despite challenges in operationalisation, associations between participation and personal factors are important considerations for practice. Modifiable personal factors, like self-concept, may be susceptible to clinical intervention; thus, clinically based research could consider whether interventions targeted at these factors can improve participation outcomes. Likewise, interventions supporting improvements in participation may positively influence modifiable personal factors and are also worthy of further exploration. The increasing number of people living in the community post-stroke (Béjot et al., Reference Béjot, Daubail, Jacquin, Durier, Guy-Victor, Rouaud and Giroud2014; Lakshminarayan et al., Reference Lakshminarayan, Berger, Fuller, Jacobs, Anderson, Steffen and Luepker2014) and greater focus on reducing participation restrictions, highlights the importance of research to understand and address this largely unmet need (Andrew et al., Reference Andrew, Kilkenny, Naylor, Purvis, Lalor, Moloczij and Cadilhac2014).
Strengths and limitations
The original sample was small (n = 62), and further reduced in some analyses due to missing data. The collection of data using a self-report, cross-sectional research design can increase risk of bias due to common method variance. Despite this, sense of self measurement can be useful in gaining insight into people’s experience of their condition, which can lead to more effective intervention (Ownsworth, Reference Ownsworth2014). Sampling bias may also be present as participants who volunteered may be more likely to represent people with higher participation levels and most participants lived in metropolitan regions of Queensland or Northern New South Wales. The HISD-III assesses a specific range of behavioural and emotional attributes, does not incorporate all domains relevant to self-concept, and reliance on retrospective data for establishing past-self is also a limitation. Data were not collected on stroke severity, a variable commonly associated with poorer participation outcomes (Ezekiel et al., Reference Ezekiel, Collett, Mayo, Pang, Field and Dawes2019), due to the explicit focus on personal factors rather than the health condition. The inclusion of people with aphasia and the different data collection methods promoted participation of a diverse, representative group. Participants indicated if someone else completed the survey on their behalf, but it is unknown if a communication partner recorded responses or if responses were completed by proxy. Proxy completion has been found acceptable for both the CIQ and SISv3 (Carod-Artal et al., Reference Carod-Artal, Coral, Trizotto and Moreira2009; Kuipers et al., Reference Kuipers, Kendall, Fleming and Tate2004; Willer et al., Reference Willer, Ottenbacher and Coad1994), but has not been established for the included self- and threat appraisal and personality variable measures. Finally, it is unknown if completion of the chosen questionnaires in the online format impacted the responses.
Despite these limitations, the study findings are a preliminary contribution towards understanding the relationships between demographic, self- and threat appraisal and personality variables and, participation post-stroke. This study adds to the expanding knowledge-base to improve understanding of factors that influence participation in the community after stroke. Future research may benefit from employing methods of data collection that increase participant numbers, decrease potential for missing data and minimise potential for biases, thereby improving power and confidence in findings.
Conclusion
This study provides preliminary evidence of associations between post-stroke participation and demographic, self-and threat appraisal and personality variables. While nearly all included personal factors demonstrated associations with participation post-stroke, these associations varied based on the way that participation was conceptualised and subsequently operationalised in each participation measure. Nevertheless, demographics, self-appraisals (i.e., self-concept and perceived recovery), explained a significant portion of the variability in post-stroke participation. This supports the value of further research to investigate associations between personal factors and participation post-stroke which may, in turn, contribute to an improved understanding of life after stroke and support future development of participation-focused interventions.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/BrImp.2022.31
Acknowledgments
This study was completed as part of a Doctor of Philosophy program at The University of Queensland. The authors thank the study’s participants, who kindly volunteered their time to provide the valuable data for this study.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sector.
Conflicts of interest
Melanie Hoyle, has no conflicts of interest to disclose. Pamela Meredith has no conflicts of interest to disclose. Tamara Ownsworth has no conflicts of interest to disclose. Asaduzzaman Khan has no conflicts of interest to disclose. Louise Gustafsson has no conflicts of interest to disclose.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Open access
