Adapting a Suicide Prevention Program for People with Psychosis Following a Psychiatric Hospitalization: Pilot Results and Lessons Learned

Abstract

Individuals with schizophrenia-spectrum disorders (SSD) are at particularly high risk for suicide, with the period immediately following a psychiatric hospitalization identified as a time of the highest suicide risk. Currently, there are limited evidence-based psychosocial interventions specifically designed to reduce suicide risk in patients with SSD following a hospitalization. The Coping Long Term with Active Suicide Program (CLASP) is a multifaceted, adjunctive suicide intervention that specifically focuses on acute transitions of care, making it a particularly promising option for people with SSD. This pilot feasibility randomized controlled trial adapted CLASP for patients with SSDs post-hospital discharge for up to 6 months. Findings indicated recruitment feasibility, success in engaging most patients in the program in the initial weeks immediately following their discharge when risk is highest, and preliminary support for reducing suicide behaviors and psychiatric symptoms. We describe important lessons learned about how to further refine the intervention to increase feasibility and acceptability to patients in future research.

Introduction

As the ninth leading cause of death in the United States for people ages 10–64, suicide is a significant public health concern (SAMHSA, 2023). Those with schizophrenia-spectrum disorders (i.e., schizophrenia, schizophreniform disorder, schizoaffective disorder, psychotic disorder unspecified; SSD) are at particularly high risk for suicide. Approximately 36% of adults with SSD in North America report a lifetime suicide attempt (Lu et al., 2019), compared with 2–9% of the general population (Nock et al., 2008). Furthermore, the suicide death rate among individuals with SSD is 4.5 times higher than that of the general population (Olfson et al., 2021). Although it is imperative to focus on this population for suicide prevention and treatment initiatives, people experiencing psychosis tend to be excluded from suicide intervention studies (Villa et al., 2020). This is particularly concerning given that research suggests that people with SSD may have unique needs (e.g., impaired cognitive functioning, increased mental health stigma; Brohan et al., 2010; Dubreucq et al., 2021; Fioravanti et al., 2012), which may limit the efficacy of existing suicide interventions for this population (Chalker et al., 2022; Rabasco et al., 2024).

Among those with SSD, the period immediately following psychiatric hospitalization is a time of heightened suicide risk (Olfson et al., 2016). In fact, one-third of suicides in patients with SSD occur immediately after hospitalization, and the risk remains high for up to one-year post-discharge (Popovic et al., 2014). Although there are promising psychosocial interventions studied for reducing suicidal thoughts and behaviors in people with psychosis in the community (Bornheimer et al., 2020), there are no well-established treatments specifically developed to reduce suicide risk in patients with SSD following hospital discharge, resulting in a significant gap in practice. The Coping Long Term with Active Suicide Program (CLASP) is a multifaceted, adjunctive suicide intervention that is delivered during acute transitions of care (Miller et al., 2022), making it a particularly promising option for patients with SSD. However, the intervention has not yet been adapted to address the unique needs of those with SSD.

CLASP facilitates suicide prevention through a combination of case management, individual psychotherapy, and optional significant other (SO; e.g., partner, parent, sibling, adult child, close friend) involvement. Originally, CLASP was developed to include brief in-person and telephone sessions delivered up to six months after hospitalization. In recent years, it has incorporated more of a telehealth format. CLASP is a patient-centered approach that helps individuals clarify and work toward individualized treatment goals that are consistent with their personal values, in the service of building a life that they want to live. Patient sessions focus on reducing suicide risk through identifying individualized risk factors, clarifying personal values and goals, planning for the future, crafting a suicide safety plan, encouraging treatment engagement and adherence, and facilitating patient and SO communication and informal problem solving (Miller et al., 2017). Using this approach, CLASP has been shown to reduce suicide attempts by 20% in patients with suicidal thoughts and behaviors following an emergency department visit (Miller et al., 2017) and by 43% following a psychiatric hospital discharge (Miller et al., 2016).

However, recent research on the traditional administration of CLASP suggests that it may be less efficacious for those with SSD (Rabasco et al., 2024), underscoring the utility of adapting CLASP for this population. To expand the CLASP intervention to the full spectrum of psychiatrically hospitalized patients, there is a critical need to examine this program for patients with SSD. Thus, the current study was a pilot randomized controlled trial (RCT) to investigate the feasibility and preliminary effects of CLASP among individuals with SSD. Specifically, this study had the following aims: 1) to adapt the treatment manual for the Coping Long-Term with Active Suicide Program for Schizophrenia (CLASP-S), an adjunctive intervention combining brief individual and telephone patient contacts with SO support to facilitate the safer transition from hospital to outpatient care and prevent subsequent suicidal behaviors; and 2) to conduct a pilot RCT of enhanced treatment as usual (ETAU) alone vs. CLASP-S for patients (N = 30) initially hospitalized with SSD and suicidality to assess feasibility and acceptability post-hospital discharge. Given the pilot nature of the trial, we did not have the statistical power to test for groups differences between conditions. However, we hypothesized that we would be able to recruit and retain sufficient patients into the study as planned, patients would engage in the proposed study sessions, study-related adverse events would be absent, and patients would demonstrate significant within-subject improvements in clinical outcomes.

Methods

CLASP-S

Rationale.

CLASP-S was adapted from the original CLASP intervention, which is described in detail elsewhere (Miller et al., 2016, 2022). Briefly, CLASP is an adjunctive intervention with three major components: 1) three longer initial meetings, 2) one meeting with a supportive significant other or family member, 3) and 11 brief follow-up phone contacts with the patient and their support person. CLASP was designed based on a “risk reduction” model because suicide is multi-determined (Cockcroft et al., 2001). By targeting and reducing modifiable risk factors for suicide, CLASP aims to prevent future suicidal behaviors. We adapted CLASP to focus on risk factors that were also theoretically related to risk in psychosis and were potentially changeable by our intervention (Potkin et al., 2003), including hopelessness (Grano et al., 2013; Klonsky et al., 2012), problem-solving deficits (Breitborde et al., 2021; Darvishi et al., 2023), treatment nonadherence (Hawton et al., 2005; Kreyenbuhl et al., 2009), positive psychotic symptoms (Bornheimer & Jaccard, 2017; de Cates et al., 2021), depression (Hawton et al., 2005), expressed emotion (Ellis et al., 2014), and comorbid substance use problems (Hawton et al., 2005). More specifically, systematic reviews and meta-analyses have indicated that, among people with SSD, depression (OR = 3.03), substance misuse (OR = 3.21), and treatment nonadherence (OR = 3.75) are all significantly associated with increased risk for suicide (Hawton et al., 2005). Furthermore, people with SSD have high levels of hopelessness (White et al., 2007), and longitudinal research shows that hopelessness predicts attempted suicide among people with psychotic disorders (Klonsky et al., 2012). People with SSD also struggle with problem solving (Morris et al., 1995; Sheffield et al., 2018), and meta-analytic evidence suggests that problem solving skills are negatively associated with suicidal behaviors (Darvishi et al., 2023). Finally, positive psychotic symptoms have been associated with suicidal ideation and behaviors among adults with SSD (Bornheimer & Jaccard, 2017; de Cates et al., 2021), and high parental expressed emotion is associated with current suicidal ideation among young people with serious mental illness (Ellis et al., 2014). Additionally, we tailored the CLASP strategies to improve their feasibility and acceptability for patients with SSD (Bradshaw et al., 2005).

Theoretical Basis.

CLASP-S integrated strategies from Acceptance and Commitment Therapy (ACT; Hayes et al., 2012) and brief family intervention for chronic illness (Bishop et al., 2014). The values clarification component of ACT was used to help patients identify important life values (e.g., family, friendships, romantic relationships, work, education) and develop linked short-term treatment goals. The therapist explored potential discrepancies between patients’ values and their daily actions and collaborated with the patient to develop a specific behavioral plan encouraging values consistency over time. ACT is effective when adapted for treating psychosis (Morris et al., 2024), and family interventions have been used to improve problem-solving among patients with chronic illnesses and their significant others (Cockcroft et al., 2001). We based our protocol on a family intervention that included an in-person family meeting followed by a series of brief telephone contacts between the therapist, patient, and SO (Bishop et al., 2014). The goal of family support was to assess and address problem areas (e.g., safety), encourage social support, decrease expressed emotion, and facilitate active problem-solving to target risk factors and support goal attainment.

Format.

CLASP-S was designed to be a hybrid between more intensive in-person only therapy versus brief case management alone. Thus, it is used as an adjunctive therapy to support and extend patients’ other treatments, such as medication management, individual and group psychotherapy, and long-term case management. The CLASP-S format was similar to the original CLASP intervention, with tailored content specifically designed to address the needs of patients with SSD (see details below). In this study, CLASP-S consisted of the following elements: a) 3 initial sessions (lasting 40 – 50 minutes) beginning during hospitalization (when time permitted) and completed within the first week or two post-discharge, b) 1 in-person meeting with both the patient and SO held after discharge, and c) 11 brief follow-up phone contacts (15–30 minutes) held separately with the patient and their SO (when available) for up to a 6-month period, when suicidality risk is heightened. If no SO was available, then these SO sessions were excluded. Sessions were held in person or remotely via phone or secure videoconference depending on circumstances and preferences of the patient and SO.

Initial Patient Sessions.

Meeting 1 focused on rapport building, history taking, and psychoeducation. Suicide risk and safety plan (Stanley & Brown, 2012) were also reviewed in the first session. The safety plan consisted of the following: identifying warning signs; listing internal coping strategies, social situations and people who can provide distraction, and people and professionals to ask for help; reviewing ways to make the environment safe; and identifying reasons for living (Stanley & Brown, 2012). Values clarification was briefly introduced to motivate change and instill hopefulness (Wilson & Murrell, 2004). In Meeting 2, the therapist reviewed the values clarification discussion using ACT strategies (e.g., the use of the Compass metaphor) and helped the patient define a list of short-term, values-consistent goals. The majority of Meeting 3 was spent formulating an individualized “Life Plan.” The Life Plan provided dedicated space to record the patient’s previously specified values and accompanying goals, documenting specific actions to help the patient achieve greater consistency between values and behaviors, as well as linking values and goals to the patient’s previously developed safety plan to address any future suicidality.

Joint Patient-SO Session.

Given the importance of family-based interventions for schizophrenia and the deleterious effect of expressed emotion (Dixon et al., 2000), Meeting 4 included the therapist, patient, and an SO. The therapist reviewed psychoeducational information regarding suicide and encouraged the SO to discuss his/her perspectives on the patient’s recent hospitalization, in a safe and constructive way that decreased the tendency for criticism and hostility (i.e., expressed emotion). The majority of the session was spent reviewing and discussing the patient’s Life Plan. The therapist educated the patient and SO about the negative role of expressed emotion (i.e., criticism, hostility, and over-involvement) and how to foster healthy communication styles. The therapist facilitated a discussion between the SO and the patient about the SO’s role in the treatment plan to support the patient’s goals. If severe patient-SO conflict/expressed emotion was identified, referrals for formal family therapy were provided. If patients were unable or unwilling to identify an SO to participate, they continued to receive phone sessions individually.

Follow-up Contacts.

After the initial sessions, the therapist conducted follow-up contacts with the patient and their SO individually. Each contact was 15–30 minutes in length, based on the number and severity of clinical issues identified. Sessions occurred via phone or secure video conferencing (i.e., telehealth) on a schedule of decreasing frequency for up to 11 sessions over 6 months, based on need. At the start of each contact, the therapist conducted a brief assessment of risk factors since last contact, including psychotic symptoms and any patient-specific factors (e.g., non-suicidal self-injury, substance use). If risk was identified, this was discussed first and the safety plan created in a prior session was reviewed. The therapist also facilitated active problem-solving and encouraged productive communication between the patient and SO to resolve any identified issues. The therapist discussed problems in terms of values-consistent strategies based on reviewing the previously developed Life Plan document. If a problem could not be solved in this manner, the therapist obtained permission to contact the patient’s primary treatment provider directly or temporarily increased the frequency of the contacts during times of crisis. The SO phone contacts proceeded in a similar manner, with a focus on supporting the patient’s treatment goals and managing risk. General support and encouragement were also provided to the SO.

Overview of adaptations for psychosis.

At the start of the project, we received feedback from two experts in schizophrenia, which we incorporated into the final treatment manual and training program. Patients with suicidality and co-occurring psychosis represent a unique clinical population requiring a tailored approach. Our originally developed CLASP protocol did not address psychosis (Miller et al., 2016). However, previous research on treatments for psychosis has demonstrated that psychosocial treatments, including ACT, can be successfully delivered to patients with acute psychosis (Gaudiano & Herbert, 2006). Thus, the CLASP-S protocol used in the current study was adapted from the original CLASP protocol to increase its feasibility, acceptability, and relevance for patients with psychosis, based on other interventions in this population (Bradshaw et al., 2005).

We carefully adapted CLASP-S to consider those with active psychosis, co-occurring problems (e.g., substance use, depression), lower levels of education, current cognitive deficits, medication side effects, significant positive/negative symptoms of psychosis, and functional impairment. The role of expressed emotion was assessed and addressed in the CLASP-S family meeting and follow-up contacts based on strategies previously found to be efficacious (Pitschel-Walz et al., 2001), including psychoeducation about the symptoms, course and treatment of psychosis, communication skills training, problem solving strategies, stress management techniques, and a discussion of relapse prevention between the patient and SO. In addition, the original CLASP protocol was modified (see Table 2) to support patients with psychosis by adding or increasing attention to the following strategies: a) rapport building and support across sessions, b) structuring of session material (by creating a clear agenda at session start and reviewing material at the end of each session), c) modeling and behavioral skill rehearsal by the therapist (including role playing of communication skills), d) cognitive tailoring of content based on individual need (facilitating the values discussion using simple stories and metaphors or focusing on the patient’s valued activities to make the discussion more concrete), e) use of handouts written at an appropriate reading level (including the Life Plan and other values clarification worksheets), f) attention to treatment adherence (by assessing this at each session and problem solving any identified barriers), g) reminders and self-monitoring strategies to support learning (via text message, emails, and letters), h) discussion of stigma and other illness-related treatment barriers (e.g., limited resources) (to address expressed emotion as incorporated into the initial session and family meeting), i) motivational enhancements to promote medication adherence (by linking values to adherence goals), j) linkage with community resources to promote out-of-session practice (by providing referrals to local support groups), k) ongoing assessment of psychotic and other symptoms that may interfere with treatment (at the beginning of each session), and l) personalizing of treatment goals based on individual needs and abilities (via the Life Plan document).

Table 2.

Protocol Adaptations for CLASP-S

Primary Factors for Adaptation: Consideration of those with active psychosis, co-occurring problems (e.g., substance use, depression), lower levels of education, current cognitive deficits, medication side effects, significant positive/negative symptoms of psychosis, and functional impairment
	CLASP	CLASP-S Additions
Structure	3 initial sessions (40 – 50 minutes) beginning during hospitalization, completed ≤ 2 weeks post discharge; 1 in-person meeting with both the patient and SO post discharge; 11 follow-up phone contacts (15–30 minutes) held separately with patient and SO for up to a 6-month period	Flexibility for sessions to be held in person or remotely via phone or secure videoconference; Option for sessions to be shorter in duration based on patient presentation and needs
Process	Rapport building, history taking, psychoeducation about suicide risk, suicide risk assessment, completion/review of safety plan, values clarification/goal setting	Rapport building and support across sessions; Emphasis on structure; Use of handouts written at an appropriate reading level; Motivational enhancements to promote medication adherence; Ongoing assessment of psychotic and other symptoms that may interfere with treatment; Modeling and behavioral skill rehearsal by the therapist; Linkage with community resources to promote out-of-session practice; Discussion of relapse prevention between the patient and SO
Content	The Life Plan (dedicated space to record the patient’s values and accompanying goals)	Psychoeducation about the symptoms, course, and treatment of psychosis; Stigma and other illness-related treatment barriers; The role of expressed emotion; Cognitive tailoring of content based on individual need (e.g., Communication skills training; Problem solving strategies; Stress management techniques); Attention to treatment adherence; Reminders and self-monitoring strategies to support learning; Personalizing of treatment goals based on individual needs and abilities

Note. SO = Significant Other.

Clinician training.

All CLASP-S sessions were led by master’s level clinicians, referred to as “Treatment Advisors.” These Advisors were trained in the CLASP-S protocol through a comprehensive process that included: 1) reviewing the treatment manual; 2) attending a recorded workshop on the theoretical basis, research evidence, and intervention methods used in the approach; 3) watching role plays of CLASP-S sessions; and 4) meeting with investigators to ask questions and receive feedback. Advisors were also trained in suicide safety planning procedures, suicide risk assessment, and crisis intervention. Advisors were then “certified” to deliver CLASP-S after successfully completing at least two mock role plays with study staff. These mock sessions were recorded and assessed by certified Advisors to ensure fidelity and competence with the protocol. Throughout the study, Advisors participated in weekly group supervision on their cases, provided by the licensed psychologists who were experienced in the intervention.

Enhanced Treatment as Usual (ETAU)

Patients in both conditions received unrestricted care, including typical inpatient treatment: medication management, occupational therapy, group therapy, and aftercare planning. Discharge plans included a scheduled appointment with a psychiatrist for outpatient pharmacotherapy and appointments or referrals for other appropriate community treatments. After discharge, community treatment was unrestricted. Patients in ETAU alone received enhanced safety and monitoring, as risk assessments were completed at each assessment time point, and safety plans were constructed for all participants at baseline. If the patient already had an existing safety plan completed as part of their hospital admission, this plan was reviewed with the patient and modified as needed. If one was not previously completed, a study interventionist completed a safety plan with the participant. After obtaining permission, assessment summaries were sent to the patient’s primary community treatment providers to support care coordination. In addition to ETAU, patients assigned to the CLASP-S condition received the intervention protocol described above.

Participants

Psychiatric inpatients were recruited from a hospital in the New England region of the United States. To be eligible for the study, participants had to have experienced suicidal ideation, plan, or a suicide attempt in the past week prior to hospital admission; have a diagnosis of schizophrenia, schizoaffective, schizophreniform disorder, schizophrenia unspecified, or mood disorder with psychotic features (bipolar disorder or major depressive disorder), as determined by the Structured Clinical Interview for DSM-5 (SCID-5; First et al., 2015); be 18+ years of age; and be able to speak and read English. Participants were excluded if they had a substance-induced psychotic disorder or if psychosis was due to a medical condition; if they did not have phone access (for follow-up); or if they had been discharged to a long-term residential facility. Patients’ SOs were also recruited when available. SOs (e.g., partner, parent, sibling, adult child, close friend) were broadly defined in the study. To participate, SOs had to be 18 years or older; be able to speak and read English; and be in regular contact with the subject.

Measures

Structured Clinical Interview for DSM-5 (SCID-5; First et al., 2015).

The SCID-5 is a semi-structured clinical interview assessing for major DSM-5 diagnoses. In the current study, it was used to confirm participant eligibility and psychiatric diagnoses. The SCID-5 has demonstrated strong reliability and validity, especially for patients with more severe psychiatric disorders (First et al., 2015).

Columbia Suicide Severity Rating Scale (C-SSRS; Posner et al., 2011).

The C-SSRS is a semi-structured clinical interview. The number of suicide attempts over one’s lifetime and lifetime suicidal ideation intensity and behaviors were measured at the baseline assessment, and past 3-month suicidal ideation intensity/past 3-month suicidal behaviors were assessed at follow-up using the C-SSRS. The suicidal ideation intensity subscale is a measure of ideation severity and is composed of 5 items (frequency, duration, controllability, deterrents, and reasons for ideation), each rated on a 5-point scale, with higher scores indicating higher levels of severity. The suicidal behavior composite included suicide attempts, interrupted attempts, aborted attempts, and preparatory behaviors (e.g., writing a suicide note; collecting pills). This suicidal behaviors composite variable was calculated for the 3-month follow-up; it reflected the number of overall suicidal behaviors. The C-SSRS has demonstrated validity and reliability (Posner et al., 2011), including for adults with serious mental illness (Madan et al., 2016).

Brief Psychiatric Rating Scale (BPRS; Overall & Gorham, 1962; Shafer, 2005).

Overall psychiatric symptom severity was assessed using the BPRS. The BPRS is an 18-item semi-structured clinical interview measure of psychopathology, including symptoms of psychosis (e.g., paranoia, hallucinations, delusions), anxiety, and depression over the past week. Each item is rated on a scale from 1 (not present) to 7 (extremely severe). A higher score on the BPRS suggests higher severity of psychopathology. The BPRS has established validity and reliability with psychiatric inpatients (Hofmann et al., 2022).

Alcohol Use Disorders Identification Test (AUDIT; Babor et al., 2001; Saunders et al., 1993).

Screening for alcohol use was completed using the AUDIT self-report measure. The AUDIT has 10 items that assess for alcohol intake, potential dependence on alcohol, and experience of alcohol-related harm over the past year. Scores range from 0–40, with higher scores indicating more hazardous alcohol use. It has been used in a variety of settings and patient groups, including patients with psychosis, showing good reliability and validity (Nesvag et al., 2010; Reinert & Allen, 2007).

Drug Use Disorders Identification Test (DUDIT; Berman et al., 2003; Berman et al., 2005).

Screening for drug use was completed using the DUDIT self-report measure. The DUDIT is 11 items and assesses the amount and frequency of substance use and signs of hazardous use or dependence, along with questions about which substances the participant used, including illicit drugs and misuse of prescription drugs. Scores range from 0–44, with higher scores indicating higher substance use and potential dependence. The DUDIT has shown good reliability and validity, including with individuals who have psychosis (Nesvag et al., 2010).

Brief Adherence Rating Scale (BARS; Byerly et al., 2008).

The BARS is a brief clinician-administered medication adherence assessment consisting of 4 items. It assesses for medication adherence over the past month. The item assessing the proportion of medication doses taken by the patient in the last month (0%−100%) was used for analyses. The BARS provides valid, reliable, sensitive, and specific estimates of antipsychotic medication adherence among patients with psychotic disorders (Byerly et al., 2008).

Procedures

Electronic medical records were screened for patients potentially meeting study criteria (after obtaining a HIPAA waiver for this purpose). Following permission from the attending physician, patients were asked to complete the pre-screening and formal consent process. Interested patients were consented and the baseline assessment was conducted to confirm eligibility. As the study was conducted in the midst of the COVID-19 pandemic, consent and baseline procedures were often conducted remotely. Following informed consent procedures and confirmation of eligibility, patients were randomized to either the enhanced treatment as usual (ETAU) alone condition or to the Coping Long Term with Active Suicide Program-Schizophrenia (CLASP-S) condition. Randomization was stratified by sex at birth (male/female) and baseline suicidality (attempt/active ideation) via a computer-generated list with concealment. Additionally, patients assigned to CLASP-S provided permission to contact the SO for study participation. Research staff contacted the SO, explained the study, and completed informed consent procedures.

Assessments were administered initially at baseline (shortly after admission) and at 3- and 6-months post-discharge (see assessment schedule below). Assessments were conducted in person (when possible) or remotely due to the COVID-19 pandemic. If patients were unable to attend in-person appointments, assessments were conducted over the phone and questionnaire packets were mailed to them. Assessments were conducted by an assessor trained to initial interrater reliability (>.80) on study measures. Study investigators reviewed diagnostic interviews to confirm accuracy. Participants were followed up to 6 months regardless of their continued participation in CLASP-S sessions or subsequent suicide attempts or hospitalizations. Patients were compensated $40 for completing each of the 3 study assessments but were not paid for attending treatment sessions.

Data Analysis Plan

Analyses were conducted using IBM SPSS Statistics version 26. CLASP-S and ETAU alone groups were compared on baseline characteristics. Independent-samples t-tests were used to compare means across conditions, and chi-square tests were used to compare frequencies. Due to the small sample size, in cases with expected cell counts < 5, Fisher’s Exact Test was employed. Alpha was set at < .05 without adjustments due to the exploratory nature of the study and reduced sample size. The primary aim of the study was to assess feasibility/acceptability for future research by evaluating the following: our ability to identify and consent patients meeting our study criteria, retention of patients through post-treatment and follow-up, completion of study assessments and treatment sessions, the presence of any study-related adverse events, improvements in our preliminary clinical outcomes related to suicidality and symptoms, and informal feedback gleaned from patients about the intervention. As part of assessing feasibility/acceptability, we examined clinical outcomes to identify possible intervention effects to inform future clinical trials. The smaller sample size prevented meaningful comparisons between groups; therefore, we present within-group changes over time from baseline to follow-up based on t-tests. Missing data were not imputed given the small sample size.

Results

Recruitment/Retention Feasibility

Of the 64 patients approached, 37 consented to participate in the study, which is an approach/consent rate of 57.8%. A total of 7 out of 37 (18.9%) consented patients were determined to not meet study criteria following baseline assessment. Thus, 30 patients were enrolled and randomized to conditions: ETAU alone = 13 and CLASP-S = 17. In the CLASP-S group, a total of 4 patients (23.5%) withdrew from the study, stating lack of time as the reason, and 0 (0%) withdrew from ETAU alone (χ²₍₁₎ = 3.5, Fisher’s Exact Test p = 0.113). Of those continuing in the study (n = 13 for ETAU alone, n = 13 for CLASP-S), retention rates at 3-month follow-up were 61.5% for CLASP-S (n = 8 completed, n = 5 lost to follow-up) and 69.2% for ETAU alone (n = 9 completed, n = 4 lost to follow-up; χ²₍₁₎ = 0.17, Fisher’s Exact Test p = 1.00). For those continuing in the study through 6-month follow-up, 8 participants (61.5%) completed assessments at 6 months (n = 5 lost to follow-up) in the ETAU condition, while 4 participants (30.8%) completed 6-month follow-up (n = 9 lost to follow-up) in the CLASP-S condition (χ²₍₁₎ = 2.48, Fisher’s Exact Test p = 0.238).

Baseline Characteristics

Regarding full sample demographics (n = 30), 40% of participants indicated female birth sex, 33% identified as a racial minority, and 27% were Hispanic/Latino. The average age of the sample was 36.1 years, and the average level of education completed was 12.8 years. No significant differences were found between the conditions on baseline variables (See Table 1 for details).

Table 1.

Sample demographic and clinical characteristics

Variable	Full Sample (n = 30)	CLASP-S (n = 17)	ETAU (n = 13)	χ2	t
Age M(SD)	36.1 (13.1)	37.5 (13.3)	34.2 (13.0)		−.67
Years of school M(SD)&	12.8 (2.1)	12.9 (2.6)	12.8 (1.5)		−.18
Birth Sex n(%)				.36
Male	18 (60.0)	11 (64.1)	7 (53.8)
Female	12 (40.0)	6 (35.3)	6 (46.2)
Gender Identity n(%)$ (Yes/No)
Male	16 (53.3)	9 (52.9)	7 (53.8)	.00
Female	10 (33.3)	6 (35.3)	4 (30.8)	.07
Transmale	3 (10.0)	1 (5.9)	2 (15.4)	.74
Non-Binary	1 (3.3)	0 (0.0)	1 (7.7)	1.35
Race/Ethnicity n(%)$ (Yes/No)
American Indian/Alaska Native	1 (3.3)	1 (05.9)	0 (0.0)	.79
Black or African American	4 (13.3)	2 (11.8)	2 (15.4)	.08
White	21 (70.0)	10 (58.8)	11 (84.6)	2.33
Multiple Races	5 (16.7)	4 (23.5)	1 (7.7)	1.33
Hispanic	8 (26.7)	6 (37.5)	2 (15.4)	1.76
Marital Status n(%)$				2.87
Married	1 (3.6)	0 (0.0)	1 (7.7)
Single, never married	20 (71.4)	10 (66.7)	10 (76.9)
Separated/Divorced/Widowed	7 (25.0)	5 (33.3)	2 (15.4)
Employment Status n(%)$				5.98
Employed full or part time	6 (21.4)	3 (20.0)	3 (23.1)
In school/training	1 (3.6)	0 (0.0)	1 (7.7)
Unemployed	8 (28.6)	6 (40.0)	2 (15.4)
Retired	2 (7.1)	1 (6.7)	1 (7.7)
Disabled	11 (39.3)	5 (33.3)	6 (46.2)
Household Income n(%)$				8.13
< $20,000	18 (69.2)	10 (71.4)	8 (66.7)
$20,000 – $59,000	4 (15.4)	3 (21.4)	0 (0.0)
≥ $60,000	4 (15.4)	0 (0.0)	4 (33.3)
Clinical Characteristics M(SD)
BPRS Total	56.6 (12.2)	54.7 (14.0)	58.8 (9.9)		.89
AUDIT	4.8 (9.4)	6.8 (12.1)	2.7 (4.8)		−1.1
DUDIT	10.3 (15.3)	13.1 (17.4)	7.5 (13.1)		−.89
Suicide History
Number lifetime SA M(SD)	13.3 (29.8)	8.6 (23.6)	19.5 (36.5)		.99
Past SA n(%)	21 (70.0)	12 (70.6)	9 (69.2)	.00

Note.

^&Denotes a variable where data was missing (n is < 30). Percentages for these variables are out of the variable total, rather than the sample total; M(SD) indicates variables for which means and standard deviations are depicted; n(%) indicates number and percentage.

^$Indicates cases where participants could check all that apply. BPRS = Brief Psychiatric Rating Scale; AUDIT/DUDIT = Alcohol/Drug Use Disorders Identification Test; SA = suicide attempt; SI = suicidal ideation.

For baseline clinical characteristics, the full sample endorsed high levels of baseline psychiatric symptoms (BPRS Total M = 56.6, SD = 12.2, a score of ≥ 53 indicates severe symptomatology; Leucht et al., 2005) and 70% reported a lifetime history of a suicide attempt. Similarly, the sample mean for the measure assessing drug use (DUDIT: M = 10.3, SD = 15.3) was above the threshold indicative of drug-related problems (M ≥ 6 for men, M ≥ 2 for women; Berman et al., 2003). There were no statistically significant differences in clinical characteristics at baseline between CLASP-S and ETAU alone (see Table 1).

Non-Study Medication Management and Adherence

In addition to intervention and assessment provided as part of the study, participants received routine hospital treatment and referrals for community follow-up post-discharge. Through self-report and record review, we identified that among participants with medication information available at baseline (n = 26), 50% were taking psychotropic medication prior to their hospitalization. At discharge, among patients with medication information available (n = 27), 96% were prescribed a psychotropic medication when leaving the hospital. The average self-reported adherence rating in the previous month at baseline was 56.0% of prescribed doses taken and at 3-month follow-up was 86% of prescribed doses taken.

Engagement with CLASP-S

The median number of CLASP-S sessions completed was 3 sessions, and the mean was 3.5 sessions (SD = 3.7; range = 12). Of the 17 participants in the treatment condition, a total of 70.6% (n = 12) of patients received at least 1 CLASP-S session, and 29.4% (n = 5) of patients did not complete any sessions with their CLASP-S therapist. Specifically, session completion rates were as follows: 2 participants (11.8%) engaged in only 1 session, 3 participants (17.6%) completed 3 sessions, 3 participants (17.6%) completed 4 sessions, and 4 participants (23.5%) completed 5 – 12 total sessions. Among the patients who engaged with CLASP-S (n = 12), a total of 4 SOs (33.3%) consented to participate with patients and completed 1 to 3 CLASP-S sessions.

Serious Adverse Events: Rehospitalizations and Suicide Attempts

No serious adverse events (SAEs; rehospitalization, suicide attempt, incarceration) were determined to be related to study participation, and no deaths were identified during the study. Given the nature of the sample, psychiatric rehospitalizations and suicide attempts were considered expected events. In CLASP-S (n = 17), 29.4% (n = 5) of participants were re-hospitalized and 5.9% (n = 1) experienced a suicide attempt over follow-up. For ETAU alone (n = 13), 46.2% (n = 6) of participants were re-hospitalized and 15.4% (n = 2) experienced a suicide attempt over follow-up. There were no significant differences between groups regarding the percentage of participants with an inpatient hospitalization during follow-up (χ²₍₁₎ = 0.89, Fisher’s Exact Test p = 0.45) or a suicide attempt (χ²₍₁₎ = 0.74, Fisher’s Exact Test p = 0.57). One participant from the ETAU alone group (7.7%) was incarcerated following their baseline assessment (an unexpected but study unrelated serious adverse event) and thus administratively removed from the study based on IRB procedures.

Preliminary Treatment Outcomes

Given the lower number of participants with 6-month data, we were only able to meaningfully examine outcomes from baseline to 3-month follow-up. First, we graphically depicted individual-level scores for the CLASP-S group, including C-SSRS rated past 3-month highest suicidal ideation “intensity” (a subscale of the C-SSRS measuring ideation severity) and past 3-month suicide behaviors composite, to illustrate changes over time (see Figure 1). Visual examination of the graphs indicates that most participants showed improvements from baseline to 3-month follow-up on the BPRS and C-SSRS, with only one person showing worsening on the BPRS.

Figure 1. CLASP-S individual level change from baseline to 3-month follow-up.

Note. BPRS = Brief Psychiatric Rating Scale; Suicidal ideation intensity measured using the Columbia Suicide Severity Rating Scale.

Regarding improvements in clinical outcomes, we analyzed within-group changes via t-tests because the small sample size precluded meaningful comparisons between groups (See Table 3). In terms of suicidality, the CLASP-S group demonstrated a significant decrease in C-SSRS rated past 3-month suicidal ideation intensity (n = 8; t₍₇₎ = 2.78, p < .014, Cohen’s d = 1.00), as well as past 3-month suicide behaviors composite scores (n = 8; t₍₇₎ = 7.48, p < .001, Cohen’s d = 2.65). Additionally, there was a significant decrease in the BPRS total score over time (n = 8; t₍₇₎ = 3.15, p = .008, Cohen’s d = 1.11). The ETAU alone group also demonstrated a significant decrease in past 3-month suicidal ideation intensity (n = 8; t₍₇₎ = 5.28, p < .001, Cohen’s d = 1.87), suicide behaviors composite score (n = 8; t₍₇₎ = 3.21, p = .007, Cohen’s d = 1.14), and BPRS total score (n = 8; t₍₇₎ = 6.26, p < .001, Cohen’s d = 2.21).

Table 3.

Within-Group Changes from Baseline and Follow-up

	CLASP-S					ETAU
Variable	Time	n	M (SD)	t	d	n		t	d
Symptoms BPRS
Total	BL	8	55.4 (14.4)			8	61.3 (10.0)
	3M	8	38.8 (13.5)	3.2**	1.1	8	35.1 (11.8)	6.3***	2.2
Suicidality CSSRS
Past 3-month SI intensity	BL	8	16.3 (2.2)			8	18.1 (3.7)
	3M	8	8.9 (5.8)	2.8*	1.0	8	3.8 (7.0)	5.3***	1.8
Past 3-month suicide behaviors	BL	8	2.1 (1.0)			8	1.3 (0.9)
	3M	8	0.1 (0.4)	7.5***	2.7	8	0.1 (0.4)	3.2**	1.1
Serious Adverse Events			N (%)				N (%)	χ2	p
Hospitalization	6M	17	5 (29.4%)			13	6 (46.2%)	.89	.45
Suicide Attempt	6M	17	1 (5.9%)			13	2 (15.4%)	.74	.57
Incarceration	6M	17	0 (0.0%)			13	1 (7.7%)	1.35	.43

Note. BL = Baseline assessment; 3M = 3-month follow-up assessment; BPRS = Brief Psychiatric Rating Scale; CSSRS = Columbia Suicide Severity Rating Scale; Suicide behaviors composite scores consist of suicide attempt, aborted attempt, interrupted attempt, and preparatory behaviors. Serious Adverse Events include chart review and are therefore inclusive of the full intent-to-treat sample.

^*p < .05

^**p < .01,

p < .001.

Chi-square p is Fisher’s Exact Test.

Discussion

To date, there has been very little research on evidence-based suicide prevention programs for individuals with schizophrenia and related conditions (Villa et al., 2020). Even less is known about how to prevent suicides in this population following a recent psychiatric hospitalization when risk is at its highest (Olfson et al., 2016). Results from this pilot RCT demonstrated that the CLASP-S intervention was safe, feasible, and potentially effective for decreasing suicide risk and psychiatric symptoms following an acute psychiatric hospitalization. We showed that recruiting patients into the program was feasible during their hospital stay. We also showed acceptability by successfully engaging most patients in the program in the initial weeks immediately following their discharge, fostering a safe and more successful transition to the community. The first few weeks following a hospital discharge are a period of 200–300 times greater risk of suicide (Chung et al., 2017), representing a critical time for increased contact.

Overall, the approach/consent rate of 57.8% demonstrated good acceptability by patients and feasibility of recruitment for future studies, and the demographics characteristics suggested a largely representative sample. Furthermore, patients in CLASP-S showed significant improvements in the large effect size range for changes in suicidal ideation and behaviors and total psychiatric symptoms by 3-month follow-up. Patients in the ETAU alone condition also evidenced reductions in suicidality and BPRS scores over time, but the sample size was too small to compare group effects, and confidence intervals around effects were too wide to conclude that one treatment was superior to the other. Although not statistically significant, it is important to note that the rates of both suicide attempts and rehospitalizations were numerically lower in the CLASP-S condition, and the effect size for reductions in suicide behaviors was greater compared to ETAU alone, suggesting the utility of conducting larger, fully-powered studies in the future. Additionally, informal feedback from patients to their treatment advisors suggested that they found the program helpful in supporting their hospital aftercare because someone was following up with them and assessing their recovery on a regular basis to help prevent crises. However, we noted that some patients reported difficulty engaging with the remote session format due to concentration. Conducting the pilot RCT provided important lessons that will further enhance the intervention for testing in future, larger-scale trials with this population, which we describe below.

Comparisons with previous studies

Until recently, most previous suicide research has excluded patients with SSDs from clinical trials (Villa et al., 2020); thus, the current project was able to support the feasibility of recruiting and providing a suicide prevention program to patients with SSDs following their hospitalization. There has been an increase in recognition of the need to include complex patient populations in suicide research and adapt interventions specifically for psychosis. Importantly, recent reviews have highlighted insufficient evidence for the effectiveness of current suicide prevention strategies for patients with schizophrenia (Penadés et al., 2022). One recent study of the efficacy of CLASP showed that the intervention did not demonstrate significant effects on suicide behaviors in the subgroup with SSD compared to those with other psychiatric diagnoses (Rabasco et al., 2024), but the number of patients with SSD was underpowered for examining this subgroup. Current findings highlight the importance of adapting our evidence-based treatments for suicidality to those with psychosis.

There are several promising treatments studied for reducing suicidal thoughts and behaviors in individuals with SSD, particularly cognitive and behavioral interventions (Bornheimer et al., 2020). For example, Cognitive-Behavioral Suicide Prevention for psychosis (CBSPp) is one treatment that has been tailored to people with SSD. Although some studies have shown positive effects for CBSPp (Bornheimer et al., 2023; Tarrier et al., 2014), a recent a multisite clinical trial in the United Kingdom found no direct effects of the treatment on reducing suicidal thoughts and behaviors (Gooding et al., 2025). This research group identified an indirect effect for improving appraisals of social support, a concept also emphasized in the CLASP-S intervention. Future research on CPSPp is needed. Additionally, the use of technology (mobile applications) to enhance social connection and treatment engagement is a promising avenue for suicide prevention in patients with SSDs, and is being examined by our group (Gaudiano et al., 2025) and others (Depp et al., 2023). Therefore, additional attention to strategies for promoting social support and leveraging mobile technology to encourage patient engagement may prove beneficial in future applications of the CLASP-S protocol.

Furthermore, combined treatment with psychotherapy and pharmacotherapy is a well-supported approach for SSDs, as psychosocial intervention can increase coping strategies and improve functioning while antipsychotic medications may target specific symptoms of psychosis (e.g., hallucinations; Modesti et al., 2023). More specifically, clozapine is an antipsychotic medication that has been approved for reducing suicidal behaviors among patients with SSDs, and a recent review has highlighted its effectiveness in reducing suicidality when compared to other antipsychotic medications (Masdrakis & Baldwin, 2023). However, this medication’s potentially dangerous side effect profile is a known limitation, thus making it inappropriate for some patients (Iqbal et al., 2003). Future research could explore the combination of CLASP-S and clozapine for reducing suicide in patients with SSDs for which this treatment is indicated.

Qualitative interviews and community-based participatory research also suggest that addressing suicidality among patients with psychosis is important (Bornheimer et al., 2022). The clarity of treatment goals, a strong therapeutic alliance, and flexibility in intervention delivery are key components that have been identified to increase engagement and acceptability (Harris et al., 2023). Below, we outline lessons learned from this pilot RCT and ideas for future adaptations and implementation of CLASP-S.

Lessons learned

First, CLASP-S is designed as an adjunctive intervention. Thus, increasing communication between the CLASP-S therapist and patients’ community treatment providers may help to improve patient follow-up and intervention engagement, especially over the longer term. Better integrating CLASP-S with patients’ usual care could provide more opportunities for providers to synergistically support each other.

Second, patients with SSD experience disproportionate concerns related to issues with social determinants of health when compared to patients with other psychiatric conditions (Jester et al., 2023). Global prevalence rates of psychosis among homeless populations are estimated to be between 19% to 29% (Ayano et al., 2019). Therefore, the CLASP-S intervention may benefit from additional attention to assessing and addressing various social factors (e.g., economic stability, education quality, healthcare access, neighborhood safety, and community support) and case management needs, such as that found in the evidence-based Critical Time Intervention (Herman et al., 2011).

Third, patients with SSD may need further adjustments to the treatment protocol to increase the acceptability and feasibility of completion. Specifically, the number and length of CLASP-S sessions may have been challenging for patients to adhere to, which is consistent with previous research demonstrating poorer therapy engagement in those with psychosis (Gooding et al., 2025). Briefer but more frequent sessions in the first month of CLASP-S may be a better approach, especially given our findings of more successful engagement during the first month following hospital discharge when risk is highest, and the subsequent decline in engagement we found after three months. Typical SSD symptoms also encourage this pivot, as shorter sessions may mitigate the impact of patients with difficulty concentrating, and more frequent contact during the initial weeks post-discharge may help to increase rapport building.

Fourth, we found that many participants preferred to complete interviews and sessions in person, and our interventionists noted some difficulties with engaging participants over the phone due to cognitive deficits. The COVID-19 pandemic necessitated the increased use of remote interactions during the study, which may have decreased the potency of our intervention. While phone contacts and telehealth options can increase accessibility, this population may benefit more from in-person sessions at the start of the intervention to build a stronger therapeutic alliance, increase social support, and foster commitment to longer-term engagement.

Limitations

This study was impacted by some limitations. First, our ability to recruit the planned sample size and to follow patients over 6 months was negatively impacted by the COVID-19 pandemic. Staffing shortages, unit closures, and restrictions for in-person follow-up visits significantly disrupted the study protocol. Nevertheless, we adapted our procedures during the pandemic (e.g., to include telehealth appointments and remote sessions), and we were able to achieve most of the original goals of this pilot project. The relatively small sample and extent of missing follow-up data also limited our ability to conduct meaningful between-group comparisons of follow-up data. Despite this constraint, we believe the within-group longitudinal comparisons provide meaningful data to inform future research and larger, next-step trials. Results suggest that the primary target of suicidal thoughts and behaviors meaningfully improved over the first three months post-discharge in those receiving CLASP-S. We planned to assess satisfaction in the CLASP-S condition; however, due to a mistake in our assessment battery, we did not collect this measure during earlier sessions as intended. Additionally, there were too few participants available at the 6-month follow-up to complete the satisfaction assessment measures. Future investigations should more formally assess treatment satisfaction using formal qualitative interviews to identify relevant themes. Conducting formal qualitative interviews and mixed-methods research could also provide additional insights into how to adapt CLASP-S to make it more feasible and acceptable for patients. While we consulted with schizophrenia experts when designing the protocol, the lack of community perspectives to inform our adaptations was a limitation. Future work should seek to include feedback from persons with lived experience during all phases of intervention refinement and implementation.

Future directions

Some patients reported difficulty engaging with the primarily remote/telehealth format we had to employ due to the COVID-19 pandemic. In the future, we plan to implement more frequent, but shorter, in-person sessions to promote patient engagement with the intervention. Second, while we originally planned to deliver up to six months of treatment, some participants reported feeling as though this timeframe was too long. Therefore, we plan to make CLASP-S more targeted by focusing on the first one to three months post-discharge, when risk is at its highest and patients may benefit from the additional assistance. Further, modifications to the CLASP-S protocol would be helpful in addressing cognitive limitations and active psychotic symptoms as part of safety planning. In the future, we plan to refine the protocol further and conduct a subsequent full-scale RCT of CLASP-S with the aforementioned modifications.