Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Jul 22;56(4):e70126. doi: 10.1111/sltb.70126

Examining the Longitudinal Effects of Therapeutic Alliance Based on Session Rating Scale Scores on Suicidal Desire in Homebound Older Adults

Cheolwoo Park 1, Laura Shannonhouse 2,✉, Alexander Kirpich 3, Matthew C Fullen 4, Erika L LeBlanc 5, Hailey Clark 2, Jared Rivera 2, Fang Xue 6
PMCID: PMC13389554  PMID: 42482621

ABSTRACT

Introduction

The Session Rating Scale (SRS) is widely used to assess therapeutic alliance in clinical settings, yet its application in paraprofessional‐delivered interventions for older adults remains underexplored. This study examined whether therapeutic alliance predicted improvements in suicide desire among community‐dwelling older adults participating in a warm‐calling intervention.

Methods

Binomial generalized linear mixed‐effects models were used with dichotomized high‐score responses for thwarted belongingness (TB) and perceived burdensomeness (PB) to evaluate the role of therapeutic alliance (SRS) over time. Participants were assigned to one of two 8‐week conditions: (1) BE, emphasizing empathy and connection through befriending and reminiscence; or (2) BE + ASIST, which incorporated an aging‐adapted LivingWorks Applied Suicide Intervention Skills Training (ASIST). Trained paraprofessionals delivered all sessions.

Results

For PB, the odds of being classified as elevated decreased over time, and this decrease was statistically significant. For TB, a three‐way interaction among condition, time, and SRS further showed that higher SRS scores predicted greater reductions in the odds of elevated TB within the BE + ASIST condition.

Conclusion

Findings underscore the importance of therapeutic alliance in enhancing paraprofessional‐led interventions for late‐life suicide prevention.

Keywords: moderation, older adults, suicide desire, therapeutic alliance

Older adults are at greater risk for suicide and frequently encounter barriers to timely mental health care, particularly when chronic health conditions, mobility limitations, and social isolation limit access to services (Conwell et al. 2011; Drapeau and McIntosh 2025; Guidry and Cukrowicz 2016; Marty et al. 2010). Beyond formal mental health services, supportive community‐based interventions may represent an important pathway for addressing psychosocial distress in this population.

One relational factor that may be especially salient within these interventions is the therapeutic alliance, which describes the feelings and attitudes between the treatment provider and the client, and the way those are expressed (Gelso and Carter 1985). A stronger therapeutic alliance has been consistently linked to greater engagement and better treatment outcomes (Duncan et al. 2003; Martin et al. 2000). Emerging evidence further reveals that a strong therapeutic alliance has been associated with decreased suicidal thoughts (Fartacek et al. 2023; Huggett et al. 2021). As such, understanding the role of therapeutic alliance may be particularly significant when supportive services are delivered by paraprofessionals rather than licensed mental health professionals.

In this study, paraprofessionals refer to Aging Services Network (ASN) providers, who are not licensed but rather staff or volunteers, routinely interact with older adults through home‐ and community‐based services, and can be trained to deliver structured supportive interventions and foster genuine connections (Shannonhouse et al. 2022; Westcott et al. 2022). Because paraprofessionals frequently work with individuals who may face barriers to accessing traditional psychotherapy, the quality of the helping relationship that they establish may be especially important. Although these services do not constitute formal psychotherapy, paraprofessionals may nonetheless foster empathy, consistency, and meaningful interpersonal connection through sustained supportive engagement (Shannonhouse et al. 2022; Westcott et al. 2022; Gould et al. 2013; Rodgers 2010). Therefore, further research is needed to better understand the therapeutic alliance in paraprofessional and other nonclinical intervention contexts, particularly for clients with accessibility needs, such as those who have difficulty traveling to a clinic or treatment setting.

1. Literature Review

1.1. Working Alliance and Suicide Desire

Thwarted belongingness and perceived burdensomeness are central interpersonal constructs within Joiner's Interpersonal Theory of Suicide and are theorized to underlie the desire for suicide (Joiner 2005; Van Orden et al. 2010). These experiences may be particularly salient among older adults, for whom mobility limitations, chronic health issues, and diminishing social networks can heighten feelings of interpersonal disconnection and burden (Fullen et al. 2020; Crosby et al. 2020; Dienst et al. 2023; Hill and Pettit 2014; Rasmussen and Wingate 2011).

The therapeutic alliance may be especially relevant to these interpersonal experiences because it reflects the extent to which individuals feel understood, supported, and collaboratively engaged in care. Indeed, a systematic review found that therapeutic alliance was associated with suicide‐related outcomes in 11 of the 12 studies examined (Dunster‐Page et al. 2017). In addition, individuals with lived experience of suicidal ideation or attempts have described helpful responders as genuine, accepting, empathic, respectful, and compassionate (Staniforth et al. 2023), suggesting that relational qualities may play a meaningful role in suicide‐related care. Goger et al. (2025) also indicated that a decrease in perceived burdensomeness was associated with reduced suicidal ideation, and this relationship was buffered by the strength of the therapeutic alliance. Therefore, there is evidence that alliance plays a vital role in addressing interpersonal vulnerabilities by providing a relational space where clients can begin to feel valued, understood, and connected (Schechter et al. 2019).

The Session Rating Scale (SRS; Duncan et al. 2003) was developed to evaluate clients' perceptions of their working alliance with treatment providers and the overall effectiveness of each session. This measure is widely utilized across helping fields to evaluate and track the relationship along a continuum (Gould et al. 2001; Klimek et al. 2017). Such standardized measures can help in identifying subtle declines in client engagement or other negative shifts in daily functioning, enabling adjustments on the part of the clinician or paraprofessional to enhance treatment outcomes (Hafkenscheid et al. 2010). Furthermore, as structured treatment approaches provided by paraprofessional helpers gain recognition for their therapeutic benefits among nonclinical populations (Montgomery et al. 2010; Shannonhouse et al. 2022; Westcott et al. 2022; Wyman et al. 2008), the SRS, or related measures ought to also be used to assess the therapeutic alliance in these contexts.

The SRS is a brief, four‐item measure of the therapeutic alliance that captures core alliance domains from Bordin's (1979) original framework, including therapeutic bond, goal agreement, task agreement, and it also includes an item reflecting clients' confidence or overall satisfaction with the session, which has been linked to treatment outcomes (Gaston 1991). In practice, the SRS provides a structured and efficient way for clients to give session‐related feedback, allowing providers to identify concerns early, adjust their approach, and monitor alliance quality and client progress over time, thereby supporting collaborative, responsive, and outcome‐focused care (Duncan et al. 2003).

1.2. The SRS and Paraprofessionals

Although “natural helpers” (Wyman et al. 2008) may lack formal clinical training, their established relationships with community members position them well to foster a strong therapeutic relationship and identify changes in behavior, warning signs, and needs, including in ways formal therapeutic settings cannot. The SRS has the potential for benefits with nonclinical samples, yet there remains a scarcity of research specifically examining the application of the SRS within nonclinical populations, particularly in the context of older adults classified as homebound who may have difficulty accessing traditional therapy services (Fullen et al. 2020; Marti et al. 2020; Sterling‐Fox 2018).

The relationship between providers and older adults is crucial in determining intervention success (Boschann et al. 2024; Mace et al. 2018; Thompson et al. 1986). For example, the quality of therapeutic relationships was a significant contributor to improved psychotherapy outcomes for older adults with geriatric depression and executive dysfunction (Mace et al. 2018). The impact of older adult clients' perceptions of their relationships with their helpers on therapeutic outcomes underscores the alliance as a critical factor in older adults' treatment experience (Boschann et al. 2024).

Furthermore, because paraprofessionals often provide ongoing support to older adults who may face challenges in accessing traditional psychotherapy, the quality of the relationship facilitated by natural helpers may play an important role in their treatment engagement, well‐being, and intervention outcomes. In this context, the SRS could serve as an essential metric for assessing engagement and relational quality in paraprofessional interventions. Investigating the SRS in this population may further broaden the therapeutic alliance literature beyond formal psychotherapy contexts and offer important insight into the relational processes that contribute to effective paraprofessional support for older adults.

1.3. The Standardized and Manualized BE WITH Program

BE WITH is one such paraprofessional‐led treatment program for older adults receiving HCBS. BE WITH consists of an 8‐week warm‐calling intervention designed for older adults across urban, racially diverse, suburban, and/or rural settings. It is delivered in 20‐ to 30‐min “dosages” under two specific conditions: (1) BE condition and (2) BE + ASIST condition (Shannonhouse et al. 2022; Westcott et al. 2022). First, the BE condition involved services provided by trained aging network providers (i.e., paraprofessionals) who received a 2.5‐h training on how to promote belongingness and empathy. This training was grounded in principles from the befriending literature (Fakoya et al. 2021; Wiles et al. 2019) and narrative reminiscence techniques (Yousefi et al. 2015), both beneficial practices when working with older adults that may be left out of services provided to them through the Aging Services Network (ASN).

More specifically, the befriending literature includes five components: (1) reciprocity (i.e., a mutual sense of benefit between both parties); (2) intimacy (i.e., a willingness to share deeply, beginning with surface‐level interactions to build rapport, followed by deeper connections that drive meaningful outcomes); (3) reliability and respect (i.e., maintaining consistency by calling at scheduled times signaling that the older adults matter); (4) proximity (i.e., a sense of connection within the community); and (5) autonomy (i.e., feeling both parties engage in the interaction willingly and meaningfully). Narrative reminiscence techniques included inviting older persons to recall and share their memories about a variety of predetermined themes (e.g., critical decisive life events, beliefs on the meaning and goals of life; Yousefi et al. 2015). Finally, a brief module on suicide intervention is also included in the BE condition to equip trained providers to identify risk and support the older adult by getting additional help. Overall, these trained providers offered a “small dose of sincere connection” throughout the treatment. While this may be sufficient for most older adults, others may need more than social connection and support as approximately 20% of older adults receiving HCBS have been found to have at least some risk of suicide (Fullen et al. 2020; Shannonhouse et al. 2021), highlighting the need for targeted interventions.

Second, the BE + ASIST condition builds upon the foundational BE training by incorporating an aging variant of LivingWorks ASIST (Applied Suicide Intervention Skills Training; Lang et al. 2013), which spans 14 h and equips ASN providers with essential suicide intervention skills, tailored especially for older adults (Shannonhouse et al. 2022; Westcott et al. 2022). It incorporates the Interpersonal Theory of Suicide (Joiner 2005), which posits that thwarted belongingness and perceived burdensomeness contribute to suicide desire, whereas acquired capability is conceptually distinct and refers to an individual's capacity to enact lethal self‐injury (Van Orden et al. 2010).

Similar to CPR, which enables lay people to assist someone drowning, LivingWorks ASIST equips lay providers to provide “suicide first aid,” ensuring non‐clinicians can provide critical support to individuals at risk (Lang et al. 2013). Often, those experiencing suicidality, and even their helpers, perceive only two options: (a) to die by suicide or (b) to live. The Pathway for Assisting Life Model (PAL) taught during ASIST introduces a third option: to stay safe for now (Lang et al. 2013). By engaging in the PAL model, the needs of persons with thoughts of suicide are paired with the corresponding caregiver tasks needed to address them. Trained paraprofessionals help participants explore their ambivalence about dying, acknowledge their confusion, and identify personal reasons for living (Lang et al. 2013). Importantly, ASIST does not seek to take away a person's autonomy. Instead, it restores autonomy by helping at‐risk individuals identify their own motivations for living and integrate them into a personalized safety plan (Lang et al. 2013). People with thoughts of suicide can always choose to end their life. However, this pause provides an opportunity for older adults to get in touch with their own ambivalence, and reasons for living.

In sum, the aforementioned empirical evidence regarding the link between therapeutic relationships and outcomes indicates that the SRS could be significantly associated with the treatment outcomes of older adult participants of programs such as BE WITH. In other words, given that the SRS provides a structured and consistent means of evaluating the therapeutic alliance, its application may offer valuable insights into the dynamics of paraprofessional‐older adult relationships and their influence on treatment outcomes.

Thus, the present study aimed to further this line of inquiry by examining the applicability of the SRS in a nonclinical sample of older adults receiving an 8‐week standardized and manualized connections treatment provided by trained ASN providers.

2. Methods

The trial and data collection were approved by the Georgia State University Institutional Review Board (IRB H21359), and older adults provided informed consent for voluntary participation prior to the start of treatment.

2.1. Treatment

In this study, we trained 60 paraprofessionals to provide an 8‐week standardized, manualized warm‐calling treatment aimed at enhancing social connection and well‐being (BE WITH). These individuals aligned with the criteria for natural helpers noted by Wyman et al. (2008), in that they were recommended by ASN personnel due to having positive helper characteristics, such as being “warm” and “empathic,” with “older adults naturally gravitating towards them.” They also had regular and ongoing interactions with older adults who were potentially at risk of suicide, positioning them as ideal to provide train the BE WITH treatment.

2.2. Participants

Four hundred and sixty‐three older adults initiated treatment across both treatment groups. The analysis of the control group (N = 205) was not performed, since SRS scores were not collected from the control group due to them not receiving treatment. The entire sample demographics collected include age, gender, religious views, firearm ownership, race, highest educational level, average monthly income, and veteran status. Participants averaged approximately 74 years in age, identified as predominantly Black (56.4%), and reported low income (44.5%). The majority identified as Christian (85.5%), and had a GED or less than 12 years of school (55.0%). A full breakdown of the demographics for two analyzed treatment conditions only is provided in Table 1. While participants were randomized between the intervention groups, potential differences between the BE and BE WITH conditions were also evaluated. A t‐test was used for continuous variables (i.e., age), and a chi‐squared test was applied for all other categorical demographic characteristics. No significant differences were found between the groups in terms of demographic characteristics.

TABLE 1.

Demographic characteristics.

Characteristic Condition
BE (N = 236) BE WITH (N = 227) Total (N = 463)
Age
Mean [Min, Max] 73.7 [54.0, 96.0] 74.0 [53.0, 100] 73.9 [53.0, 100]
Gender
Female 150 (63.6%) 149 (65.6%) 299 (64.6%)
Male 86 (36.4%) 77 (33.9%) 163 (35.2%)
Transgender, male to female 0 (0%) 1 (0.4%) 1 (0.2%)
Religious views
Christian 196 (83.1%) 200 (88.1%) 396 (85.5%)
Jewish 2 (0.8%) 2 (0.9%) 4 (0.9%)
Muslim 3 (1.3%) 2 (0.9%) 5 (1.1%)
Something else 28 (11.9%) 18 (7.9%) 46 (9.9%)
Prefer not to say 5 (2.1%) 3 (1.3%) 8 (1.7%)
No response 2 (0.8%) 2 (0.9%) 4 (0.9%)
Firearm ownership
No 184 (78.0%) 166 (73.1%) 350 (75.6%)
Yes 52 (22.0%) 60 (26.4%) 112 (24.2%)
Race
No response 2 (0.8%) 1 (0.4%) 3 (0.6%)
Black 132 (55.9%) 129 (56.8%) 261 (56.4%)
White 88 (37.3%) 83 (36.6%) 171 (36.9%)
American Indian/Alaskan Native 1 (0.4%) 2 (0.9%) 3 (0.6%)
Native Hawaiian or Pacific Islander 0 (0%) 0 (0%) 0 (0%)
More than one race 7 (3.0%) 5 (2.2%) 12 (2.6%)
Other 6 (2.5%) 7 (3.1%) 13 (2.8%)
Highest education level
No response 0 (0%) 1 (0.4%) 1 (0.2%)
Less than 12 years 51 (21.6%) 51 (22.5%) 102 (22.0%)
High school diploma or GED 72 (30.5%) 81 (35.7%) 153 (33.0%)
Trade school 11 (4.7%) 16 (7.0%) 27 (5.8%)
Some college 50 (21.2%) 43 (18.9%) 93 (20.1%)
Associate's degree 13 (5.5%) 10 (4.4%) 23 (5.0%)
Bachelor's degree 25 (10.6%) 14 (6.2%) 39 (8.4%)
Master's degree 9 (3.8%) 8 (3.5%) 17 (3.7%)
Doctorate 5 (2.1%) 3 (1.3%) 8 (1.7%)
Average monthly income
No response 6 (2.5%) 6 (2.6%) 12 (2.6%)
Under $500 7 (3.0%) 6 (2.6%) 13 (2.8%)
Between $500 and $1000 92 (39.0%) 101 (44.5%) 193 (41.7%)
Between $1000 and $1500 89 (37.7%) 70 (30.8%) 159 (34.3%)
Between $1500 and $2000 24 (10.2%) 27 (11.9%) 51 (11.0%)
Between $2000 and $2500 9 (3.8%) 11 (4.8%) 20 (4.3%)
Between $2500 and $3000 3 (1.3%) 4 (1.8%) 7 (1.5%)
Over $3000 6 (2.5%) 2 (0.9%) 8 (1.7%)
Veteran status
No response 0 (0%) 1 (0.4%) 1 (0.2%)
Not a veteran 212 (89.8%) 203 (89.4%) 415 (89.6%)
Army 18 (7.6%) 17 (7.5%) 35 (7.6%)
Navy 3 (1.3%) 3 (1.3%) 6 (1.3%)
Air force 3 (1.3%) 2 (0.9%) 5 (1.1%)
Open in a new tab

2.3. Procedures

The presented analysis was based on a large clinical trial dataset funded by the United States Department of Health and Human Services (Grant number: HHS‐2020‐ACL‐AOA‐INNU‐0403). For the purpose of this study, the authors focused on the subset of the original trial data (i.e., treatment groups only excluding the control group) and procedures that related to the aforementioned research questions regarding the applicability of the SRS. In the trial, for both treatment groups (i.e., callers trained in BE or BE + ASIST) as well as for the control group, older adults were contacted and administered baseline assessments, which included assessment of suicide desire (i.e., thwarted belongingness and perceived burdensomeness). Participants' data were scored, assigned a risk tier, and randomly assigned to one of the three conditions. Older adults assigned to either of the two treatment conditions were then matched with a calling partner (i.e., a trained ASN provider who delivered the designated intervention, BE or BE + ASIST). Follow‐up assessments occurred every 2 weeks, resulting in five measurement occasions (baseline plus four follow‐ups). Thwarted belongingness and perceived burdensomeness were assessed at baseline and at each follow‐up. The SRS was collected only following intervention contacts and therefore was not collected at baseline or for control participants.

2.4. Measures

2.4.1. Therapeutic Alliance

The 4‐item Session Rating Scale (SRS; Duncan et al. 2003) was used to assess participants' perceptions of the therapeutic session across four key domains (e.g., “Did you feel heard, understood, and respected by the volunteer?”). Participants marked their responses on a Likert‐Scale, ranging from 0 to 10, representing: (a) the quality of the therapeutic relationship, (b) agreement on goals and topics, (c) the relevance and fit of the therapist's approach, and (d) the overall effectiveness of the session. The scores for each domain were summed to generate a total score. In this study, the 4‐item SRS demonstrated acceptable internal consistency (α = 0.78). The SRS scores, standardized by sample characteristics (i.e., centered on the sample mean and divided by the sample standard deviation), were used as predictors in both models.

2.4.2. Thwarted Belongingness and Perceived Burdensomeness

The Interpersonal Needs Questionnaire (INQ; Van Orden et al. 2012) was used to evaluate thwarted belongingness and perceived burdensomeness through 15 items rated on an adapted 3‐point Likert‐scale version tailored specifically for older adults ranging from 1 (Not at all true for me) to 3 (Very true for me). The instrument includes two sets of characteristics: (1) the 9‐item Thwarted Belongingness (e.g., “I feel disconnected from other people.”); and (2) the 6‐item Perceived Burdensomeness (e.g., “I think my death would be a relief to the people in my life.”). Six questions from the Thwarted Belongingness subscale were reverse‐coded when the overall score across questions was computed due to the nature of those questions. In particular, the total score on the INQ is calculated by summing the questions responses to all items and reversing the appropriate ones. As a result, greater scores on each subscale reflect higher levels of the respective construct being measured. Within each question set in this study both thwarted belongingness and perceived burdensomeness showed good internal consistency (α = 0.88; α = 0.86, respectively).

2.5. Data Analysis

Two binomial generalized linear mixed‐effects models were used for dichotomized binary high‐score responses, with analyses conducted separately for the thwarted belongingness (TB) and perceived burdensomeness (PB) scores. The goal was to evaluate the role of the therapeutic alliance (SRS) in a longitudinal within‐subjects study with two conditions: (1) BE focused on fostering empathy and connection through befriending and reminiscence techniques, and (2) BE + ASIST which included the previous condition and added an aging variant of LivingWorks Applied Suicide Intervention Skills Training (ASIST). Both models (for TB and PB) included a random effect for each individual to account for between‐person variability, along with fixed effects for treatment, standardized SRS score, survey time increments, and their two‐ and three‐way interactions to evaluate the impact of therapeutic alliance on suicidal desire over time. In addition, extended versions of the models were considered for both TB and PB.

For the TB and PB responses, the combined measures (i.e., summed scores) of thwarted belongingness and perceived burdensomeness from the longitudinal within‐groups treatment study were used to produce high (coded as 1) and low (coded as 0) risk responses for the two separate mixed models. More specifically, the sample distributions for both raw summed scores for TB and PB were skewed with a few high values. The use of normal and log‐normal models for both scores was deemed inappropriate since the assumptions of linear models were not satisfied. Instead, participants at or above the 80th percentile were classified as elevated relative to the sample distribution. This approach reflected relative standing within the study sample. To evaluate robustness, more stringent cutoffs were also examined using higher percentile thresholds. For PB, the cutoff score at the 80th percentile was seven, such that scores of seven or higher were coded as elevated perceived burdensomeness. Similarly, for TB, the 80th percentile corresponded to a sum score of 14, such that scores of 14 or higher were coded as elevated thwarted belongingness. These dichotomous indicators were coded 1 for elevated and 0 for low‐to‐moderate at the given survey. Diagnostics were performed for each resulted model to ensure there was no lack of fit.

Missing data were handled within the mixed‐effects modeling framework. Mixed‐effects models were estimated using all completed observations for a given individual. Observations with missing values on the outcome or any model predictors were excluded for that specific model at that timepoint, rather than excluding an entire participant. Because the SRS was only collected after intervention sessions, baseline assessments were not included in models that included the SRS, and small differences in PB versus TB sample sizes reflected occasional skipped items when computing summed scores.

3. Results

3.1. Descriptive Statistics

Descriptive statistics for the combined raw summed scores used to generate binary high‐risk responses are provided in Table 2. Specifically, the combined score for PB averages around 6.5, while the mean score for TB is approximately 12. Both distributions were highly skewed, with a few large values corresponding to individuals in the upper tail. The resulting number of upper quintile responses over time for PB was identified as 385 out of 1523, while for TB it was 449 out of 1460, across all time measurements and individuals. The total number of responses for TB and PB differed since not all questions which form the score were answered by all individuals.

TABLE 2.

Descriptive statistics for perceived burdensomeness and thwarted belongingness by condition and time.

Condition Time n INQ‐PB mean INQ‐PB SD n INQ‐TB mean INQ‐TB SD
BE Baseline 235 6.83 1.93 233 12.28 4.24
BE Survey 1 195 6.67 1.77 196 12.05 4.14
BE Survey 2 184 6.59 1.89 184 11.76 4.08
BE Survey 3 172 6.66 1.74 172 12.11 3.92
BE Survey 4 180 6.49 1.65 181 11.71 3.94
BE + ASIST Baseline 226 6.85 1.9 225 11.96 4.06
BE + ASIST Survey 1 188 6.73 2.05 188 11.63 3.5
BE + ASIST Survey 2 182 6.66 1.89 182 11.4 3.72
BE + ASIST Survey 3 172 6.45 1.66 172 11.36 3.62
BE + ASIST Survey 4 174 6.66 2.06 176 11.26 3.58
Open in a new tab

Note: Survey time points represent sequential assessments.

Abbreviations: INQ‐PB, INQ perceived burdensomeness; INQ‐TB, INQ thwarted belongingness. SD, standard deviation.

To ensure transparency of the initial conditions, z‐tests comparing proportions at baseline were conducted separately for thwarted belongingness and perceived burdensomeness. None were significant, indicating no evidence that the initial baseline conditions differed between the BE and BE WITH groups.

3.2. Model Diagnostics

Model fit diagnostics were performed using the DHARMa package in R, and the checks did not indicate any deficiencies. The use of the proposed models was therefore deemed appropriate. The performed checks included Kolmogorov–Smirnov tests, Q‐Q plots for simulated residuals, overdispersion checks, zero‐inflation checks, and outlier checks. None of these tests indicated any model deficiencies. These findings collectively suggest that key assumptions were not violated, and the utilized binomial GLMMs offered an appropriate framework for modeling change in elevated TB or PB classification over time while accounting for the nested data structure within individuals.

3.3. Main Effects

For perceived burdensomeness, the complete model with all interaction terms identified a significant main effect of time (OR = 0.69, 95% CI [0.51; 0.93], p = 0.016) keeping other predictors in the model fixed (see Table 3), indicating that the odds of being classified in the elevated perceived burdensomeness group decreased by 31% with each unit increase in time. However, the main effects of intervention condition (i.e., BE + ASIST) as well as SRS were not statistically significant (p > 0.05), indicating that there is no evidence that either differences between the proposed interventions or the quality of therapeutic alliance are associated with high perceived burdensomeness. Likewise, for thwarted belongingness, no significant main effects were found in the complete model with all interaction terms for time, intervention condition, and SRS when each was considered alone. There is no evidence from the data that time, intervention condition, and therapeutic alliance are associated with high thwarted belongingness.

TABLE 3.

Main and interaction effects from binomial linear mixed‐effects models.

Fixed effects Perceived burdensomeness Thwarted belongingness
Odds ratio 95% CI p Odds ratio 95% CI p
Intercept 0.00 [0.00, 0.01] < 0.001*** 0.00 [0.00, 0.01] < 0.001***
Condition 1.04 [0.25, 4.39] 0.958 1.80 [0.42, 7.61] 0.428
Time 0.69 [0.51, 0.93] 0.016* 0.86 [0.65, 1.14] 0.297
SRS 1.02 [0.38, 2.73] 0.963 0.44 [0.16, 1.19] 0.106
Condition × Time 0.98 [0.63, 1.52] 0.932 0.65 [0.43, 1.00] 0.048*
Condition × SRS 1.36 [0.36, 5.18] 0.648 3.93 [1.12, 13.83] 0.033*
Time × SRS 1.08 [0.72, 1.65] 0.684 1.26 [0.85, 1.88] 0.245
Condition × Time × SRS 0.71 [0.41, 1.22] 0.216 0.59 [0.36, 0.96] 0.034*
Open in a new tab

Note: Condition = BE + ASIST condition. *p < 0.05, **p < 0.01, ***p < 0.001.

3.4. Interaction Effects

Significant interaction effects were identified in the thwarted belongingness model. A significant three‐way interaction among condition, time, and SRS was found (OR = 0.59, 95% CI: [0.36; 0.96], p = 0.034) (see Table 3), indicating that change over time in the odds of elevated thwarted belongingness differed by condition and varied as a function of SRS. Because this higher‐order interaction was present, the two‐way interaction terms were interpreted as conditional effects at the reference level of the third variable. Within this context, the interaction between treatment condition (i.e., BE + ASIST) and time was significant (OR = 0.65, 95% CI: [0.43; 1.00], p = 0.048) (see Table 3), indicating that the time‐related decrease in the odds of elevated thwarted belongingness was stronger in the BE + ASIST intervention group than in BE intervention group when SRS was held at its reference value. Additionally, a significant interaction effect between treatment conditions and standardized SRS was found for thwarted belongingness (OR = 3.93, 95% CI: [1.12; 13.83], p = 0.033) (see Table 3), indicating that the association between SRS and the odds of elevated thwarted belongingness differed between conditions when time is held at its reference value. In contrast, for perceived burdensomeness, none of the two‐way or three‐way interaction terms were statistically significant (p > 0.05).

4. Discussion

Our study findings extend prior research on the importance of therapeutic alliance, largely established in professional therapy contexts, by demonstrating its moderating role in paraprofessional‐delivered interventions for older adults. Consistent with literature linking therapeutic alliance to suicide‐related outcomes (e.g., Dunster‐Page et al. 2017; Huggett et al. 2021; Staniforth et al. 2023), the present study identified that higher SRS scores shaped the trajectory of treatment effects over time, particularly by mitigating thwarted belongingness. Although SRS scores alone were not associated with perceived burdensomeness or thwarted belongingness, the alliance significantly interacted with intervention condition and time to predict changes in the odds of being classified in the upper tail of thwarted belongingness scores, a factor theorized to underlie late‐life suicide desire (Joiner 2005; Van Orden et al. 2010). Importantly, these classifications reflect relative elevation within this sample distribution rather than a clinical risk threshold. Overall, the results highlight therapeutic alliance as a process‐oriented factor that may strengthen paraprofessional‐delivered intervention effects over time on interpersonal indicators of suicide desire.

Administering the SRS at each session within BE WITH interventions (i.e., BE and BE + ASIST) can help paraprofessionals evaluate client satisfaction, identify areas for improvement, foster collaboration, and strengthen therapeutic relationships. Specifically, the significant three‐way interaction among intervention condition, time, and SRS provides insight into relational mechanisms of change. Older adult clients in the BE + ASIST condition who reported higher‐quality therapeutic sessions showed greater reductions in the odds of being classified in the upper tail of thwarted belongingness over time, underscoring that improvements were not due to time or intervention alone but depended on the sustained presence of a strong therapeutic alliance. ASIST training, designed to equip natural helpers with suicide‐specific skills (e.g., collaborative safety planning, recognizing ambivalence about dying, respecting choices about life and death), may have further fostered supportive, empathic relationships. In turn, these skills likely enhanced participants' sense of being respected, connected, and collaboratively engaged.

For older adults who often face loss, isolation, and stigma related to mental health, such relational attunement may be particularly critical. Supportive and trustworthy paraprofessionals can validate older adult clients' worth and belonging, directly countering interpersonal risk factors of suicide desire. The current research findings suggest that the effectiveness of suicide interventions depends not only on the interventions but also on the quality of the helper‐client relationship. This interdependence underscores the importance of relational competence, particularly when working with relationally sensitive groups, such as homebound older adults.

In addition, the odds of being classified in the elevated perceived burdensomeness group decreased over time, regardless of condition or therapeutic alliance. This finding suggests that repeated engagement with a paraprofessional helper through consistent social contact and structured conversation may offer therapeutic benefits in decreasing the internalized sense of being a burden. The time effect may reflect the cumulative impact of ongoing connection and support, even in the absence of significant relational depth or therapeutic alliance strength. This finding underscores the value of simply showing up for older adult clients, especially those who are homebound or socially isolated, and highlights the potential of brief, consistent interventions to alleviate internalized suicide risk.

Importantly, the current research also demonstrated a significant interaction between intervention condition and time in predicting thwarted belongingness. Specifically, older adults in the BE + ASIST condition experienced a greater time‐related decrease in the odds of elevated thwarted belongingness compared to those in the BE condition. This pattern indicates that the addition of suicide‐specific training (ASIST) may have enhanced paraprofessionals' ability to foster meaningful social connection, beyond what was offered through the BE condition alone. While both groups received warm‐calling interventions, the added structure, guidance, and relational techniques offered through ASIST appear to have created a more impactful relational experience. These findings underscore the importance of not just relational contact but also relational competence in facilitating change, especially in domains related to social disconnection and suicide desire.

Finally, the finding that BE + ASIST participants with higher therapeutic alliance scores showed greater odds of being in the high‐risk group for thwarted belongingness (when time was not modeled) may reflect a potential Simpson's paradox wherein the apparent relationship between SRS and risk status is reversed or misleading when the time dimension is not taken into account. Specifically, this interaction was observed in the absence of the time component, which was later shown to significantly moderate the effect of SRS on thwarted belongingness through a three‐way interaction. One possible explanation is that the BE + ASIST group included participants who began the intervention with more severe interpersonal risk profiles, thus inflating the odds of high‐risk classification at baseline, even among those with stronger therapeutic alliance perceptions. When accounting for change over time, however, the three‐way interaction revealed that higher therapeutic alliance in the BE + ASIST group was actually associated with greater decrease in thwarted belongingness. These findings highlight the significance of interpreting interaction within the full temporal context and considering baseline group differences that may obscure the direction of effects when not modeled longitudinally.

4.1. Limitations

Results from the current investigation offer insight into the importance of therapeutic alliance; nevertheless, they need to be interpreted within the context of the study's limitations. First, although the study was positioned within the context of older adult suicide prevention, the outcomes examined were interpersonal indicators of suicide desire, specifically thwarted belongingness and perceived burdensomeness. The present study did not include direct assessment of suicide risk outcomes (e.g., suicidal ideation severity, suicide attempts, and/or other indicators of actual risk). Accordingly, the findings should not be interpreted as demonstrating reduced suicide risk or reduced risk for death by suicide. Second, paraprofessional callers were recommended by ASN personnel based on strong helper characteristics, such as warmth and empathy. This selection process may limit generalizability to paraprofessional workforces that do not use similar screening or that include helpers with a broader range of baseline relational strengths. Third, variability in callers' skills and relational capacity may have introduced provider effects, such that some paraprofessionals may have been more effective than others in establishing and maintaining a strong therapeutic alliance. Fourth, therapeutic alliance was assessed from the participant's perspective using the Session Rating Scale (SRS) affording an “other rated” measure of the therapeutic relationship, as felt and experienced by the participants. However, additional sources of alliance data (e.g., observer ratings from supervisors of the trained professionals) could provide unique triangulation of data and more comprehensive measures of relational quality. Finally, the demographics of our sample (e.g., predominantly women, Black, lower education, lower income) may limit the generalizability of these findings.

Demographic covariates were briefly considered in the presented models one at a time and in the resulting expanded models showed limited and inconsistent associations with the outcomes. Age, gender, firearm access, and military service were not significant predictors in the models. Education at the master's and doctorate levels was associated with reduced odds; however, very few participants in the overall dataset fell into these categories, raising concerns about the stability and reliability of these estimates. Although for predictors such as religion, race, and income level estimates were statistically significant, the wide confidence intervals suggest substantial uncertainty, and these findings should therefore be interpreted with caution and not presented in the results section.

4.2. Implications for Practice

These findings demonstrate the potential benefit for standardized and manualized innovative treatment programming for older adults, and they underscore the critical role of the therapeutic alliance in predicting and enhancing treatment outcomes. These findings support the documented efficacy of relationship‐based interventions such as BE WITH for older adults receiving Home and Community‐Based Services (HCBS), as well as training paraprofessionals in the ASN in manualized protocols that foster reciprocally caring relationships, trust, empathy, and collaboration. The ASN may consider investing in this significant volunteer workforce (i.e., home delivered meal providers, information and referral services, assessment coordinators, caregiver support, etc.) in meeting the mental health needs of older adults in the face of reduced funding, and workplace cuts. These findings align with prior research emphasizing the importance of therapeutic alliance in interventions that target suicide‐related outcomes and interpersonal indicators of suicide desire. A strong therapeutic alliance may support reductions in interpersonal indicators of suicide desire, particularly thwarted belongingness, by reinforcing the need for warm, consistent, and responsive relational support provided by trained paraprofessionals to the highest risk age cohort for suicide, older adults.

These findings also support the use of the Session Rating Scale (SRS) for real‐time monitoring of therapeutic relationships and treatment quality. Incorporating measures such as the SRS into paraprofessional practice may enhance treatment fidelity, which may increase the likelihood of achieving meaningful improvements in mental health outcomes such as perceived burdensomeness and thwarted belongingness, particularly in underserved and isolated populations such as older adults with chronic health conditions and mobility issues. In addition, when time and treatment conditions are factored in, it makes good sense that older adults who received treatment from paraprofessionals trained in BE + ASIST emerged with significantly reduced thwarted belongingness, and this supports the use of training the aging variant of LivingWorks ASIST in this context. This standardized, manualized, and evidence‐based program may be implemented within the volunteer workforce of the Aging Services Network to promote life with older adults at a time needed the most.

4.3. Implications for Future Research

The current findings support further investigation into the role of the therapeutic alliance within paraprofessional‐delivered interventions. While the SRS has previously shown utility in clinical settings, this is the first study to our knowledge that evaluates its application and validity in paraprofessional contexts. Additional research is needed to corroborate these findings. Future research could examine the psychometric properties and sensitivity of the SRS in other diverse, nonclinical populations, as well as its ability to predict meaningful treatment outcomes across varying degrees of intervention intensity. Additionally, fluctuations in therapeutic alliance over time may impact client engagement, well‐being, and suicide risk, especially among older adults experiencing chronic isolation; therefore, closer examinations at particular measurement occasions in similar longitudinal studies merit further attention. Future longitudinal clinical trials may incorporate more routine feedback and therapeutic alliance scores to trained paraprofessionals, allowing these treatment providers to make timely adjustments and even prevent treatment failure. Research may also investigate the specific mechanisms by which paraprofessionals foster effective therapeutic alliances through the use of recording treatment dosages to better understand the unique relational components that most strongly contribute to therapeutic alliance development. Finally, particular therapeutic alliance factors such as the therapeutic bond, shared goals/hopes for participation, perceived empathy, etc. may or may not influence reductions in suicide desire, and further examination could provide valuable insight into tailoring interventions for at‐risk populations.

5. Conclusion

In sum, the research findings provide evidentiary support that therapeutic alliance is not merely an ancillary feature of intervention, but a core mechanism of change even in nonclinical, community‐based treatment settings. The results support ongoing investment in paraprofessional training that centers relational quality alongside content fidelity and suggest that alliance‐monitoring tools like the SRS can serve as practical instruments to ensure treatment responsiveness and success. As the demand for accessible, community‐based mental health support continues to grow, especially among aging populations with limited access to traditional therapy, ongoing research on the implementation of alliance‐focused, relationship‐based practices remains critical for improving the quality and effectiveness of care, and therapeutic alliance should be prioritized as both an outcome predictor and intervention enhancer.

Funding

This work was supported by the U.S. Department of Health and Human Services (90INNU0049‐01‐00).

Ethics Statement

This study was approved by the Georgia State University Institutional Review Board (IRB H21359) and conducted in accordance with the ethical guidelines.

Consent

All participants were provided written informed consent prior to participation, and confidentiality was maintained throughout the study.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgments

This project is supported by the Administration for Community Living (ACL), U.S. Department of Health and Human Services (HHS) as part of a financial assistance award totaling $4,107,807 (Grant# 90INNU0049‐01‐00) with 73% funded by ACL/HHS and $1,107,918 million amount and 27% funded by nongovernment source(s). The contents are those of the author(s) and do not necessarily represent the official views of, nor an endorsement by, ACL/HHS, or the U.S. Government.

Data Availability Statement

Research data are not shared.

References

  1. Bordin, E. S. 1979. “The Generalizability of the Psychoanalytic Concept of the Working Alliance.” Psychotherapy: Theory, Research & Practice 16, no. 3: 252–260. 10.1037/h0085885. [DOI] [Google Scholar]
  2. Boschann, A. , Staats H., and Wiegand‐Grefe S.. 2024. “Older Patients' Perspectives on the Therapeutic Relationship With Young Psychotherapists.” PLoS One 19, no. 5: e0295834. 10.1371/journal.pone.0295834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Conwell, Y. , Van Orden K. A., and Caine E. D.. 2011. “Suicide in Older Adults.” Psychiatric Clinics of North America 34, no. 2: 451–468. 10.1016/j.psc.2011.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Crosby, E. S. , Zuromski K. L., and Witte T. K.. 2020. “Perceived Burdensomeness Is a Curvilinear, Short‐Term Predictor of Suicide Ideation in a Community Sample of Adults.” Suicide and Life‐Threatening Behavior 50, no. 6: 1205–1213. 10.1111/sltb.12716. [DOI] [PubMed] [Google Scholar]
  5. Dienst, F. , Forkmann T., Schreiber D., and Höller I.. 2023. “Attachment and Need to Belong as Moderators of the Relationship Between Thwarted Belongingness and Suicidal Ideation.” BMC Psychology 11, no. 1: 50. 10.1186/s40359-023-01080-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Drapeau, C. W. , and McIntosh J. L.. 2025. “U.S.A. Suicide: 2023 Official Final Data.” https://jmcintos.pages.iu.edu/2023datapgsv1a.pdf.
  7. Duncan, B. L. , Miller S. D., Sparks J. A., et al. 2003. “The Session Rating Scale: Psychometric Properties of a “Working” Alliance Measure.” Journal of Brief Therapy 3, no. 1: 3–12. [Google Scholar]
  8. Dunster‐Page, C. , Haddock G., Wainwright L., and Berry K.. 2017. “The Relationship Between Therapeutic Alliance and Patient's Suicidal Thoughts, Self‐Harming Behaviours and Suicide Attempts: A Systematic Review.” Journal of Affective Disorders 223: 165–174. 10.1016/j.jad.2017.07.040. [DOI] [PubMed] [Google Scholar]
  9. Fakoya, O. A. , McCorry N. K., and Donnelly M.. 2021. “How Do Befriending Interventions Alleviate Loneliness and Social Isolation Among Older People? A Realist Evaluation Study.” PLoS One 16, no. 9: e0256900. 10.1371/journal.pone.0256900. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fartacek, C. , Kunrath S., Aichhorn W., and Plöderl M.. 2023. “Therapeutic Alliance and Change in Suicide Ideation Among Psychiatric Inpatients at Risk for Suicide.” Journal of Affective Disorders 323: 793–798. 10.1016/j.jad.2022.12.028. [DOI] [PubMed] [Google Scholar]
  11. Fullen, M. C. , Shannonhouse L. R., Mize M. C., and Miskis C.. 2020. “Mental Health Distress in Homebound Older Adults: Importance of the Aging Network.” Aging & Mental Health 25, no. 8: 1580–1584. 10.1080/13607863.2020.1758920. [DOI] [PubMed] [Google Scholar]
  12. Gaston, L. 1991. “Reliability and Criterion‐Related Validity of the California Psychotherapy Alliance Scales—Patient Version.” Psychological Assessment: A Journal of Consulting and Clinical Psychology 3, no. 1: 68–74. 10.1037/1040-3590.3.1.68. [DOI] [Google Scholar]
  13. Gelso, C. J. , and Carter J. A.. 1985. “The Relationship in Counseling and Psychotherapy: Components, Consequences, and Theoretical Antecedents.” Counseling Psychologist 13, no. 2: 155–243. 10.1177/0011000085132001. [DOI] [Google Scholar]
  14. Goger, P. , Guo R., Jo S., et al. 2025. “Moderators and Mediators of Treatments for Suicidal Thoughts and Behaviors: A Review of More Than 50 Years of Randomized Controlled Trials.” Behavior Therapy 56, no. 2: 241–260. 10.1016/j.beth.2024.10.003. [DOI] [PubMed] [Google Scholar]
  15. Gould, D. , Kelly D., Goldstone L., and Gammon J.. 2001. “Examining the Validity of Pressure Ulcer Risk Assessment Scales: Developing and Using Illustrated Patient Simulations to Collect the Data.” Journal of Clinical Nursing 10, no. 5: 697–706. 10.1046/j.1365-2702.2001.00525.x. [DOI] [PubMed] [Google Scholar]
  16. Gould, M. S. , Cross W., Pisani A. R., Munfakh J. L., and Kleinman M.. 2013. “Impact of Applied Suicide Intervention Skills Training on the National Suicide Prevention Lifeline.” Suicide & Life‐Threatening Behavior 43, no. 6: 676–691. 10.1111/sltb.12049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Guidry, E. T. , and Cukrowicz K. C.. 2016. “Death Ideation in Older Adults: Psychological Symptoms of Depression, Thwarted Belongingness, and Perceived Burdensomeness.” Aging & Mental Health 20, no. 8: 823–830. 10.1080/13607863.2015.1040721. [DOI] [PubMed] [Google Scholar]
  18. Hafkenscheid, A. , Duncan B. L., and Miller S. D.. 2010. “The Outcome and Session Rating Scales: A Cross‐Cultural Examination of the Psychometric Properties of the Dutch Translation.” Journal of Brief Therapy 7, no. 1 & 2: 1–12. [Google Scholar]
  19. Hill, R. M. , and Pettit J. W.. 2014. “Perceived Burdensomeness and Suicide‐Related Behaviors in Clinical Samples: Current Evidence and Future Directions.” Journal of Clinical Psychology 70, no. 7: 631–643. 10.1002/jclp.22071. [DOI] [PubMed] [Google Scholar]
  20. Huggett, C. , Gooding P., Haddock G., and Pratt D.. 2021. “The Relationship Between the Therapeutic Alliance and Suicidal Experiences in People With Psychosis Receiving Therapy.” International Journal of Environmental Research and Public Health 18, no. 20: 10706. 10.3390/ijerph182010706. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Joiner, T. E. 2005. Why People Die by Suicide. Harvard University Press. [Google Scholar]
  22. Klimek, L. , Bergmann K. C., Biedermann T., et al. 2017. “Visual Analogue Scales (VAS): Measuring Instruments for the Documentation of Symptoms and Therapy Monitoring in Cases of Allergic Rhinitis in Everyday Health Care.” Allergo Journal International 26, no. 1: 16–24. 10.1007/s40629-016-0006-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lang, W. A. , Ramsay R. F., Tanney B. L., Kinzel T., Turley B., and Tierney R. J.. 2013. ASIST Trainer Manual. 11th ed. LivingWorks Education Incorporated. [Google Scholar]
  24. Mace, R. A. , Gansler D. A., Suvak M. K., et al. 2018. “Therapeutic Relationship in the Treatment of Geriatric Depression With Executive Dysfunction.” Journal of Affective Disorders 214: 130–137. 10.1016/j.jad.2017.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Marti, C. N. , Kunik M. E., and Choi N. G.. 2020. “The Reciprocal Relationship Between Depression and Disability in Low‐Income Homebound Older Adults Following Tele‐Depression Treatment.” International Journal of Geriatric Psychiatry 36, no. 6: 802–810. 10.1002/gps.5480. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Martin, D. J. , Garske J. P., and Davis K. M.. 2000. “Relation of the Therapeutic Alliance With Outcome and Other Variables: A Meta‐Analytic Review.” Journal of Consulting and Clinical Psychology 68, no. 3: 438–450. 10.1037/0022-006X.68.3.438. [DOI] [PubMed] [Google Scholar]
  27. Marty, M. A. , Segal D. L., and Coolidge F. L.. 2010. “Relationships Among Dispositional Coping Strategies, Suicidal Ideation, and Protective Factors Against Suicide in Older Adults.” Aging & Mental Health 14, no. 8: 1015–1023. 10.1080/13607863.2010.501068. [DOI] [PubMed] [Google Scholar]
  28. Montgomery, E. C. , Kunik M. E., Wilson N., Stanley M. A., and Weiss B.. 2010. “Can Paraprofessionals Deliver Cognitive‐Behavioral Therapy to Treat Anxiety and Depressive Symptoms?” Bulletin of the Menninger Clinic 74, no. 1: 45–62. [DOI] [PubMed] [Google Scholar]
  29. Rasmussen, K. A. , and Wingate L. R.. 2011. “The Role of Optimism in the Interpersonal‐Psychological Theory of Suicidal Behavior.” Suicide and Life‐Threatening Behavior 41, no. 2: 137–148. 10.1111/j.1943-278x.2011.00022.x. [DOI] [PubMed] [Google Scholar]
  30. Rodgers, P. 2010. Review of the Applied Suicide Intervention Skills Training Program (ASIST): Rationale, Evaluation Skills, and Direction for Future Research. LivingWorks Education Inc. www.livingworks.net/dmsdocument/274. [Google Scholar]
  31. Schechter, M. , Ronningstam E., Herbstman B., and Goldblatt M. J.. 2019. “Psychotherapy With Suicidal Patients: The Integrative Psychodynamic Approach of the Boston Suicide Study Group.” Medicina 55, no. 6: 303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shannonhouse, L. , Fullen M., Leroux A., Franck Love S., and Westcott J.. 2022. “Addressing Social Isolation, Loneliness, and Elevated Suicidality Among Older Adults During Covid‐19 and Beyond: A Partially Nested Randomized Control Trial of the BE WITH Innovation.” (Research Brief Requested by the Administration of Community Living). https://acl.gov/sites/default/files/nutrition/BE‐WITH_ResearchBrief_508.pdf.
  33. Shannonhouse, L. , Snyder J. D., Fullen M., et al. 2021. “Resource Loss, Religious Coping, and Suicide Predictors/Outcomes During the COVID‐19 Pandemic: A Prospective Study of U.S. Racially Diverse Older Adults.” Journal of Psychology and Christianity 39, no. 4: 313–327. [Google Scholar]
  34. Staniforth, B. , Ali B., and Adamson C.. 2023. “‘I Just Want You to Listen’: People Who Have Experienced Suicidal Ideation/Attempts Talk About What They Want From Their Crisis Teams.” Qualitative Social Work 23, no. 6: 951–969. 10.1177/14733250231212114. [DOI] [Google Scholar]
  35. Sterling‐Fox, C. 2018. “Access to Five Nonprimary Health Care Services by Homebound Older Adults: An Integrative Review.” Home Health Care Management & Practice 31, no. 1: 55–69. 10.1177/1084822318810384. [DOI] [Google Scholar]
  36. Thompson, L. W. , Davies R. D., Gallagher D., and Krantz S.. 1986. “Cognitive Therapy With Older Adults.” Clinical Gerontology 15: 245–279. [Google Scholar]
  37. Van Orden, K. A. , Cukrowicz K. C., Witte T. K., and Joiner T. E. Jr. 2012. “Thwarted Belongingness and Perceived Burdensomeness: Construct Validity and Psychometric Properties of the Interpersonal Needs Questionnaire.” Psychological Assessment 24, no. 1: 197–215. 10.1037/a0025358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Van Orden, K. A. , Witte T. K., Cukrowicz K. C., Braithwaite S. R., Selby E. A., and Joiner T. E. Jr. 2010. “The Interpersonal Theory of Suicide.” Psychological Review 117, no. 2: 575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Westcott, J. B. , Fullen M. C., Tomlin C. C., et al. 2022. “‘Listen Closer’: Home‐Delivered Meal Volunteers' Understanding of Their Role in Suicide Intervention.” Ageing and Society 44, no. 3: 642–660. 10.1017/S0144686X22000368. [DOI] [Google Scholar]
  40. Wiles, J. , Morgan T., Moeke‐Maxwell T., et al. 2019. “Befriending Services for Culturally Diverse Older People.” Journal of Gerontological Social Work 62, no. 7: 776–793. 10.1080/01634372.2019.1640333. [DOI] [PubMed] [Google Scholar]
  41. Wyman, P. A. , Brown C. H., Inman J., et al. 2008. “Randomized Trial of a Gatekeeper Program for Suicide Prevention: 1‐Year Impact on Secondary School Staff.” Journal of Consulting and Clinical Psychology 76, no. 1: 104–115. 10.1037/0022-006X.76.1.104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Yousefi, Z. , Sharifi K., Tagharrobi Z., and Akbari H.. 2015. “The Effect of Narrative Reminiscence of Elderly Women.” Iranian Red Crescent Medical Journal 17, no. 11: e19612. 10.5812/ircmj.19612. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Research data are not shared.

Articles from Suicide & Life-Threatening Behavior are provided here courtesy of Wiley

RESOURCES