Caregivers of individuals with borderline personality disorder: The relationship between leading caregiver interventions and psychological distress/positive mental well-being
Abstract
Burden and psychological distress are higher in informal caregivers (ICs) of people with severe emotional and behavior dysregulation who have been given a diagnosis of borderline personality disorder (BPD) compared with non-caregivers. The current cross-sectional study examines the difference in outcomes of ICs of people with BPD who have received the intervention Family Connections (FC) and who also led interventions for other caregivers (caregiver-leaders) compared with those who have attended FC but not led caregiver interventions (non-leader-FC participants). The sample for this research is from a larger study (Hayes et al., 2023, Borderline Personality Disorder and Emotion Dysregulation, 10, 31). Data for 347 participants who self-reported receiving FC and completed the McLean Screening Instrument for BPD-Carer Version, the Brief COPE, the Multidimensional Scale of Perceived Social Support, the Kessler Psychological Distress scale, the WHO-5 Well-being Index, and the Coronavirus Anxiety Scale were analyzed. The results found that being a caregiver-leader was associated with higher positive mental well-being and lower psychological distress compared with non-leader-FC participants. Being a caregiver-leader was also associated with significantly greater use of the coping strategy of positive reframing and lower use of behavioral disengagement and self-blame than non-leader-FC participants. The study provides preliminary evidence that for those who have received FC, becoming an intervention leader is associated with better outcomes than caregivers who do not become leaders and provides support for caregiver-led rollout of FC across services.
Borderline personality disorder (BPD) is a characterized by dysregulated emotions and behavior (American Psychiatric Association, 2022). It is associated with high rates of self-harm and suicidality (Kirtley et al., 2019; Paris & Zweig-Frank, 2001). BPD is also referred to as emotionally unstable personality disorder (EUPD; World Health Organization, 2016) and personality disorder, borderline pattern (World Health Organization, 2019). Many countries recommend community-based specialist treatment with occasional brief crisis inpatient admissions (National Health and Medical Research Council, 2013; National Institute for Health and Care Excellence [NICE], 2009), leading to reliance on community support from informal caregivers (ICs) (family members and significant others within the person's support system) for individuals with BPD (Acres et al., 2019; Hoffman et al., 2007).
Caregiver research suggests a transactional and interdependent caregiver-care–recipient relationship (Bidwell et al., 2017). This is of particular importance for ICs of people with BPD. Theoretical models highlight the prominent role of relationships in BPD etiology and maintenance, for example, the Biosocial model's ‘invalidating environment’ (Linehan, 1993). Interventions for people with BPD that include family members are associated with positive outcomes for the person with BPD (Sutherland et al., 2020), signifying the potentially important role of caregivers in the well-being of people with BPD.
Unfortunately, research indicates that ICs of people with BPD evidence high levels of burden, grief, and psychological distress (Bailey & Grenyer, 2015; Kirtley et al., 2019; Seigerman et al., 2020). Supporting a person with BPD is associated with changes in caregiver life roles, reduced attention to IC personal well-being (Buteau et al., 2008; Dunne & Rogers, 2013; Meshkinyazd et al., 2020), and feelings of hopelessness, shame, guilt, sadness, and despair (Buteau et al., 2008; Ekdahl et al., 2011; Giffin, 2008; Kay et al., 2018). National guidelines outline that ICs of people with BPD may require support (National Health and Medical Research Council, 2013; NICE, 2009). Emerging evidence suggests that interventions developed for ICs of people with BPD may provide some improvement in caregiver outcomes (Guillén et al., 2021; Sutherland et al., 2020). In a systematic review, Sutherland et al. (2020) concluded that compared with five other interventions, Family Connections (FC) had the greatest evidence of efficacy to date. Guillén et al. (2021) also systematically reviewed 11 different interventions and described FC as the most advanced concerning content and program delivery.
Family Connections (Hoffman et al., 2005) is a 12-week group psychoeducation and skills training program for ICs of people with BPD, comprising six modules including psychoeducation about the characteristics and etiology of BPD and dialectical behavior therapy-informed intra and interpersonal skills to improve emotion regulation, well-being, and family relationships (Fruzzetti et al., 2023; Hoffman et al., 2005). An initial study found pre- to post-intervention reduction in burden and grief, increased mastery, and no change in depression scores; results remained or improved after 3 months (Hoffman et al., 2005). Hoffman et al. (2007) replicated the findings, including improvement in depression scores, and results were sustained or improved, except for mastery scores, at 3-month follow-up. Flynn et al. (2017) compared the treatment group to optimized treatment-as-usual and again reported post-treatment improvements in burden, grief, depression, and mastery. Liljedahl et al. (2019) found improvements in family functioning as well as burden and psychological distress post-treatment and at 6–7-month follow-up. These results provide preliminary evidence for FC effectiveness and efficacy for ICs of people with BPD.
Family connections was initially a peer-delivered intervention. Trained family members (ICs) delivered the FC curriculum and provided skill-use examples from their own lived experience (Hoffman et al., 2005). Currently, FC programs may be IC-led, delivered by mental health clinicians, or team-led (co-facilitated by an IC and a clinician). In this study, ICs who have received FC and delivered FC are referred to as caregiver-leaders.
While there is currently no published research, anecdotal reports from caregiver-leaders suggested that leading FC programs may have additional benefits above participating in the program for these individuals (Courey et al., 2021), thus identifying a research avenue into factors that may improve outcomes for the IC, and due to the possible transactional nature of the relationship between the IC and the person with BPD, the person with BPD. This is of interest as there remains a scarcity of research on factors associated with differential outcomes for this caregiving population. Accordingly, the first study's aim was to investigate if being an IC who is also a caregiver-leader is associated with different outcomes compared with ICs who have received FC and have not been intervention leaders (non-leader–FC participants).
The current study is part of a larger study that investigated the relationship between IC social support, coping strategies, and psychological distress, and positive mental well-being (Hayes et al., 2023), informed by the stress-process model of caregiving (Pearlin et al., 1990). Details of the measures chosen, and a summary of the literature are detailed in Hayes et al. (2023). The authors found that coping strategies were important predictors of IC psychological distress and positive mental well-being. Based on this finding, a second aim herein was to investigate if the relationship between being a caregiver-leader and IC outcomes was partially mediated by the use of coping strategies.
METHOD
This study used a cross-sectional design with online survey data collection.
Participants
In this study, the term IC refers to a person who identified as the primary caregiver and family member or significant other of a person with BPD. Participants met the following inclusion criteria: (a) identified as a primary caregiver of a person with BPD/EUPD, (b) had a score of seven or above on the adapted version of the McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD; Zanarini et al., 2003; see next section), (c) identified as a relative or significant other of a person with BPD/ EUPD who was minimum 12 years of age, (d) was aged 18 years or older, (e) an English speaker, and (f) indicated they had or were currently attending FC.
The sample included herein is part of a larger study (Hayes et al., 2023). Of 1207 participants, data of 347 were analyzed; 860 were excluded because the score on the adapted MSI-BPD was <7 (n = 216), they were not primary caregivers (n = 105), were caring for a person under the age of 12 (n = 2), had more than 40% missing data on a measure (n = 21), or they did not state attendance of FC (n = 516).
Table 1 includes the sociodemographic descriptive statistics. ICs were between 20 and 82 years old. The majority resided in North America (n = 319; 91.9%), were female (n = 308, 88.8%), and were parents of the person with BPD (n = 291, 83.9%).
| Caregiver | |
|---|---|
| Age | |
| Mean (SD) | 55.5 (9.9) |
| Gender | |
| Male | 39 (11.2%) |
| Female | 308 (88.8%) |
| Education level | |
| Less than a graduate degree | 81 (23.3%) |
| Graduate degree | 127 (36.6%) |
| Postgraduate degree | 139 (40.1%) |
| Employment | |
| Working | 230 (66.3%) |
| Not working | 117 (33.7%) |
| Relationship to person with BPD | |
| Parent | 291 (83.9%) |
| Child | 11 (3.2%) |
| Sibling | 7 (2.0%) |
| Spouse/partner | 29 (8.4%) |
| Significant other | 9 (2.6%) |
| Living with a person with BPD | |
| Yes | 206 (59.4%) |
| Mean years caregiving | |
| Mean (SD) | 12.9 (9.1) |
| Mean hours contact per week in last month | |
| Mean (SD) | 44.8 (51.8) |
| Caregiver-leader | |
| Yes | 64 (18.4%) |
| Chronic illness | |
| Yes | 93 (26.8%) |
| Person with BPD | |
|---|---|
| Age | |
| Mean (SD) | 28.0 (12.2) |
| Gender | |
| Male | 73 (21.0%) |
| Female | 259 (74.6%) |
| Non-binary | 14 (4.0%) |
| Prefer not to say | 1 (0.3%) |
| Employment | |
| Working | 98 (28.2%) |
| Not working | 249 (71.8%) |
Measures
The survey included the sociodemographic questions, Table 1, and the following measures.
The McLean Screening Instrument for BPD-Carer Version (MSI-BPD-C; Goodman et al., 2011), adapted from the MSI-BPD (Zanarini et al., 2003), is a 10-item screening instrument to validate IC report that the cared-for person has BPD. The adapted version has been used in many studies of caregivers of people with BPD (e.g., Bailey & Grenyer, 2015; Buteau et al., 2008; Goodman et al., 2011; Kirtley et al., 2019). The MSI-BPD is reported to have acceptable test–retest reliability (ICC = 0.72–0.78) and diagnostic accuracy (AUC = 0.73–0.83) with a score of seven or more (Chanen et al., 2008; Zanarini et al., 2003). This minimum score was adopted as inclusion criteria for this study.
The Brief COPE (Carver, 1997) includes 14 coping strategies with reported Cronbach's alphas of 0.5–0.9. In this study, wording was adapted for situational coping, relevant to the research question, as recommended by Carver (1997). The question read “to deal with the stress associated with my role as a caregiver for a person with Borderline Personality Disorder (Emotionally Unstable Personality Disorder), I've…” Mean inter-item correlations were between 0.35 and 0.91 for 13 coping strategies, demonstrating adequate internal consistency (Briggs & Cheek, 1986). The self-distract subscale mean inter-item correlation was 0.17.
The Kessler Psychological Distress scale (K10; Kessler et al., 2002) comprises 10 questions focussed on negative experiences, scored on a 1–5 Likert scale, about how a participant has been feeling over the last 30 days. Total scores range from 10 to 50. Higher scores indicate greater levels of psychological distress. Andrews and Slade (2001) reported convergent validity of the K10 and a significant correlation with the General Health Questionnaire (ρ = 0.5) and Short Form-12 (ρ = −0.6). Kessler et al. (2002) reported an internal consistency of 0.93; Cronbach's alpha for the current sample was 0.92.
The World Health Organization – Five Well-being Index (WHO-5) was included as a measure of positive mental well-being. It asks five questions about the frequency of positive experiences over the previous fortnight on a 0–5 Likert scale and is one of the most widely used measures of psychological well-being (Topp et al., 2015). Summed scores transform to a final range from 0 to 100; higher scores indicate higher positive mental well-being. Hall et al. (2011) reported high content validity. Blom et al. (2012) demonstrated construct validity compared with the Beck's Depression Inventory (r = −0.49). Adequate internal consistency (alpha = 0.84–0.89) has been reported in population studies (Bech et al., 2003; Maguire et al., 2019). Chronbach's alpha in the current study was 0.92.
The Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., 1988) comprises 12 questions, answered on a 1–7 Likert Scale, to measure perceived social support. Cartwright et al. (2022) reported adequate internal consistency (alpha = 0.92) and 1-month test–retest reliability (ICC = 0.90) in another sample of caregivers. In the current sample, Cronbach's alpha for the total score was 0.94.
Data were collected during the COVID-19 pandemic. Therefore, the analysis controlled for COVID-19 anxiety measured using the Coronavirus Anxiety Scale (CAS; Lee, 2020). The CAS asks five questions, answered on a 0–4 Likert scale, about participants' experience of COVID-19-related anxiety symptoms. Higher scores indicate higher anxiety. Previous studies (Lee, 2020; Lee et al., 2020) have demonstrated construct validity and adequate internal consistency (0.92–0.93). The current study Cronbach's alpha was 0.85.
Procedure
Following receipt of institutional ethical approval by the Clinical Psychology Research Ethics Committee, University College Cork, a Qualtrics online survey was posted on a public mental health service webpage and an international online forum and by emailing a register held by two national organizations that help caregivers of people with BPD and a register of past and current attendees of a local program for caregivers of people with BPD. Responses were gathered between August and November 2020.
Statistical analysis
Participants missing >40% of data on one measure (1.7% of respondents) were excluded. Multiple imputation was used to manage >5% missingness in the remaining data (Jakobsen et al., 2017). Psychological distress and positive mental well-being were the dependent variables (DVs). The independent variable (IV) was whether the participant identified as a caregiver-leader. Coping strategies were included as possible mediator variables. Non-normally distributed variables were analyzed using nonparametric tests; parametric test results are reported unless there was a discrepancy with the nonparametric results. Bivariate analyses of the relationship between the sociodemographic variables and the two DVs were conducted. Independent t-tests compared DV differences between those who were and were not caregiver-leaders.
To investigate the relationship between the IV and the two DVs, two hierarchical multiple regression models were undertaken. COVID-19 anxiety and sociodemographic variables that had a significant relationship with the DVs were controls. The 14 coping strategies were introduced in step 2 of each model. Mediation analyses were undertaken using model 4 of Process version 3.5 (Hayes, 2022), using 5000 bootstrap intervals.
RESULTS
Table 2 provides the descriptive statistics of the measures. Table 3 details the relationship between the sociodemographic variables and the DVs.
| M | SD | Median | Minimum | Maximum | |
|---|---|---|---|---|---|
| Psychological distress (K10) | 23.9 | 7.9 | 23 | 10 | 46 |
| Positive mental well-being (WHO-5) | 45.1 | 20.8 | 44 | 0 | 92 |
| Coping strategies | |||||
| Self-distraction | 5.7 | 1.4 | 6 | 2 | 8 |
| Active coping | 6.0 | 1.6 | 6 | 2 | 8 |
| Denial | 2.5 | 1.1 | 2 | 2 | 8 |
| Substance use | 3.1 | 1.7 | 2 | 2 | 8 |
| Use emotional support | 5.4 | 1.7 | 5 | 2 | 8 |
| Use instrumental support | 5.3 | 1.8 | 5 | 2 | 8 |
| Behavioral disengagement | 3.1 | 1.4 | 3 | 2 | 8 |
| Venting | 4.3 | 1.4 | 4 | 2 | 8 |
| Positive reframing | 5.0 | 1.6 | 5 | 2 | 8 |
| Planning | 6.3 | 1.6 | 6 | 2 | 8 |
| Humor | 3.1 | 1.4 | 3 | 2 | 8 |
| Acceptance | 6.6 | 1.4 | 7 | 2 | 8 |
| Religion | 4.7 | 2.1 | 4 | 2 | 8 |
| Self-blame | 4.4 | 1.8 | 4 | 2 | 8 |
| Perceived social support (MSPSS) | 4.8 | 1.4 | 4.9 | 1 | 7 |
| Coronavirus (COVID-19) anxiety | 1.9 | 3.0 | 1 | 0 | 19 |
| MSI-BPD-C | 8.7 | 1.1 | 9 | 7 | 10 |
- Abbreviations: K10, Kessler Psychological Distress scale; MSI-BPD-C, McLean Screening Instrument for BPD-Carer Version; MSPSS, Multidimensional Scale of Perceived Social Support; WHO-5, World Health Organization – Five Well-being Index.
| Test | Psychological distress | Positive mental well-being | |||
|---|---|---|---|---|---|
| Statistic | p | Statistic | p | ||
| Caregiver variables | |||||
| Age a | r | −0.18 | 0.001 | 0.13 | <0.05 |
| Gender b | F(1, 345) | 1.28 | 0.26 | 0.53 | 0.47 |
| Education b | F(2, 344) | 1.01 | 0.37 | 0.78 | 0.46 |
| Employment c | t(345) | 0.75 | 0.46 | 0.02 | 0.98 |
| Caregiver-leader c | t(345) | 4.42 | <0.001 | −4.80 | <0.001 |
| Chronic Illness c | t(345) | −4.03 | <0.001 | 3.36 | 0.001 |
| Coronavirus (COVID-19) anxiety, CAS | r | 0.30 | <0.001 | −0.23 | <0.001 |
| Perceived social support, MSPSS | r | −0.37 | <0.001 | 0.30 | <0.001 |
| Person with BPD variables | |||||
| Age a | r | 0.08 | 0.13 | 0.02 | 0.71 |
| Gender b | F(3, 343) | 0.70 | 0.55 | 1.60 | 0.19 |
| Employment c | t(345) | −0.57 | 0.57 | 1.91 | 0.06 |
| Relationship variables | |||||
| Relation to person with BPD b | F(4, 342) | 3.06 | 0.02 | 1.46 | 0.21 |
| Living with person with BPD b | t(345) | −1.51 | 0.13 | 3.56 | <0.001 |
| Hours of contact d | r | 0.15 | <0.01 | −0.19 | <0.001 |
| Years of caregiving a | r | −0.04 | 0.49 | 0.03 | 0.63 |
- Abbreviations: CAS, Coronavirus Anxiety Scale; MSPSS, Multidimensional Scale of Perceived Social Support.
- a Pearson correlation.
- b One-way ANOVA.
- c T-test.
- d Spearman correlation.
Relationship between being the IV and the DVs
Psychological distress was significantly lower in caregiver-leaders (M = 20.5, SD = 6.6) compared with non-leader-FC participants (M = 24.7, SD = 7.9), t(108.45) = 4.42, p < 0.001. Positive mental well-being was significantly higher in caregiver-leaders (M = 56.0, SD = 19.6) compared with non-leader-FC participants (M = 42.6, SD = 20.3), t(345) = −4.80, p < 0.001.
To investigate the relationship between caregiver-leader and psychological distress, the following controls were included in the regression based on a significant relationship with the DV (Table 3): IC age, chronic illness, perceived social support, and COVID-19 anxiety, IC relationship with the person with BPD, and hours of contact with the person with BPD. To investigate the relationship between caregiver-leader and positive mental well-being, the following variables were included in the regression (Table 3): IC age, chronic illness, perceived social support and COVID-19 anxiety, cohabitation, and hours of contact with the person with BPD. Tables 4 and 5 summarize the regression results for each DV. Being a caregiver-leader and the sociodemographic variables contributed significantly to both regression models. Being a caregiver-leader was a significant predictor, uniquely explaining 2.68% of the variance in psychological distress and 4.39% of the variance in positive mental well-being, when controlling for the other variables. The addition of coping strategies in step 2 explained an additional 20.5% of the variance in psychological distress and 13.5% in positive mental well-being. This change was significant in both models, ΔR 2 = 0.21, ΔF(14, 325) = 9.36, p < 0.001 (psychological distress) and ΔR 2 = 0.14, ΔF(14, 325) = 4.96, p < 0.001 (positive mental well-being). The inclusion of coping strategies in step 2 caused a reduction in the variance in psychological distress explained by caregiver-leader from 2.68% in step 1 to 0.20% in step 2 and a reduction in the variance in positive mental well-being explained by caregiver-leader from 4.39% in step 1 to 0.83% in step 2.
| B | SE B | β | p | |
|---|---|---|---|---|
| Step 1 | ||||
| (Constant) | 39.46 | 2.94 | <0.001 | |
| Age | −0.15 | 0.04 | −0.19 | <0.001 |
| Caregiver-leader | −3.38 | 0.95 | −0.17 | <0.001 |
| Chronic illness | 2.82 | 0.84 | 0.16 | <0.01 |
| Hours per week of contact | 0.00 | 0.01 | 0.01 | 0.87 |
| Relationship (spouse/partner) | 0.29 | 1.43 | 0.01 | 0.84 |
| Perceived social support, MSPSS | −1.79 | 0.28 | −0.31 | <0.001 |
| Coronavirus (COVID-19) anxiety, CAS | 0.56 | 0.13 | 0.21 | <0.001 |
| Step 2 | ||||
| (Constant) | 24.95 | 3.66 | <0.001 | |
| Age | −0.13 | 0.04 | −0.16 | <0.001 |
| Caregiver-leader | −1.01 | 0.89 | −0.05 | 0.25 |
| Chronic illness | 1.91 | 0.75 | 0.11 | 0.01 |
| Hours per week of contact | 0.00 | 0.01 | 0.02 | 0.56 |
| Relationship (spouse/partner) | 0.09 | 1.26 | 0.00 | 0.94 |
| Perceived social support, MSPSS | −1.35 | 0.30 | −0.23 | <0.001 |
| Coronavirus (COVID-19) anxiety, CAS | 0.42 | 0.11 | 0.16 | <0.001 |
| Self-distraction | 0.26 | 0.25 | 0.05 | 0.30 |
| Active coping | 0.30 | 0.28 | 0.06 | 0.28 |
| Denial | 0.83 | 0.34 | 0.11 | 0.01 |
| Substance use | 0.39 | 0.20 | 0.08 | 0.05 |
| Use emotional support | −0.09 | 0.29 | −0.02 | 0.76 |
| Use instrumental support | 0.34 | 0.25 | 0.08 | 0.18 |
| Behavioral disengagement | 0.68 | 0.28 | 0.12 | 0.02 |
| Venting | −0.01 | 0.26 | 0.00 | 0.98 |
| Positive reframing | −0.78 | 0.22 | −0.16 | <0.001 |
| Planning | 0.31 | 0.27 | 0.06 | 0.26 |
| Humor | 0.01 | 0.24 | 0.00 | 0.98 |
| Acceptance | −0.16 | 0.27 | −0.03 | 0.56 |
| Religion | −0.20 | 0.16 | −0.05 | 0.21 |
| Self-blame | 1.20 | 0.20 | 0.28 | <0.001 |
- Note: Step 1 R 2 adj = 0.27, F(7, 339) = 19.48, p < 0.001; Step 2 R 2 adj = 0.46, F(21, 325) = 14.98, p < 0.001.
| B | SE B | β | p | |
|---|---|---|---|---|
| Step 1 | ||||
| (Constant) | 19.65 | 7.62 | 0.01 | |
| Age | 0.23 | 0.10 | 0.11 | 0.03 |
| Caregiver-leader | 11.44 | 2.60 | 0.21 | <0.001 |
| Chronic illness | −5.26 | 2.30 | −0.11 | 0.02 |
| Hours per week of contact | 0.01 | 0.02 | 0.03 | 0.61 |
| Cohabitation | −7.64 | 2.35 | −0.18 | <0.01 |
| Perceived social support, MSPSS | 3.77 | 0.76 | 0.24 | <0.001 |
| Coronavirus (COVID-19) anxiety, CAS | −1.24 | 0.35 | −0.18 | <0.001 |
| Step 2 | ||||
| (Constant) | 35.52 | 10.50 | <0.001 | |
| Age | 0.23 | 0.10 | 0.11 | 0.03 |
| Caregiver-leader | 5.41 | 2.61 | 0.10 | 0.04 |
| Chronic illness | −4.24 | 2.22 | −0.09 | 0.06 |
| Hours per week of contact | 0.00 | 0.02 | 0.01 | 0.92 |
| Cohabitation | −6.48 | 2.22 | −0.15 | <0.01 |
| Perceived social support, MSPSS | 2.56 | 0.87 | 0.17 | <0.01 |
| Coronavirus (COVID-19) anxiety, CAS | −1.09 | 0.34 | −0.16 | <0.01 |
| Self-distraction | −1.02 | 0.73 | −0.07 | 0.16 |
| Active coping | 0.79 | 0.82 | 0.06 | 0.34 |
| Denial | −0.31 | 1.00 | −0.02 | 0.75 |
| Substance use | −1.21 | 0.58 | −0.10 | 0.04 |
| Use emotional support | 0.70 | 0.86 | 0.06 | 0.42 |
| Use instrumental support | −0.30 | 0.73 | −0.03 | 0.69 |
| Behavioral disengagement | −0.95 | 0.83 | −0.06 | 0.25 |
| Venting | 0.20 | 0.77 | 0.01 | 0.79 |
| Positive reframing | 2.28 | 0.64 | 0.18 | <0.001 |
| Planning | −2.69 | 0.81 | −0.20 | <0.01 |
| Humor | 1.10 | 0.70 | 0.07 | 0.12 |
| Acceptance | 0.46 | 0.80 | 0.03 | 0.56 |
| Religion | 0.60 | 0.48 | 0.06 | 0.20 |
| Self-blame | −1.71 | 0.58 | −0.15 | <0.01 |
- Note: Step 1 R 2 adj = 0.22, F(7, 339) = 14.79, p < 0.001; Step 2 R 2 adj = 0.33, F(21, 325) = 9.04, p < 0.001.
Do coping strategies partially mediate the IV-DV relationship?
Mediation analyses were conducted as the variance explained by the IV (caregiver-leader) reduced in both regression models when coping strategies were added, suggesting that the effect of caregiver-leader may be partially mediated by coping strategies. Table 6 identifies coping strategies selected for investigation as mediator variables based on a significant relationship between the coping strategy and the IV and the DVs. Mediation analysis, summarized in Tables 7 and 8, showed that the use of emotional support, behavioral disengagement, positive reframing, acceptance, and self-blame coping partially mediated the relationship between caregiver-leader and both DVs.
| Coping strategy | Caregiver-leader v coping strategy | Coping strategy v psychological distress | Coping strategy v positive mental well-being | |||
|---|---|---|---|---|---|---|
| r | p | r | p | r | p | |
| Self-distraction | 0.09 | 0.08 | 0.15 | <0.01 | −0.11 | <0.05 |
| Active coping | 0.07 | 0.19 | 0.05 | 0.41 | 0.07 | 0.19 |
| Denial | −0.07 | 0.18 | 0.33 | <0.001 | −0.21 | <0.001 |
| Substance use | −0.03 | 0.60 | 0.19 | <0.001 | −0.16 | <0.01 |
| Use emotional supporta,b | 0.16 | <0.01 | −0.18 | <0.01 | 0.21 | <0.001 |
| Use instrumental support | 0.00 | 0.97 | −0.21 | <0.001 | 0.02 | 0.76 |
| Behavioral disengagementa,b | −0.13 | <0.05 | 0.35 | <0.001 | −0.27 | <0.001 |
| Venting | 0.02 | 0.73 | 0.08 | 0.12 | −0.00 | 0.93 |
| Positive reframinga,b | 0.23 | <0.001 | −0.18 | <0.01 | 0.22 | <0.001 |
| Planning | −0.05 | 0.33 | 0.17 | <0.01 | −0.17 | <0.01 |
| Humor | 0.05 | 0.37 | 0.02 | 0.74 | 0.07 | 0.18 |
| Acceptancea,b | 0.22 | <0.001 | −0.21 | <0.001 | 0.19 | <0.001 |
| Religionb | 0.21 | <0.001 | −0.08 | 0.14 | 0.12 | <0.05 |
| Self-blamea,b | −0.16 | <0.01 | 0.47 | <0.001 | −0.33 | <0.001 |
- Note: Selected to investigate if mediator variable for the relationship between caregiver-leader and psychological distress (a) or positive mental well-being (b).
| Mediator (M) | Variable/effects | b | SE | t | p | 95% CI | |
|---|---|---|---|---|---|---|---|
| LL | UL | ||||||
| Use of emotional support | IV->K10 | −3.70 | 1.07 | −3.46 | <0.001 | −5.81 | −1.60 |
| IV->M | 0.70 | 0.23 | 3.08 | <0.01 | 0.25 | 1.15 | |
| M->K10 | −0.70 | 0.25 | −2.83 | <0.01 | −1.20 | −0.21 | |
| Effects Direct | −3.70 | 1.07 | −3.46 | <0.001 | −5.81 | −1.60 | |
| Indirect | −0.50 | 0.27 | −1.13 | −0.07 | |||
| Behavioral disengagement | IV->K10 | −3.38 | 1.02 | −3.32 | <0.01 | −5.38 | −1.37 |
| IV->M | −0.44 | 0.19 | −2.3 | <0.05 | −0.82 | −0.07 | |
| M->K10 | 1.86 | 0.29 | 6.48 | <0.001 | 1.29 | 2.42 | |
| Effects Direct | −3.38 | 1.02 | −3.32 | <0.01 | −5.38 | −1.37 | |
| Indirect | −0.82 | 0.36 | −1.59 | −0.17 | |||
| Positive reframing | IV->K10 | −3.58 | 1.09 | −3.29 | <0.01 | −5.72 | −1.44 |
| IV->M | 0.97 | 0.22 | 4.37 | <0.001 | 0.53 | 1.40 | |
| M->K10 | −0.65 | 0.26 | −2.50 | <0.05 | −1.15 | −0.14 | |
| Effects Direct | −3.58 | 1.09 | −3.29 | <0.01 | −5.72 | −1.44 | |
| Indirect | −0.32 | 0.31 | −1.32 | −0.09 | |||
| Acceptance | IV->K10 | −3.41 | 1.08 | −3.16 | <0.01 | −5.54 | −1.29 |
| IV->M | 0.80 | 0.19 | 4.24 | <0.001 | 0.43 | 1.17 | |
| M->K10 | −0.99 | 0.30 | −3.27 | <0.01 | −1.58 | −0.39 | |
| Effects Direct | −3.41 | 1.08 | −3.16 | <0.01 | −5.54 | −1.29 | |
| Indirect | −0.79 | 0.29 | −1.38 | −0.26 | |||
| Self-blame | IV->K10 | −2.79 | 0.97 | −2.89 | <0.01 | −4.70 | −0.89 |
| IV->M | −0.74 | 0.25 | −2.94 | <0.01 | −1.24 | −0.25 | |
| M->K10 | 1.89 | 0.20 | 9.30 | <0.001 | 1.49 | 2.30 | |
| Effects Direct | −2.79 | 0.97 | −2.89 | <0.01 | −4.70 | −0.89 | |
| Indirect | −1.41 | 0.48 | −2.39 | −0.49 | |||
| Mediator (M) | Variable/effects | b | SE | t | p | 95% CI | |
|---|---|---|---|---|---|---|---|
| LL | UL | ||||||
| Use of emotional support | IV->WHO-5 | 2.98 | 0.70 | 4.26 | <0.001 | 1.60 | 4.35 |
| IV->M | 0.70 | 0.23 | 3.08 | <0.01 | 0.25 | 1.15 | |
| M->WHO-5 | 0.54 | 0.16 | 3.30 | <0.01 | 0.22 | 0.86 | |
| Effects Direct | 2.98 | 0.69 | 4.26 | <0.001 | 1.60 | 4.35 | |
| Indirect | 0.38 | 0.18 | 0.08 | 0.78 | |||
| Behavioral disengagement | IV->WHO-5 | 2.95 | 0.68 | 4.31 | <0.001 | 1.60 | 4.29 |
| IV->M | −0.44 | 0.19 | −2.34 | <0.05 | −0.82 | −0.07 | |
| M->WHO-5 | −0.92 | 0.19 | −4.77 | <0.001 | −1.30 | −0.54 | |
| Effects Direct | 2.95 | 0.68 | 4.31 | <0.001 | 1.60 | 4.29 | |
| Indirect | 0.41 | 0.20 | 0.07 | 0.86 | |||
| Positive reframing | IV->WHO-5 | 2.82 | 0.71 | 3.98 | <0.001 | 1.42 | 4.21 |
| IV->M | 0.97 | 0.22 | 4.37 | <0.001 | 0.53 | 1.40 | |
| M->WHO-5 | 0.56 | 0.17 | 3.32 | <0.01 | 0.23 | 0.89 | |
| Effects Direct | 2.82 | 0.71 | 3.98 | <0.001 | 1.42 | 4.21 | |
| Indirect | 0.54 | 0.22 | 0.17 | 1.01 | |||
| Acceptance | IV->WHO-5 | 2.94 | 0.71 | 4.14 | <0.001 | 1.54 | 4.34 |
| IV->M | 0.80 | 0.19 | 4.24 | <0.001 | 0.43 | 1.17 | |
| M->WHO-5 | 0.52 | 0.20 | 2.61 | <0.01 | 0.13 | 0.91 | |
| Effects Direct | 2.94 | 0.71 | 4.14 | <0.001 | 1.54 | 4.34 | |
| Indirect | 0.41 | 0.19 | 0.06 | 0.81 | |||
| Self-blame | IV->WHO-5 | 2.74 | 0.68 | 4.05 | <0.001 | 1.41 | 4.07 |
| IV->M | −0.74 | 0.25 | −2.94 | <0.01 | −1.24 | −0.25 | |
| M->WHO-5 | −0.83 | 0.14 | 5.84 | <0.001 | −1.11 | −0.55 | |
| Effects Direct | 2.74 | 0.68 | 4.05 | <0.001 | 1.41 | 4.07 | |
| Indirect | 0.62 | 0.23 | 0.20 | 1.10 | |||
| Religion | IV->WHO-5 | 3.15 | 0.71 | 4.40 | <0.001 | 1.74 | 4.55 |
| IV->M | 1.18 | 0.29 | 4.07 | <0.001 | 0.61 | 1.75 | |
| M->WHO-5 | 0.17 | 0.13 | 1.35 | 0.18 | −0.08 | 0.43 | |
| Effects Direct | 3.15 | 0.71 | 4.40 | <0.001 | 1.74 | 4.55 | |
| Indirect | 0.21 | 0.17 | −0.10 | 0.57 | |||
DISCUSSION
The current study aimed to investigate if (1) being an IC who has received FC and identified as having led interventions (caregiver-leader) is associated with different outcomes compared with ICs who have received FC but not identified as leaders (non-leader-FC participants) and (2) coping strategies partially mediated the relationship between being a caregiver-leader and IC outcomes. In response to the first aim, this study found that caregiver-leaders had greater positive mental well-being and lower psychological distress compared with non-leader-FC participants. This supports a suggestion by Courey et al. (2021) of anecdotally improved outcomes in ICs who went on to facilitate FC. In response to the second aim, the results suggest that differences in coping strategy use may explain better outcomes for caregiver-leaders compared with non-leader-FC participants. Positive reframing, use of emotional support, acceptance coping, self-blame, and behavioral disengagement partially explained the positive relationship between being a caregiver-leader and positive mental well-being and the negative relationship between being a caregiver-leader and psychological distress.
Hayes et al. (2023) previously found that positive reframing and self-blame were the strongest positive and negative coping predictors of caregiver outcomes respectively. Therefore, the results herein suggest that being a caregiver-leader is associated with greater use of beneficial and reduced use of maladaptive coping strategies.
While further research is required to understand the link between coping strategies measured and FC teachings, the findings of the current study make sense considering the FC program content. The current results suggest that the coping strategies linked with positive cognitive restructuring, acceptance coping, and positive reframing (Folkman & Moskowitz, 2004), partially mediated the relationship between being a caregiver-leader and IC outcomes. The finding that being a caregiver-leader, compared with non-leader-FC participants, was associated with higher use of acceptance is consistent with the FC manual addressing the concept of radical acceptance in detail. While the terminology of positive reframing is not found within the FC manual, it is likely that in-depth knowledge of the FC assumption of adopting the most benign interpretation and the teaching on dialectical thinking equipped caregiver-leaders with the capacity to engage in positive reframing. Indeed, Herley's (2015) study reported that post-intervention FC participants noted that the program helped them to reappraise the person with BPD and to increase acceptance of their situation and their loved one, indicating that FC may aid cognitive reappraisal.
Family Connections material on the functions of emotions, wise mind, and effectiveness, and on the importance of reducing emotional vulnerability and emotional reactivity may have contributed to the finding of caregiver-leaders being higher on the use of emotional support compared with non-leader-FC participants. Being a caregiver-leader was associated with lower self-blame. This likely reflects caregiver-leaders being skilled in practicing a non-judgemental stance, which is an integral part of FC. The lower levels of behavior disengagement in caregiver-leaders are consistent with understanding how family functioning and individual functioning are inextricably linked, and a deeper understanding of the concept of reciprocity. These concepts are threaded throughout FC.
Family Connection studies have reported different outcomes in relation to mastery with post-intervention increases (Hoffman et al., 2005) and decreases (Hoffman et al., 2007) in mastery at 3-month follow-up. Flynn et al. (2017) reported an increase in mastery post-FC but no further mastery increase at 19-month follow-up, noting that participants identified challenges with consolidating and generalizing skills and that additional resources may be required to support skill strengthening post-FC (Flynn et al., 2017). Courey et al. (2021) suggested that being an FC caregiver-leader may afford ICs the opportunity to develop mastery and practice skills learned during the FC program.
While further research is required, collectively the findings herein provide preliminary support for Courey et al.'s (2021) suggestion. Compared with non-leader-FC participants, caregiver-leaders had many skilled coping strategies more firmly within their coping repertoires. The finding is understandable as caregiver-leaders are more likely to have consolidated coping skills learned in the FC program when they subsequently facilitated FC. As ICs initially engage with a FC program, they typically focus on the process of skills acquisition. It is likely that as individuals then become caregiver-leaders, ongoing teaching and role play of FC skills in FC groups provides opportunities for additional skill rehearsal and supports further skills generalization. Having generalized the FC skills, it corresponds that FC caregiver-leaders would report more frequent skills use. It is highly desirable that ICs would attain skills generalization, allowing them to effectively and readily draw on FC skills in complex situations. Given the outlined challenges of the 12-week FC programs in getting ICs to a place of significantly enhanced skills mastery, inviting ICs who have completed FC programs to become FC leaders may provide a cost-effective pathway to support ICs to achieve skills generalization.
The findings herein should consider limitations relating to the cross-sectional design, limiting statements of causality, and sample generalizability, limited by the majority being North American, female, and parents of the person with BPD. There is a possible positive association between the high percentage of females and parents participating. Typically females, often mothers, undertake most family caregiving. However, there is emerging evidence of increased male caregiving (Sharma et al., 2016), highlighting the need for research on male caregivers and exploration of gender differences. Recruitment in treatment settings may support increased male participation; there were higher rates of male research participation at treatment locations (e.g. Flynn et al., 2017; Seigerman et al., 2020) than in this online study.
Despite these limitations, the current research presents preliminary evidence of better caregiver outcomes for caregiver-leaders than non-leader-FC participants among caregivers who have or are attending FC and that these outcomes may be partially mediated by coping strategy use. Accordingly, this research provides preliminary support for adopting a caregiver-led model for FC delivery, which may also be more cost-effective than clinician-led programs (Acri et al., 2017). In a climate of limited mental health resources, adopting caregiver-led delivery may also enable mental health services to increase access to FC and reduce wait times. To inform these decisions, future research is required to assess the efficacy of caregiver-led compared with clinician-led FC programs on participant outcomes as previously recommended (Boritz et al., 2021; Courey et al., 2021; Guillén et al., 2021); to investigate if differences between caregiver-leaders and non-leader-FC participants remain if both groups attend the same number of FC sessions; and if there are differences in outcomes between current and past attendees of FC.
ACKNOWLEDGMENTS
The authors would like to respectfully acknowledge the support of the National Education Alliance for Borderline Personality Disorder, the Sashbear Foundation, and BPDFamily.com who assisted with recruitment of participants.
