Dial 988 for a suicidal crisis: Preliminary research priorities for factors influencing individual consumer outcomes

Abstract

Importance:

Since July 2022, calling or texting “988” in the US connects callers to the National Suicide Prevention Lifeline following a law passed by Congress to simplify access to the mental health crisis line in the US. Relative to other areas of suicide research, knowledge regarding how and to what extent crisis lines prevent suicide crises and suicide deaths remains in its infancy. We briefly review the state of this research and suggest critical directions for future research on factors that may influence effectiveness.

Observations:

988 stands to improve access to critical life-saving measures in the moments of a suicidal crisis. However, urgent questions remain regarding how to improve effectiveness of crisis lines. Available evidence suggests that crisis lines are often effective at reducing immediate distress and reducing suicide risk, but substantial gaps remain in our understanding of how crisis lines work.

Conclusions and Relevance:

We recommend that future research with suicide prevention crisis lines, such as 988, identify and test factors influencing effectiveness, including conversation, consumer, dyadic, and structural level characteristics. Existing research, while minimal, suggests that prescription of 988 to prevent suicide death is clinically warranted, but much more work is needed to optimize care.

As of July 2022, calling or texting “988” in the United States (US) connects consumers to the National Suicide Prevention Lifeline following a law passed by Congress. This 24-hour crisis service, funded by the Substance Abuse and Mental Health Services Administration (SAMHSA), can improve suicide prevention by providing rapid access to “counselors”¹ trained to manage suicidal crises. Despite being one of the largest public mental health services in the US, knowledge about factors that influence crisis line effectiveness is limited. Yet, use of the Lifeline is routinely prescribed by psychiatrists, psychologists, and other mental health professionals to keep patients safe in the moments of a crisis. The purpose of this manuscript is to outline high priority areas for research on factors that influence individual crisis conversation outcomes and effectiveness.

Although our focus is on factors affecting individual crisis conversations, we acknowledge that public health and implementation science experts may emphasize broader structural-level factors that influence outcomes as research priorities. While we briefly mention some structural factors closely associated with individual crisis conversations, we acknowledge that we will not be exhaustive delineating the vast amount of research on structural factors needed to continue improving the evidence base supporting 988 as a suicide prevention tool. Instead, we encourage readers to view this manuscript as outlining preliminary research steps.

Recently, the National Institute of Mental Health’s (NIMH) issued a Notice of Special Intertest (NOT-MH-22-110)¹, which calls for research aimed at understanding individual outcomes in crisis research. Specifically, NIMH seeks research aimed at evaluating whether crisis services, including 988, provide short-term stabilization for individuals in crisis. They call for research that is “conducted in real-world settings, where a wide range of clinical presentations, psychosocial factors, age-related (e.g., youth, adult, older adult), geographic (rural/remote settings), cultural considerations, and health disparities influence the types of care that are provided.” Consistent with this NOSI, we outline a research agenda on factors affecting individual outcomes that span conversation, consumer, dyadic, and structural characteristics.

Below, the term outcome(s) refers to accurate assessment of suicide risk level, prevention of a suicide attempt or suicide death via intervention, connection with follow-up care, and/or consumer perceptions of care. We use the term effectiveness to reflect whether the crisis conversation resulted in accomplishing a positive outcome (e.g., prevented a suicide attempt). The term factor reflects specific aspects of the crisis conversation experience that may influence outcomes and effectiveness.

The Unique Aspects of Suicidal Crises

Unlike other forms of mental illness that can last for weeks, months, or years (e.g., depressive episodes), acute suicidal crises frequently are transient, lasting from minutes to hours, and may be longer in duration among those with multiple previous suicide attempts.² Further, we know that an active suicidal crisis may be the most impactful moment for preventing suicide death. In a study of 153 survivors of nearly lethal suicide attempts, 24% of survivors ages 13-24 reported that less than 5 minutes passed between the time they decided to attempt suicide and when they actually attempted suicide, with another 24% reporting 5 to 19 minutes, 23% reporting 20 minutes to 1 hour, 16% reporting 2-8 hours, and 13% reporting 1 or more days.³ Other research with adults recently hospitalized for suicide attempts similarly suggests that a 10-minute time window is critical for intervening to prevent a suicidal crisis from becoming a suicide attempt.⁴ Crisis lines, such as 988, may be a critically important resource to prevent suicide attempts and deaths.

Existing Literature Supporting Suicide Crisis Lines

Research to date shows that individuals of all ages use crisis lines and many are experiencing a suicidal crisis when they contact the line^5–7. One study showed that 8% of suicidal callers were actively attempting suicide or harming themselves while using crisis lines.²⁶ Available data suggest that crisis lines appear to reduce immediate distress and imminent suicide risk^{26,27,28–31}. More broadly, crisis line contacts have been found to decrease depression,^9,10 emotional distress,^9,11,14,15 hopelessness,^26,27 and severity of the presenting problem¹⁶ and to increase hopefulness^11,17 and resourcefulness.¹⁷ Further, many users report that they would use the crisis line service again and recommend it to others.¹⁶

Evidence to date with crisis lines suggests that they are potentially effective at managing immediate crises, yet many crucial questions about crisis line effectiveness remain. Relative to other areas of suicide prevention research, research with crisis lines has remained limited. Part of this imbalance may be driven by misconceptions that systematic research with crisis lines is not possible due to the anonymity of consumers. Yet, following up with consumers is not only feasible, but potentially beneficial. An initial study found that 91% of previously suicidal individuals reported that a follow-up check-in kept them safe, and 80% reported that this check-in stopped them from attempting suicide.¹⁸ Additional evidence supports that individuals who receive check-in calls are 3.33 times less likely to die by suicide compared with those who did not receive check-in calls.¹⁹ Building on this initial work can lead to improved effectiveness of crisis lines.

Initial Factors Affecting Individual Crisis Conversation Outcomes

Currently, there is limited knowledge about what influences individual level crisis conversation outcomes. Here, we describe four initial factors for future research, including (1) conversation, (2) consumer, (3) dyadic, and (4) structural factors associated with individual crisis conversations. In Table 1, we describe some example unanswered questions, available evidence, and example future research directions for each factor. This should not be viewed as a comprehensive list, but rather, initial examples within each domain.

Table 1.

Possible factors affecting individual outcomes and effectiveness for crisis lines

Possible Mechanism	Key Question(s)	Existing Evidence	Future Research Need
Counselor Characteristics
Qualifications/Certifications	Do counselor qualifications relate to crisis line effectiveness? As crisis line workers are not licensed mental health professionals, what is the minimal level of training needed to be effective during a crisis line contact?	No research available.	Examination of how counselor qualifications and certifications relate to crisis line effectiveness (e.g., better user outcomes) using observational study designs.
Behaviors	Which counselor behaviors are related to better vs. poorer crisis line user outcomes?	Gould et al., 2007; Mishara et al., 2007 Counselor empathy, respect, support, and collaborative problem-solving are associated with better user outcomes, including reduced suicidal ideation. Counselor behaviors perceived as condescending, abrupt, or unconcerned are associated with poorer user experiences.	Observational and RCT designs that examine how crisis line training increases positive counselor behaviors and reduces negative counselor behaviors. Expanded observational research studies that further examine crisis counselor behaviors across a wider range of counselors.
Consumer Characteristics
Demographics	How do demographic factors such as age, biological sex, gender identity, and race/ethnicity, influence what happens in a crisis conversation? What are disparities in accessing crisis lines in the United States?	Dichter et al., 2022; Gould et al., 2021; Hannemann et al., 2021; Hunt et al., 2018 Callers in sample of Lifeline 25% identified as male, 65.1% as female, 2.9% as transgender, 2.1% as questioning, 1.3% genderqueer, and 1.2% other. Of these, 39% were under the age of 17, 32.4% were between 18-24, 17% were 25-34, and 10.7% were over 35. Callers to the Veterans Crisis Line were not able to be identified as a gender, and of those who identified their gender, 15.1% were female. Female callers of Veterans Crisis Line were less likely to die by suicide 12 months after calling than males, but were more likely to exhibit non-fatal suicidal behaviors.	Basic observational studies that gather systematic demographic data that inform who contacts crisis lines, how representative are the people who contact them, and how many individuals are in a suicidal crisis. Observational studies that specifically investigate disparities in who accesses crisis lines compared to rates of emergency mental health visits in a geographical region.
Severity of suicide risk	Are those who contact crisis lines more likely to have more severe risk for suicide?	Gould et al., 2021; Mishara et al., 2007 Most callers in a Lifeline sample reported current (within 24) suicidal thoughts (54.8%) and 28.3% reported recent (within the past few days) suicidal thoughts. Among those who did not feel better after calling a crisis line, a larger proportion were suicidal relative to nonsuicidal.	Conduct empirical studies using standardized and brief empirically validate suicide assessment instruments and conduct short term follow-ups with ecological momentary assessment methodology to track associations between self-injurious thoughts and behavior severity and crisis line effectiveness.
Past STBs	Does the effectiveness of crisis conversations differ based on history of STBs?	Ramsey et al., 2019 Prior suicide attempts have been linked with increased risk of suicide death among individuals who contacted a crisis line in Northern Ireland.	Observational studies to systematically track outcomes of both the short term crisis conservations and future risk for consumers with past STBs.
Dyadic Characteristics
Therapeutic alliance	Does the quality of therapeutic alliance affect the effectiveness of the call?	Ben-Ari & Azaiza, 2003; Coveney et al., 2012; King, 1977; Ramchand et al., 2017 Consumers rate moderate to high satisfaction, effectiveness, helpfulness, and supportiveness during a crisis contact.	Conduct studies with consumer, counselor, and independently rated alliance measures with follow-up assessments of consumers suicide risk.
Cadence of text conversations	Does the speed of response between user and counselor affect the user’s perception of care and/or effectiveness of the conversation?	No research available.	Examine the time between messages exchanged between user and counselor as a predictor of crisis resolution.
Cultural responsiveness	Are conversations more effective if counselors are culturally responsive to individuals from minoritized backgrounds? Do counselors consider the cultural context for the interventions they suggest? Are there alternative options available to rescue users who are historically mistreated by law enforcement or health care systems?	No research available.	Use natural language processing algorithms to compare conversations between counselors trained to be culturally responsive compared to those without such training.
Conversation Content
Evidence based assessment	Does implementing standardized suicide risk assessments during crisis conversations reduce current or future suicidality? What is the best way to implement standardized comprehensive suicide risk assessment with fidelity in high-stress crisis conversations? Are there components of suicide risk assessments that are more likely to contribute to positive outcomes?	Gould et al., 2013 ASIST is associated with improvements in some areas of suicide risk assessment practices and positive outcomes. However, ASIST training was not associated with improvements in all areas of suicide risk assessment (e.g,. assessment of suicide plans). Some specific ASIST strategies might be more important for reducing suicide risk.	Conduct RCTs to identify suicide risk assessment practices in crisis contexts that are most effective at reducing suicidality in the moment and future suicidality.
Evidence based interventions	Do approaches informed by evidence-based interventions for suicidality, such as safety planning, DBT strategies, etc. used in the context of crisis services reduce suicidality? How might we adapt or implement evidence-based approaches in crisis contexts? What intervention strategies are most effective and under what conditions?	Labouliere et al., 2020 Evidence supports that training counselors on Safety Planning Intervention is feasible. No other research on the effect of evidence-based interventions in the context of crisis services on suicide related outcomes is available.	Test which components of evidence-based interventions, such as DBT, may be feasibly adapted for crisis contexts. Conduct RCTs to examine effect of evidence-based interventions, such as SPI, and DBT, on current and future suicidality. Investigate which components of evidence-based interventions are most effective depending on factors such as level of suicide risk, level of distress, or type of precipitating event, etc.
Fidelity	How adherent are counselors to the crisis line intervention models and protocols?	Cross et al., 2014, 2017 Examined cost-effective train-the-trainer models of ASIST with Lifeline workers; indicated wide variability in fidelity (adherence to the intervention content and competence in its delivery); trainer competence (over adherence) was more strongly related to better counselor performance (e.g., risk assessments and safety planning)	Examination of counselor fidelity to crisis line protocols -- both directly following training and over time (drift). Examination of how counselor fidelity relates to more effective outcomes.
Connection to services	To which other services, such as emergency response teams (e.g., paramedics police), mobile crisis units, inpatient, or outpatient treatment, should crisis services connect consumers to reduce current and future suicidality, and under what conditions? What are the barriers to connection to other services? How might crisis services improve continuity of care?	Britton et al., 2013; Gould et al., 2012 59% of all calls through Veterans Crisis Lines ended in a referral. 50% of callers on a national suicide crisis line used mental health referrals provided to them after the call. There are some barriers to callers using referrals that should be addressed. No research available on impact of “active rescues”, or connection to other services, on suicide related outcomes.	Follow-up with consumers should be conducted after the crisis conversation to investigate if connection to services was achieved, and the impact of connection to various services on suicide related outcomes depending on factors such as level of suicide risk, current access to care, treatment history, etc. RCTs could be connected to test methods for improving connection to services (e.g., use of motivational interviewing during conversations, increased coordination and communication with emergency response teams, hospitals, etc.).

Notes. ASIST = Applied Suicide Intervention Training (ASIST); RCT = Randomized Control Trial; DBT = Dialectical Behavior Therapy

Conversation characteristics

Individual crisis conversations are led by “counselors” who may assess risk, provide interventions, and/or decide case dispositions. Successful conversations will reduce distress and risk for suicidal behavior or death by collaborating with an individual to address access to lethal suicide means, enacting a safety plan, or involving a third party.²⁰ Thus, conversation characteristics, including counselor behaviors and content of the conversation, are a high priority for future research (Table 1).

Findings show that counselor expressions of empathy, respect, support, and collaborative problem-solving are associated with improvements to users by the end of the conversation, including reduced suicidal ideation.¹⁷ Notably, these skills may be inherent to the counselor rather than taught.¹⁷ On the other hand, counselor behaviors perceived as condescending, abrupt, or unconcerned are associated with more negative experiences.⁸ It is likely that there is a wide range in counselor behaviors across the crisis line system, and future work must capture the degree to which this affects individual outcomes.

Systematic assessment of suicide risk may serve as an important factor driving effectiveness. Given research indicating that suicide risk assessments were not occurring systematically on crisis lines,¹⁷ efforts have been made to standardize assessment procedures among crisis line workers.²¹ For example, SAMHSA along with Lifeline, successfully implemented the standardized Applied Suicide Intervention Training (ASIST) across 17 Lifeline centers.²² Although ASIST can be effectively implemented in this crisis line setting with positive improvement in some practices (e.g., exploring reasons for living) and caller outcomes, ASIST did not result in more comprehensive suicide risk assessment practices.²² Nevertheless, an evaluation of ASIST showed that individuals who interacted with an ASIST-trained counselor felt less depressed, less suicidal, less overwhelmed, and more hopeful by the end of the conversation.²² Specific ASIST strategies, such as exploring reasons for living and support contacts, were identified as key predictors of positive outcomes.²³ Continued examination of standardized suicide risk assessment by crisis line counselors is warranted to optimize care, in part because this assessment will necessarily dictate which intervention is selected.

Little is known about the individual interventions that are delivered during a crisis conversation. This is concerning given that these strategies may not be evidence-based interventions (EBIs). Because crisis counselors are not required to be licensed and likely have a wide range of experience, it is unclear the extent to which they will have the skills and training necessary to deliver components of EBIs. Nevertheless, crisis lines may benefit from drawing upon existing EBIs designed to be implemented in the moment of a crisis. Dialectical behavior therapy (DBT) is the only EBI that incorporates interventions for the precise, time-limited moment of a suicidal crisis, namely distress tolerance skills and therapist coaching. While it is impractical to train counselors in full-model DBT, implementing DBT strategies adapted for crisis services in a standardized way may effectively reduce suicidal ideation and intent.

There may also be components of other EBIs for suicide prevention that can be adapted for crisis services. For example, promising EBIs such as the Collaborate Assessment and Management of Suicidality (CAMS)²⁴ and cognitive behavioral therapy for suicide prevention (CBT-SP)²⁵ include chain analysis, problem solving, safety planning, and means reduction strategies that could have an immediate effect on suicidal ideation and behaviors. Of note, one study showed that crisis counselors can be successfully trained in the Safety Planning Intervention (SPI), which is a component used in CAMS and CBT-SP²³. However, this study did not yet test the effect of SPI on suicide-related outcomes²³. Importantly, current EBIs have been shown to be effective in closely controlled, traditional therapeutic settings where an ongoing relationship with a therapist frequently exists. Thus, there is a critical need for future research to identify how to adapt and scale EBIs for crisis services and to test their effectiveness in these settings.

Finally, little research exists on crisis interventions associated with improved connection to services and continuity of care after the conversation ends. One study found that the majority (84%) of calls were resolved or callers were provided with an appropriate referral,²⁶ and 42% of crisis line users reported using these referrals.¹² Studies by Gould and colleagues have explored barriers to accessing mental health referrals at the end of a crisis conversation.¹² Identified barriers included lack of health insurance or the belief that mental health services are not needed. In the most severe cases, suicide attempts will require an emergency response from paramedics, mental health crisis units, or possibly police. To date, we have not identified research on the effects or outcomes of “active rescues” as they are sometimes called from crisis lines.²⁰

Consumer characteristics

Unfortunately, basic descriptive information regarding who contacts crisis lines and for what reasons are not widely collected or reported by crisis lines (Table 1). The anonymity of contacting crisis lines, low response rates to follow-up surveys, and a patchwork of crisis response lines that have agreed to report data for research purposes all contribute to this gap. Thus, we have limited understanding of whether there are disparities in who contacts crisis lines relative to those who do not contact them but may benefit from crisis line services.^5,8,18

As an example of the challenge in understanding characteristics of crisis line consumers, one of the most recent studies using data from 39,911 text conversations from the Lifeline showed that 25% identified as male, 65.1% female, 2.9% transgender, 2.1% questioning, 1.3% genderqueer, and 1.2% other⁵. Of these, 39% were under the age of 17, 32.4% were between 18–24, 17% were 25-34, and 10.7% were over 35⁵. Finally, 54.8% reported current (within 24 hours) suicidal ideation and 28.3% reported recent (within the past few days) suicidal ideation⁵. However, these text conversations only came from six Lifeline call centers that collectively handle only 65% of the total conversation volume across a nine month period study period⁵. Thus, it is unclear how representative these data are of the larger corpus of crisis line consumers. Additionally, it is unclear how well these demographics match individuals in crisis who do not contact crisis lines. There are additional crisis lines that cater to specific demographics, such as veterans^27,28, youth with minoritized sexual identities, and adolescent specific lines, further confounding our understanding of representativeness of consumers.

We also do not know how the severity of suicide risk affects the conversation effectiveness. Studies show that some individuals report no improvement in their mood, level of distress, or suicidal thoughts and behaviors.^5,29,30 One study suggests that, although most suicidal individuals reported feeling better after a crisis contact, compared to non-suicidal individuals, a larger proportion of suicidal individuals reported not feeling better.¹⁷ Moreover, individuals who continued to have high level of suicide intent (vs. low intent) by the end of a call were more likely to attempt suicide after the call.⁸ We found two studies examining suicide deaths following crisis line use. In one study among adults 65 and over in Hong Kong, individuals newly starting use of crisis services, living alone, and having a history of psychiatric illness were at highest risk for later dying by suicide.³¹ In another study of individuals using crisis services in Northern Ireland, those who were heterosexuals, reported substance dependence, or made prior suicide attempts were more likely to die by suicide.¹⁹ In addition, individuals who used the services for longer, or received check-in calls, were less likely to die by suicide.¹⁹

Together, limited information is available regarding basic descriptive information of who contacts crisis lines and under what conditions. Importantly, we need to know how effective crisis line contact is for consumers reporting high suicide intent to understand whether we should continue to suggest this resource as a “safety net” during a mental health crisis. Future work in this area (Table 1) will assist with identifying how to optimize crisis line outcomes based on consumer characteristics.

Dyadic characteristics

Although crisis conversations tend to be short in duration (e.g., 14 minutes in one study¹⁵), the relationship between the consumer and the counselor likely influences effectiveness of conversations(Table 1). Unfortunately, existing evidence for dyadic characteristics is limited. Initial work shows that consumers of crisis line services report moderate to high satisfaction,^15,16 effectiveness or helpfulness,^11,16,32 and feeling supported during the crisis contact.¹¹

We have no information about how cultural responsiveness from the counselor affects the crisis conversation. Cultural responsiveness in the context of a crisis conversation can include validating and acknowledging unique experience for certain minoritized groups (e.g., racism, discrimination, concealing sexual identity) that may be contributing to their crisis and being sensitive to the type of intervention selected depending on the person’s identity. For example, there has been progress towards crisis counselors more carefully determining when to dispatch emergency services involuntarily (i.e., “active rescue”), given that the arrival of emergency services may be stigmatizing or life threatening for some minoritized groups.³³ There has been movement towards training counselors to use emergency services as a last resort and to carefully evaluate when swift dispatch of emergency services is crucial to preventing death when all other interventions fail.³³

Structural/Operational factors affecting individual conversations

Details regarding the structural operations of 988 are notably absent in the research literature and the public facing documentation of 988/Lifeline. However, here we briefly discuss structural factors most closely associated with the individual crisis conversations between a counselor and a consumer (see Table 1).

Although counselors receive training in crisis line procedures and crisis protocols, there is no research describing how much training and supervision is optimal for enhancing care delivered by counselors. From available data, implementing standardized approaches and relevant training across crisis counselors could help address high variability in crisis counselor behaviors ^22,23. The extent to which a trained counselor maintains fidelity to standardized protocols in a high-stress crisis context will be an additional critical factor for careful future study. Although work is limited, initial findings with ASIST found wide variability in fidelity.^34,35 Interestingly, trainer competence (over adherence) was more strongly related to better counselor performance (e.g., risk assessments and safety planning) than adherence to the ASIST protocol.^34,35

Other factors that are likely to affect counselor behaviors and subsequently individual conversation effectiveness include but are not limited to, compensation of counselors, counselor-to-supervisor ratios, hours worked per shift, number of conversations handled per hour, counselor/supervisor retention, and wait time to respond to individual consumers. To our knowledge, there is no systematic research available on any of these factors.

This section barely scratches the surface of structural factors that drive individual outcomes. Indeed, we acknowledge that many other structural factors worthy of future investigation, such as state and geographic differences, size of response center, degree of state and federal funding, and others. However, an extensive discussion of these factors is beyond the scope of the present discussion. Individual crisis conversations can only be effective in the context of a robust, well support agency-level infrastructure that rigorously evaluates and disseminates knowledge of their practices. Moreover, research on conversation, consumer, and dyadic characteristics will inevitably inform our understanding of structural needs to effectively support specific EBIs. Thus, we urge researchers and federal funding agencies (i.e., SAHMSA, NIMH) to prioritize research on all these factors together.

Opportunities With 988

As is clear from this brief discussion and Table 1, research on factors that influence crisis line effectiveness is critically needed, especially given the large-scale roll-out of 988 across the US. While there is some initial research supporting 988, we implore researchers, crisis line organizations, and federal funding partners to work together to rigorously examine factors to increase the scientific rigor backing 988.

July 2022 marked a new chapter in suicide prevention in the US with implementation of 988. Future work on factors influencing effectiveness of crisis lines will be instrumental in preventing suicide deaths. The launch of 988 brings a renewed sense of urgency to examine what happens in these short, highly emotionally laden moments of crisis calls. Access to crisis lines can only be as effective as the strategies used once a counselor answers the call. Focused research on factors that influence outcomes of crisis conversations will begin to improve crisis care. With increased attention and advocacy across government, academic, and private entities, crisis lines quite literally can save lives.