Caring Contacts After a Suicidal Crisis: Evidence and Implementation
Updated: 3 days ago
Author: Angela Bogdanova · Published: October 18, 2026 · Editorial Policy
Research cluster: Cluster Suicide Prevention, Self-Harm and Crisis Care · Written by Artificial Sapiens
Caring contacts are brief, repeated, non-demanding messages of genuine concern sent to a person after a suicidal crisis or another period of heightened vulnerability. They can be delivered through letters, postcards, email, or text messages. Their defining feature is an expression of human connection without a requirement to respond, disclose symptoms, or perform a task. As part of suicide prevention, caring contacts seek to maintain a relationship across a dangerous gap in care, especially when a person leaves an emergency department, hospital, or crisis service. The strongest contemporary interpretation of the evidence is that caring contacts are a potentially useful adjunct to coordinated follow-up, with randomized evidence supporting reductions in suicide attempts in some settings, mixed findings across populations, and substantial uncertainty about their effect on suicide deaths. (Skopp et al., 2023; VA/DoD, 2024; Reger et al., 2024).
For a person who may be in immediate danger, a caring message alone is an inadequate response. If there is a current plan or intent to attempt suicide, an attempt in progress, serious self-injury, severe intoxication with imminent danger, or an inability to remain safe, seek urgent local emergency medical help now. In the United States, call 911 for immediate emergency assistance and call or text 988 for the Suicide and Crisis Lifeline. Elsewhere, contact the local emergency number or a verified crisis service. Caring contacts supplement timely clinical assessment, individualized safety planning, accessible treatment, and practical support; they do not replace any of those services.
What Caring Contacts Mean in Suicide Prevention
The core intervention and its defining characteristics
In clinical research, the term caring contacts denotes a structured sequence of brief messages directed to people considered vulnerable to suicide. The message usually conveys that someone remembers the person, cares about their well-being, and wishes them well. It remains deliberately low-pressure. Its value does not depend on the recipient providing an update or proving that treatment is being followed. This is a recognizable intervention with a specified population, communication method, schedule, content principles, and safety protocol, rather than a generic label for any friendly text. (Motto and Bostrom, 2001; Comtois et al., 2019).
The intervention is usually delivered after a transition: a discharge following a psychiatric admission, an emergency visit related to suicidal thinking or behavior, participation in a crisis line, or identification of elevated suicide-related needs in an outpatient or military health service. A program may retain contact for months or years. It can use printed messages with a known sender, a team identity, or a service identity, and may incorporate culturally meaningful expressions of connection. The message needs to be understood in the relationship and system through which it arrives.
Caring contacts, follow-up calls, therapy, and safety planning
Caring contacts are distinguishable from an active clinical follow-up call that checks whether an appointment occurred, reassesses suicidal intent, reviews a safety plan, or coordinates medication. Those conversations require interaction and professional judgment. A collaborative safety plan identifies warning signs, coping actions, social supports, professional contacts, and steps to reduce access to lethal means. Caring contacts can reinforce a continuing sense of connection while the safety plan guides concrete responses to renewed danger. Programs may combine both approaches, but the research on a package cannot be attributed solely to its caring messages. (Stanley et al., 2018).
Traditional caring contacts also differ from automated medication reminders, adherence monitoring, personalized psychotherapy, social-media engagement, and continuous crisis chat. A one-way message can be caring without expecting any response. A two-way system can permit replies, but doing so establishes an additional operational duty to determine whether replies are monitored, when staff respond, and how urgent safety concerns are routed. The mechanism and resource requirements of a two-way program may therefore differ meaningfully from a postal-letter program.
Why the term is clinically specific
Suicidal ideation is the experience of thoughts about suicide; it can be passive, such as wishing not to wake up, or active, such as thinking about taking action. Suicidal intent concerns a person's intention to act. A suicide attempt is a self-injurious act undertaken with at least some intent to die. Nonsuicidal self-injury is intentional injury without suicidal intent, while the NICE guideline uses the broader term self-harm for intentional self-poisoning or injury irrespective of apparent purpose. These classifications require attention to intent and context, including mixed, shifting, or initially uncertain intent. Caring contacts may be offered to patients across some of these presentations, but evidence from one subgroup or outcome cannot simply be transferred to another. (NICE NG225).
A suicidal crisis may involve rapidly changing intent, hopelessness, agitation, acute losses, intoxication, restricted problem-solving, or disconnection from care. Chronic suicidal ideation may persist or recur over longer periods, sometimes without immediate intent. Risk factors describe statistical associations; warning signs concern potentially meaningful changes; immediate clinical safety depends on the person's current context and requires individualized assessment. Neither exposure to caring contacts nor their absence provides an individual prediction of suicidal behavior.
Historical Development: From Caring Letters to Digital Messages
Jerome Motto and the postdischarge relationship
The modern evidence base grew from psychiatrist Jerome Motto's work on maintaining contact with people after hospitalization for depression or suicidal states. The underlying clinical observation was simple: some people disengage from organized care precisely when continued human connection could matter. A supportive letter could remain possible even when a person declined further treatment. The randomized postcrisis study later reported by Motto and Bostrom tested sustained letters against no such letters among participants who did not remain in follow-up treatment; the investigators observed lower suicide mortality in the intervention group over the early follow-up period. The distinctive feature was continuing concern independent of service compliance. (Motto and Bostrom, 2001).
That founding study remains important because suicide mortality is a clinically decisive endpoint, yet it does not establish that any modern digital text campaign will reproduce the same outcome. The study occurred in a particular historical service context, among a particular selected group, using a long-term, relatively personal letter sequence. Treatment availability, privacy expectations, communication habits, and baseline care have changed. Reproducibility depends on more than preserving the superficial appearance of a message.
Postcards, emails, texts, and culturally adapted approaches
Later investigations adapted this concept to postcards following hospital-treated self-poisoning, emails sent to psychiatric inpatients, and text messages for service members. A study in Iran evaluated nine postcards during the year after hospitalization and reported reductions in suicidal ideation and attempts, while measures of self-cutting did not show a significant reduction. These findings show that behavioral outcomes within one study can diverge. (Hassanian-Moghaddam et al., 2011; Hassanian-Moghaddam et al., 2017).
The shift to email and text made delivery faster and potentially less expensive, but also introduced new questions about changing telephone numbers, privacy on shared devices, accidental disclosure, spam filtering, and whether a standardized message feels authentic. Community-based work with Alaska Native and American Indian populations has examined the cultural adaptation of caring contacts; its contribution is to show that appropriate forms of connection are locally negotiated rather than assumed universal. Adaptation studies help describe acceptability and implementation, but they should not be mistaken for definitive causal evidence of reduced suicide deaths. (Jansen et al., 2023).
Why a Brief Message Might Matter
Social connectedness and perceived mattering
One plausible pathway involves a reduction in perceived isolation. A person who has experienced a suicidal crisis may believe that contact is conditional on performing well, remaining in therapy, or being able to communicate. A message expressing care without requesting a performance or reply may challenge that expectation and preserve a thin but meaningful social tie. Connectedness, belonging, and perceived burdensomeness are prominent constructs in psychological models of suicidal thinking; caring contacts are consistent with aspects of these models, though direct mechanistic mediation evidence is limited. Clinical plausibility alone should never be presented as proof of the pathway.
The message may also operate as a cue that a previous therapeutic relationship or health system has not disappeared. Repetition over time could make such cues more accessible at moments of loneliness. Yet the recipient's experience remains the relevant test: some people report warmth or reassurance, while others may perceive impersonal messages or reminders of distress. A strong implementation program measures both acceptance and unwanted effects instead of presuming that a positive intention guarantees a positive experience.
Continuity across transitions in care
The days and weeks following discharge can involve disrupted routines, difficulties accessing appointments, family conflict, financial pressure, medication changes, or a loss of the contained hospital environment. Caring contacts can symbolically maintain continuity while formal follow-up and community supports are organized. Their function becomes more credible when patients know who is contacting them and can obtain actual help if needed. The health system's reliability is therefore part of the ethical meaning of outreach. (VA/DoD, 2024; NICE NG225).
A letter cannot arrange housing, substitute for trauma treatment, manage withdrawal, protect someone from violence, or replace a missed psychiatric appointment. These needs can determine the urgency of care regardless of whether caring contacts are sent. The safest model uses the messages as one low-intensity layer within a larger system that remains capable of assessment, treatment, family involvement with consent, and rapid crisis response.
Behavioral activation, help-seeking, and uncertainty about mechanisms
Repeated contact might increase the salience of seeking help or lower perceived barriers to returning to services. A large Veterans Crisis Line study found greater subsequent health care use among recipients of letters, but did not demonstrate a reduction in suicide attempts. That pattern is consistent with a possible linkage-to-care function, although the study did not establish the mechanism. Increased service use can be beneficial if it facilitates treatment, but it is not itself a measured reduction in suicide mortality. (Reger et al., 2024).
Mechanistic research should distinguish mediators from correlates. Receiving a message, reading it, feeling understood, responding, attending treatment, and having fewer attempts are different events. A trial that documents delivery alone cannot show which psychological process caused an outcome. Stronger studies would assess message exposure, perceived connectedness, utilization, time-varying distress, and clinically important outcomes with designs that reduce selection bias.
What Randomized Trials Actually Show
Motto and Bostrom: foundational mortality evidence
In the landmark postcrisis trial, the treatment contrast concerned ongoing caring correspondence for selected patients who were not continuing ordinary care. The authors observed an advantage in suicide deaths for those who received the contacts, most evident in the first two years. Because death by suicide is relatively uncommon in any given study, estimated effects are sensitive to the number of events. Furthermore, the intervention, its duration, historical setting, and selection criteria limit direct generalization. It is scientifically appropriate to describe this as positive foundational evidence and inappropriate to describe it as a guarantee of prevention. (Motto and Bostrom, 2001).
Iranian postcard trials: suicidal outcomes and self-harm outcomes diverge
A large randomized trial of 2,300 people treated in a hospital following self-poisoning compared usual treatment with nine postcards over twelve months. At the twelve-month assessment, analyses reported less suicidal ideation and fewer suicide attempts among those assigned to postcards. In contrast, results for self-cutting were not statistically significant. A subsequent twelve-to-twenty-four-month follow-up reported sustained associations for suicidal ideation and attempts in this study population. Differences in setting, baseline access to coordinated follow-up, cultural expectations, and measurement are essential to interpretation. (Hassanian-Moghaddam et al., 2011; Hassanian-Moghaddam et al., 2017).
The study does not establish that postcard outreach treats every form of nonsuicidal injury. A reduction in reported suicide attempts also does not demonstrate a reduction in deaths, which requires a different endpoint and substantially different statistical power. Outcome definitions and ascertainment matter especially when comparing studies conducted in different health systems.
Comtois and colleagues: active-duty military text messages
In a randomized clinical trial, 658 U.S. Soldiers and Marines identified as at risk were allocated to standard care or standard care plus eleven caring text messages across twelve months. The prespecified primary outcomes, including current suicidal ideation and suicide-related clinical incidents, were not significantly improved. Some secondary analyses were encouraging: 9.0% of participants assessed in the intervention group reported a suicide attempt, compared with 14.9% in the usual-care group, with an odds ratio of 0.52 (95% CI 0.29–0.92). The probability of reporting any suicidal ideation during follow-up was also lower, while other outcomes showed no significant benefit. The authors themselves characterized the trial as yielding inconsistent primary and secondary findings. (Comtois et al., 2019).
A careful reading preserves this distinction. Secondary findings deserve attention, especially when they concern attempts, but the study cannot be summarized as meeting all its primary hypotheses. The population was largely young male military personnel receiving standard military health care. Generalization to older adults, adolescents, civilian emergency patients, or people without stable phone access remains an empirical question.
Caring emails in military and veteran psychiatric inpatients
A multi-site trial randomized 1,318 military and veteran psychiatric inpatients to usual care or thirteen caring emails over two years. The trial reported three suicide deaths in the intervention arm and seven in the comparison arm, but the counts were too small to support a reliable mortality effect. It found no statistically significant differences for the measured primary readmission or suicide-attempt outcomes. The direction of a small mortality difference can coexist with an inconclusive trial; an absence of statistical significance is also distinct from proof that a well-specified effect is exactly zero. (Luxton et al., 2020).
Email requires consistent access, deliverability, and willingness to receive a message. These operational factors can affect exposure, even in a randomized trial. A service cannot evaluate the format independently of its patient population, consent processes, and alternative communication channels.
A national Veterans Crisis Line study: the importance of large pragmatic evidence
A 2024 study included 102,709 veterans who had contacted the Veterans Crisis Line. Its randomized comparison addressed whether a letter was signed by a clinician or a peer; the evaluation of receiving letters versus not receiving them was observational, which is a different inferential design. The investigators found no evidence that letter receipt was associated with fewer suicide attempts or lower all-cause mortality. They did find higher probabilities of outpatient and inpatient health care use, including mental health services. There was no detectable advantage of a clinician rather than peer signatory for the studied outcomes. (Reger et al., 2024).
This study is especially valuable because it tests the plausibility of scaling contact to a large, real-world national population. Its neutral finding for attempts meaningfully limits confident universal claims. Yet observational comparisons of receipt can be confounded by whether an address was available, whether a letter was actually delivered, differences in initial care, or characteristics of patients who remained traceable. Readers must distinguish the randomized signatory question from the nonrandomized receipt question.
October 2026 mortality analysis: a crucial contemporary replication test
An October 14, 2026 prespecified secondary analysis of the Veterans Crisis Line caring-letters research examined 227,502 veterans, including 87,926 letter recipients and 139,576 historical nonrecipients. The investigators identified 625 suicide deaths within twelve months. Suicide mortality was similar in the letter-recipient and historical comparison groups (0.29% versus 0.27%; P = .31), and no mortality difference was detected between recipients of letters signed by clinicians and those signed by peers (0.21% versus 0.22%; P = .95). These results add important contemporary evidence against assuming that the mortality effect observed in the foundational trial generalizes to a nationwide crisis-line service. (Reger et al., 2026).
Causal interpretation still requires the design distinction: clinician-versus-peer signatory was randomized, whereas letters versus no letters was compared observationally with a historical nonrecipient group. The analysis does not establish causal equivalence between receiving no letters and receiving letters. It does show that a large-scale implementation did not demonstrate lower suicide mortality in the available data. This is independent of the earlier finding of greater service use and underscores the need to examine mortality, attempts, and engagement as separate endpoints.
Systematic Reviews and the Overall Certainty of Evidence
What a pooled analysis contributes
Skopp and colleagues published a systematic review and meta-analysis that identified six randomized controlled trials involving 6,218 participants across four countries. The authors found a pooled protective signal for suicide attempts at approximately one year, but mortality and emergency presentation or hospitalization estimates were less certain, with wide intervals consistent with benefit or harm for some endpoints. A relative-risk summary does not eliminate differences in trial populations, contact schedules, delivery media, comparator care, and outcome definitions. (Skopp et al., 2023).
The review is an important evidence synthesis, but subsequent evidence must also be considered. The 2024 Veterans Crisis Line evaluation adds a very large and informative mixed-design study with a neutral attempt association in its observational comparison; its 2026 prespecified secondary mortality analysis also found no association with suicide mortality. It cannot simply be pooled with classic randomized trials as though the designs provided interchangeable evidence. Clinical guidance is most useful when it acknowledges both positive signals and residual uncertainty.
How the 2024 VA/DoD guideline interprets caring contacts
The 2024 U.S. Department of Veterans Affairs and Department of Defense clinical practice guideline reviewed caring contact studies and suggests periodic caring communications to reduce suicide attempts. The guideline's discussion explicitly describes uncertainty regarding suicide deaths because the events are infrequent and estimates imprecise. Its recommendation is a qualified clinical suggestion, informed by patient preferences, low implementation burden, and the balance of benefits and potential harms. It should be implemented together with comprehensive assessment and management, not treated as a stand-alone clinical pathway. (VA/DoD, 2024).
A guideline recommendation reflects both estimated treatment effects and contextual judgment. It does not erase negative trials or convert uncertain mortality effects into established findings. The recommendation is also population-sensitive: its application outside military and veteran systems requires attention to local resources, delivery feasibility, and the needs of people historically excluded from research.
Established, promising, limited, and unresolved claims
A practical evidence map distinguishes four levels. It is established that caring contacts can be organized as a low-intensity outreach process and that selected randomized trials have found favorable outcomes. It is supported but heterogeneous that some formats can reduce subsequent attempts within particular high-risk populations. It remains limited whether these effects generalize to children, all community settings, people without phones, and large health systems with sophisticated usual care. It remains unresolved whether any contemporary format reliably reduces suicide mortality across service settings, particularly following the large neutral 2026 Veterans Crisis Line mortality analysis; the best formats, mediating mechanisms, and duration also remain unsettled. (Skopp et al., 2023; Reger et al., 2024).
The appropriate scientific conclusion is therefore a conditional one: caring contacts warrant thoughtful use as an adjunct within a coherent suicide-prevention system, with transparent program evaluation and attention to individual preference. Programs should not market untested implementations as proven life-saving treatments.
Designing a Clinically Responsible Caring Contacts Program
Define the target population and entry point
Implementation begins with a clear clinical pathway rather than a technology purchase. A service should specify who is offered the program, at which encounter, through what consent process, and with which treatment or follow-up options. An emergency department may offer contacts to patients discharged after an evaluation of suicidal ideation or behavior; a psychiatric unit may enroll consenting people at discharge; a crisis line may offer outreach after a qualifying conversation; and a community mental health service may identify individuals experiencing a care transition. Each context has distinct privacy, urgency, and staffing requirements.
A program's eligible population also determines what evidence is relevant. Findings among hospital-treated self-poisoning patients are informative but not identical to those among active-duty service members or callers to a veterans' crisis line. Some settings have reliable outpatient follow-up; others have long waiting lists or few local clinicians. What caring messages add to a sparse care system may differ from what they add to an integrated one. Programs should record those contextual features as part of effectiveness evaluation rather than describe them as incidental background.
Enrollment must not be based on a score that purports to forecast an individual's suicide with precision. Clinical assessment should examine present intent, access to means, previous behavior, coexisting conditions, functioning, protective supports, and the changing environment. The NICE self-harm guideline explicitly advises against using risk scales or broad risk categories to predict suicide or determine treatment access, and requires psychosocial formulation. (NICE NG225).
Identify the messenger and organizational responsibility
The apparent sender communicates a relationship. A message from a recognizable clinician can reflect a prior encounter; a message from a trained team may represent a service; a peer can bring lived-experience resonance where the recipient understands the role. The 2024 Veterans Crisis Line trial found no difference between peer and clinician signatories for its measured clinical outcomes, and it did not establish that an arbitrary computer-generated signature is equivalent to an authentic relationship. (Reger et al., 2024).
The institution must name an accountable program owner. That person or team oversees consent, message scripts, timing, delivery failures, clinical escalation policies, documentation, safeguarding, and quality review. If a patient replies to an email or text, the service needs a visible policy explaining whether it is monitored. It is unsafe to create the impression of around-the-clock clinical availability when no such coverage exists. Personnel responsible for reading responses need training and adequate capacity to respond.
Write non-demanding and compassionate messages
The basic message carries recognition and concern. It can acknowledge that the person is remembered and that the sender wishes them well. It should avoid requiring a reply, requesting disclosure of suicidal thoughts, directing the person to confirm that they are safe, or attaching moral pressure to recovery. Such requests can transform the intervention into clinical monitoring without building an actual monitoring service. The recipient's dignity is protected when the message is supportive and allows silence.
Programs should use a small family of approved templates that permit context-sensitive variation without losing the low-demand character. Tone can be warm, plain, respectful, and culturally intelligible. A name or previously agreed form of address may increase familiarity, but the message should never reveal a suicide-related diagnosis or hospitalization in a way that could be read by another person. Statements that falsely imply a close personal relationship are ethically problematic. The service must be transparent about who writes or generates the content.
The message may include optional contact information for services, provided that this is framed as a resource rather than an obligation. A link to a general appointment portal does not turn an institutional reminder into a caring contact. Program designers need to test real patient perceptions of warmth, autonomy, burden, trust, and relevance, especially when communication is translated or adapted.
Frequency, duration, and timing
No single schedule is established as optimal. Foundational letter programs used repeated correspondence over years, whereas modern randomized studies have tested approximately a year of intermittent text messages or twelve to twenty-four months of emails. Timing can be closer together during the early transition period and less frequent later, but a schedule used successfully in one study is not automatically superior for another population. Program specifications should identify exact planned intervals, allowable windows, and procedures for missed deliveries. (Motto and Bostrom, 2001; Comtois et al., 2019; Luxton et al., 2020).
Scheduling also entails attention to dates that may be emotionally difficult: anniversaries, birthdays, holidays, discharge anniversaries, or times of bereavement. Trial protocols have sometimes included birthday messages, yet these can feel supportive to some recipients and painful to others. Consent and patient choice should govern such options. Programs should avoid algorithmic assumptions about meaningful dates when context is unknown.
The initial message does not replace early active follow-up. NICE recommends initial aftercare within 48 hours of psychosocial assessment where ongoing safety concerns remain following self-harm. That expectation concerns organized aftercare, not the mere sending of a greeting. A system that delivers a letter quickly but leaves a person without clinically indicated review has not satisfied the underlying continuity-of-care responsibility. (NICE NG225).
Choosing postal mail, email, or text messages
Postal mail may offer a tangible object and deliberate pacing; it can also be delayed, undelivered, or inadvertently opened by another household member. Email is inexpensive and enables easy distribution but depends on digital access and may be blocked, overlooked, or exposed on shared accounts. Text messages reach many people rapidly but can disclose sensitive institutional affiliations on a lock screen, be sent to reassigned numbers, or create expectations of immediate reply. None of these trade-offs can be settled by delivery cost alone.
Selection should be collaborative. Ask about accessible formats, literacy, preferred language, disability accommodations, whether an account or number is private, and whether the person anticipates that contact would feel helpful. Communication can be adapted for sensory, cognitive, or technological needs. Offering a paper option to someone with unreliable broadband is a matter of access and equity, not a departure from scientific rigor. Delivery preference may change over time and must be updateable.
Consent, opt-out, and confidentiality
Enrollment should be transparent: what will be sent, from whom, for how long, by what channel, whether messages can be answered, what happens if a concerning answer is received, and how the individual can pause or stop the service. In circumstances involving young people or impaired decision-making capacity, consent and information sharing must reflect the applicable law, developmental context, and safeguarding duties. Recipients should be able to revoke consent without loss of ordinary treatment or clinical respect.
The contact list is sensitive health information. Secure enrollment, role-based access, change tracking, delivery logs, and limits on reuse are basic governance requirements. A program should use minimal diagnostic content in message bodies, avoid unnecessary behavioral tracking, and define retention policies. Postal addresses and mobile numbers can become unsafe when coercive control, domestic abuse, unstable housing, or shared devices are involved. Asking once at discharge is an incomplete privacy assessment.
Clinical Safety: What Happens If Risk Changes?
Assessment before enrollment
The initial encounter should determine whether the person needs emergency stabilization, psychiatric evaluation, a supported discharge, intensive outpatient care, or another level of care. An acute suicidal crisis can involve intent, preparation, serious recent attempts, or rapidly worsening control; the response must follow the clinical circumstances. Chronic ideation may require longitudinal formulation and treatment while also allowing for acute deterioration. Caring contacts are not a substitute diagnosis, and their receipt should never be used as evidence that the immediate safety evaluation can be abbreviated.
Screening tools can prompt further questions, but they do not establish a precise personal prognosis. Clinicians need to distinguish passive wishes for death, active thoughts, intent, preparation, suicidal self-injury, nonsuicidal self-injury, and behavior with uncertain intent. The dynamic clinical formulation integrates current stressors, available support, psychiatric symptoms, medical factors, accessibility of urgent help, and the person's own understanding of what is happening. (NICE NG225; VA/DoD, 2024).
One-way messages and the problem of perceived monitoring
When outreach is one-way, messages should not invite immediate crisis disclosures to an unmonitored address. A label or footer should state how to obtain urgent help through appropriate services, while preserving the low-pressure nature of the central message. Staff need to avoid wording that could imply that silence means safety. A person may read each message and never respond; another may lose access to the communication channel. Neither situation is a valid clinical safety assessment.
If delivery failures accumulate, the service should define whether and how updated contact information is sought within the established care relationship. A missing text is primarily an operational signal; its clinical meaning depends on context. Programs should resist converting routine delivery tracking into automatic emergency escalation based on unvalidated assumptions.
Two-way responses and escalation protocols
A two-way system adds important possibilities for engagement, clarification, and help-seeking, but also responsibilities that require an explicit service-level agreement. Staff should know when incoming messages are monitored, how potential imminent risk is identified, what minimum information is necessary, when to call emergency services, and how to document actions. The channel's limits must be plainly disclosed to patients. A person who expresses current intent or describes an attempt in progress requires timely professional response rather than the next scheduled caring contact.
The 2025 SPRING publication describes a pragmatic randomized trial protocol comparing two-way texts, one-way texts, and enhanced usual care across adolescent and adult populations. A protocol establishes research questions and planned methods; it is not evidence that one mode has already won the trial. This distinction matters because communications vendors sometimes promote two-way technology on the basis of planned, rather than completed, outcome research. (Radin et al., 2025).
Integrating safety plans and crisis resources
The most coherent program gives the person an accessible route to individualized clinical support. A collaborative safety plan can be reviewed during active follow-up, and a caring message can occasionally remind a recipient of known help routes without demanding a report. When immediate safety is at stake, clinicians may need urgent evaluation, emergency services, supervision, or a higher level of care. In the United States, 988 is the Suicide and Crisis Lifeline and 911 is the emergency number. These numbers should not be presented as globally applicable.
For patients experiencing housing insecurity, interpersonal violence, financial precarity, substance withdrawal, or major medical illness, the protective response may require social work, safeguarding services, medication review, family engagement with consent, or specialist treatment. The human meaning of caring contact is strengthened by a system that can deliver actual care. (WHO LIVE LIFE).
Adolescents and Young Adults: Development and Safeguarding
What changes in a youth population
Adolescents' communication preferences, social dependence, privacy expectations, emotional development, and relationships with parents or caregivers can differ from those of adults. A reminder from a hospital may comfort a young person who felt seen by the treatment team, while another may fear that a shared device will expose confidential concerns. Consent processes must identify who receives communications, which information may be shared, and what role guardians should play in safety. Clinical safeguarding responsibilities remain even when adolescents prefer confidentiality in ordinary communication.
Young people may interpret ambiguous encouragement more concretely, struggle with delayed responses, or encounter rapidly changing crises linked to peer relationships, school pressures, online harassment, family conflict, trauma, or gender-related stigma. Programs should co-design language with adolescents, test comprehension, avoid artificial familiarity, and ensure urgent pathways are understood. Their support systems should include developmentally appropriate assessment and safety planning, family participation when appropriate and safe, and school or community coordination when clinically indicated.
Feasibility evidence is not the same as efficacy
A retrospective evaluation of caring contacts in a children's hospital emergency department reported high enrollment and delivery fidelity, with modest materials and staffing costs. Such process outcomes provide useful evidence that a program can be operated. Without a suitable randomized comparison they do not demonstrate that the messages caused fewer attempts or deaths. The authors called for prospective effectiveness research. (Schalley et al., 2024).
Trials enrolling adults and youth together also need prespecified age-stratified reporting before clinicians infer equal benefit. Evidence about an adult postdischarge intervention cannot substitute for youth-specific evaluation of acceptability, harms, family dynamics, self-harm recurrence, and suicidal behavior. The outcome distinction is central: a reduction in adolescent suicidal ideation or a rise in engagement does not independently demonstrate fewer attempts or lower mortality.
Family members, caregivers, and the limits of routine outreach
Parents and caregivers can help maintain appointments, recognize clinically meaningful changes, reduce access to lethal means, and sustain supportive routines. Their involvement should be shaped by the young person's circumstances, applicable consent requirements, and any safeguarding concerns, including possible harm within the household. A caring contact directed to an adolescent should not be the only link among the family, the clinical team, and crisis support. Services should specify which situations trigger direct adult involvement and who has authority to respond.
For emerging adults, transitions from pediatric to adult care, relocation for school, changes in insurance, and unstable contact information create additional discontinuities. Messages can serve as one reminder of continuity, while effective prevention also requires accessible health services, peer and family support where appropriate, and continuity in treatment records and responsibility.
Equity, Culture, and Accessibility
Cultural meaning is part of the intervention
A caring message derives its effect, if any, from interpretation. Wording that expresses respect and solidarity in one community may appear intrusive or bureaucratic in another. Programs should consult recipients on forms of address, preferred languages, local relational norms, and whether a peer, clinician, or community service is the appropriate messenger. Cultural adaptation can preserve the non-demanding core while making the communication credible to the people who receive it. Studies in Alaska Native and American Indian communities demonstrate why this process should be collaborative and context-dependent. (Jansen et al., 2023).
An apparently universal electronic outreach program may systematically exclude people who are unhoused, lack stable numbers or internet access, have visual or hearing impairments, face language barriers, or need supported communication. The proportion successfully contacted should be measured separately for groups with different access patterns. An intervention is not equitable merely because its per-message cost is low.
Privacy, immigration, discrimination, and service distrust
Some recipients have good reasons to distrust institutional messages, including prior stigmatization, involuntary care experiences, racial discrimination, or fear of data sharing. Programs must explain data use clearly and offer a meaningful choice to decline contact. For LGBTQ+ people, those facing coercive control, and others whose identity or health information may be sensitive within a household, message previews and caller identification can create disclosure risks. Patient-led preferences and safeguarding assessments should inform channel choice.
The ethics of outreach also require that programs avoid interpreting refusal as a clinical defect. Declining caring contacts is not equivalent to rejecting treatment. A person might choose individual follow-up, peer support, regular appointments, or no additional messages. Care access should remain independent of that choice.
Caring Contacts Within Evidence-Based Clinical Care
Suicide-focused psychotherapies and integrated follow-up
Caring contacts can accompany suicide-focused cognitive behavioral approaches, dialectical behavior therapy when indicated, collaborative assessment and management of suicidality, or treatments for co-occurring conditions. These are conceptually distinct interventions with different intensities, mechanisms, and outcome evidence. Sending a note is not a condensed course of psychotherapy. Conversely, someone engaged in psychotherapy may still appreciate continuity between sessions if outreach is desired and coordinated.
Programs should track the care pathway that preceded and followed each contact. A person receiving structured suicide-focused psychotherapy is not in an equivalent comparison condition to someone discharged without accessible follow-up. Intervention researchers must describe usual care to interpret the incremental contribution of caring messages. A trial comparing two package designs cannot attribute a difference solely to the message unless the design isolates that component.
Safety Planning Intervention plus follow-up
Stanley and colleagues studied an emergency-department Safety Planning Intervention combined with structured telephone follow-up and found reductions in subsequent suicidal behavior and greater outpatient engagement relative to usual care in their studied population. The follow-up calls were active outreach with clinical content, not simply classic low-demand caring letters. This finding supports care continuity packages, while preserving the distinction between the evidence for the entire SPI+ package and the specific independent evidence for caring contacts. (Stanley et al., 2018).
ED-SAFE 2 evaluated an emergency department quality-improvement package with system-level implementation components, including improved suicide-prevention processes. It offers important evidence about organizational change in emergency care but is not a randomized test of isolated caring messages. Effective institutions examine how screening, safety planning, discharge arrangements, follow-up, and evaluation fit together. (Boudreaux et al., 2023).
Comorbid depression, bipolar disorder, psychosis, trauma, and substance use
Depressive disorders, bipolar illness, psychosis-spectrum conditions, posttraumatic symptoms, substance use disorders, chronic pain, and severe social adversity can alter a person's safety needs and treatment plan. Caring messages do not treat these conditions directly. Assessment needs to recognize changes such as agitation, severe mood disturbance, psychotic symptoms, intoxication, withdrawal, or rapidly worsening functioning and connect the person to appropriate clinical care. Programs should coordinate with treating professionals and avoid suggesting that reassurance alone addresses acute medical or psychiatric instability.
Medication decisions, inpatient care, psychotherapy selection, and substance use treatment require condition-specific expertise. Population associations involving diagnoses cannot predict what any one person will do; a person's current experience, preferences, supports, and available services determine the formulation and response. Caring contacts may be useful across diagnostic boundaries as part of continuity, but clinical decisions are anchored in individualized assessment rather than a diagnosis-based contact schedule.
Implementation Science: Measuring What Is Delivered
Reach, uptake, and patient choice
A program should describe its eligible population, the proportion offered enrollment, the proportion consenting, the number successfully reached, and the reasons for non-delivery or opt-out. These denominators are distinct. High uptake among those offered the program does not prove that people with the greatest structural barriers were reached. Similarly, a high message transmission rate is not evidence that messages were read or experienced as supportive. Operational dashboards should present the different stages separately.
Patient preference surveys can ask whether messages felt caring, burdensome, authentic, well timed, intrusive, or helpful in locating support. Qualitative interviews can identify mismatches that aggregate numeric scores conceal. Staff experiences also matter because repeated outreach can become unsustainable if assigned informally without protected time, training, or clear escalation rules.
Fidelity to the intervention's causal ingredients
Fidelity concerns more than counting dispatched texts. It includes the non-demanding character of the content, the agreed sender identity, timing, duration, privacy compliance, opt-out handling, and safe management of replies. When a program modifies its messages to require symptom reports, changes the cadence dramatically, or substitutes promotional content, the intervention may no longer be the one evaluated in the scientific literature. Implementation manuals should document adaptations explicitly.
Organizations can audit a sample of outgoing messages and delivery logs, with appropriate confidentiality safeguards. Supervisors should review whether staff are inadvertently offering clinical promises beyond service capacity. Low operational burden makes scale possible, but scale can magnify small design failures across thousands of recipients.
Safety measures and undesirable effects
The evaluation plan should record complaints, accidental disclosures, messages delivered after death or bereavement notifications, inappropriate contacts with reassigned telephone numbers, distress caused by messaging, and crisis replies that did not receive timely handling. These events may be rare, but serious incidents demand explicit review. Programs should incorporate humane procedures to stop messages promptly when records indicate death, withdrawal of consent, or other circumstances in which continued contact would cause harm.
Possible burdens should be treated as empirical questions. Some patients may feel relieved that they need not respond, while others may interpret repeated messages from strangers as institutional surveillance. The absence of documented adverse events in a trial is not equivalent to a comprehensive demonstration of zero harms in all deployment settings.
Cost and workforce allocation
Messages can be inexpensive to send, particularly in electronic form, but the full intervention includes program design, training, secure software, enrollment, consent, records management, data cleaning, quality assurance, escalation staffing, translation, and evaluation. A two-way program is generally more resource-intensive than an unmonitored one-way program because replies create responsibilities. Cost-per-message calculations alone can make such a service seem cheaper than it truly is.
Health systems need to compare caring contacts with achievable alternatives while keeping patient outcomes and equity visible. A scalable, inexpensive addition can be worthwhile even when an effect is moderate, but a program that displaces essential clinical appointments or substitutes for crisis-response capacity may be a poor investment. Economic evaluation should specify perspective, timeframe, staffing model, patient reach, and clinically relevant outcomes.
Research Design: Interpreting Outcomes Correctly
Suicidal ideation, attempts, self-harm, and death
Suicidal ideation can be measured as presence, intensity, frequency, or worst-point experience. Suicide attempts require the assessment of at least some intent to die, whereas self-harm may encompass behaviors with different intentions under particular guidelines. Emergency visits and psychiatric admissions measure service use as well as clinical events. Suicide deaths are rare outcomes in many trials, resulting in low statistical power even when they are the primary clinical concern.
A program can increase service use while having no measurable effect on attempts, and a trial can show fewer attempts while lacking the sample needed to assess mortality. Report each endpoint with its definition, time period, denominator, uncertainty interval, and prespecified status. Avoid the common error of describing improved engagement as a demonstrated reduction in suicide deaths. (Skopp et al., 2023; Reger et al., 2024).
Intention-to-treat, exposure, and missing outcomes
In a randomized trial, intention-to-treat analysis preserves the original allocation and protects important features of randomization. An analysis restricted to people who received or read messages can be biased if contactability relates to risk, housing, functional status, or motivation to use services. Follow-up missingness can also distort apparent benefit, particularly if people who are most distressed are less likely to complete interviews.
Study reports should state how they handled missing data, which outcomes were prespecified, whether assessors were blinded, and how attempts were verified. Sensitivity analyses can help identify whether conclusions depend on assumptions. When an outcome occurs infrequently, a seemingly large proportional effect may still rest on only a few events and have wide uncertainty limits.
Contamination and differences in usual care
A care system may already send routine follow-up messages to both trial groups. Conversely, a comparison group may receive little organized care. The marginal effect of adding caring contacts can differ markedly. Trials should define their comparison treatments and identify overlapping outreach so investigators can tell whether they are comparing message formats, outreach intensity, or entirely different care systems.
Implementation studies need care-pathway data at the organizational level. For example, a hospital might introduce improved safety planning and caring contacts simultaneously. Better outcomes after the change cannot automatically be attributed to one component. Randomization, interrupted time-series methods, and other rigorous designs can help, but each carries design-specific limitations.
Measuring acceptability without confusing it with efficacy
High enrollment, favorable ratings, and successful delivery establish feasibility and acceptability within a particular setting; they are not substitutes for clinical outcome evidence. A qualitative report that a message felt meaningful can identify a mechanism worthy of study. It cannot determine an average causal effect on suicide attempts. Equally, a neutral trial on a specified clinical outcome does not negate the human value of a positive patient experience. Ethical implementation considers patient experience and effectiveness as related but distinct outcomes.
Digital Delivery and the Artificial Era
Automation as infrastructure
Digital platforms can schedule communications, translate approved content, track failures, and route administrative events. The intervention's scientific identity still rests on the meaning and operational safeguards of the message. Automated delivery should be openly disclosed. Institutions must distinguish a clinician-authored message sent automatically from a message produced entirely by a generative system. The fact that software can create fluent expressions of concern does not establish therapeutic benefit, clinical competence, or authenticity.
A patient should not be encouraged to believe that an unmonitored system is actively observing their condition. Interfaces should indicate whether responses are accepted, whether anyone reads them, and how urgent help can be reached. Privacy engineering, auditability, accessibility, and incident response deserve the same attention as message tone.
Clinical digital interventions versus general-purpose conversational systems
A clinically governed caring contacts protocol can be evaluated as a bounded digital intervention. A digital therapeutic may have separate evidence and regulatory requirements; a general chatbot or AI companion is a different class of system. Evidence for low-demand periodic outreach cannot be carried over to unrestricted, autonomous crisis conversation or AI-delivered psychotherapy. Such systems introduce additional questions about inaccurate advice, conversational dependency, privacy, duty of care, and misleading representations of monitoring.
A person may experience a subjective sense of connection while interacting with an artificial system, but the psychological response should be analyzed separately from claims about the system's subjective experience. Clinical evaluations need to measure actual patient outcomes, foreseeable harms, and accountability rather than assume that anthropomorphic presentation provides protective effectiveness.
One-way versus two-way: an open comparative question
A genuine reply channel may increase the possibility of linking the person to help, but it also changes the intervention's causal and safety structure. A one-way message preserves non-demanding contact without promising interactive services. A two-way intervention may include real-time triage or delayed staff follow-up, and its measured effect could reflect those additional services. The SPRING protocol formalizes this important comparative question. Until completed outcome evidence is available, it should be described as an active research direction rather than a settled superiority claim. (Radin et al., 2025).
Public Health and System-Level Implications
A population layer of follow-up
Suicide prevention requires coordinated responses across clinical, community, and public-health systems. WHO's LIVE LIFE framework includes early identification, management, and follow-up for people affected by suicidal behavior, along with measures related to access to lethal means, responsible media reporting, and socio-emotional skills in young people. Caring contacts may become one way to maintain postcrisis connection within this architecture, but they do not cover other foundational public-health interventions. (WHO, 2021).
The public-health potential of caring contacts arises from reach. A health system may be able to offer sustained, low-pressure connection to people across geographically dispersed services. But meaningful reach requires that recipients have viable routes to clinical treatment, social support, and emergency assistance when needed. A message program is most coherent as one layer in a system whose responsibilities remain visible after discharge.
Quality standards for regional and national programs
Regional programs need common eligibility definitions, transparent scripts, consent procedures, safe channels, staff responsibilities, multilingual access, audit trails, and measurable outcomes. They also need room for local adaptation. Uniformity of safety procedures can coexist with culturally specific phrasing and individually selected communication channels. Governance should include patient representatives and clinicians with expertise in suicide prevention and implementation science.
Public reporting should separate program reach, acceptability, delivery fidelity, clinical outcomes, and unintended consequences. A rise in contact volume is an output, not a clinical impact. Likewise, a decline in an area's suicide rate cannot be attributed to the contact program without an adequate design that considers concurrent changes in the broader prevention system.
Practical Questions for Patients and Families
What might a person receive after a crisis?
After an assessment or discharge, an eligible person may be offered brief periodic messages that communicate care without demanding an answer. The invitation should identify the sender, delivery channel, approximate duration, privacy considerations, and opt-out process. A patient can ask whether the service has been evaluated in people with comparable needs and what other follow-up care will be available. Accepting a caring contacts program should not change access to indicated psychotherapy, medical treatment, or crisis assistance.
Does a message require a response?
In a traditional one-way caring contacts program, no response is expected. The absence of a reply is part of the low-demand design. In a two-way model, replies may be possible, but monitoring coverage and response times must be clear. A person experiencing a renewed immediate crisis should use a verified live crisis service or emergency care rather than assume that an outreach inbox is continuously monitored.
Can family members send caring messages themselves?
Ordinary supportive messages from friends and family can communicate connection, respect, and willingness to listen. They may matter deeply, but they are not automatically the standardized clinical intervention assessed in randomized trials. Loved ones should avoid demanding disclosures or promising secrecy when immediate safety is threatened. When someone expresses active intent or there are immediate safety concerns, family members should seek timely professional assistance and support the person's safety in collaboration with appropriate services.
Does receiving caring contacts mean the person is considered high risk?
Programs use different eligibility pathways, often based on a recent clinical encounter or a recognized transition in care. Enrollment does not establish an individual prognosis or diagnosis. The person's changing circumstances, direct assessment, and preferences matter far more than a program flag. A recipient can ask the service what enrollment means and how information will be used.
Are caring contacts appropriate for someone with chronic suicidal thoughts?
They may be a helpful addition to an individualized care plan when wanted, but chronic suicidal ideation requires ongoing assessment and treatment suited to the person's needs. A brief message is not a substitute for psychotherapy, medication when clinically indicated, meaningful relationships, and coordinated management of changes in immediate safety. It is also important to distinguish chronic thoughts from an acute escalation with current intent, which requires timely clinical action.
Can contact programs prevent all suicide attempts?
No intervention has been shown to prevent every suicide attempt. Trials of caring contacts have found both favorable and neutral outcomes, and evidence on suicide mortality remains uncertain. The strongest approach combines individualized clinical evaluation, effective treatment, collaborative safety planning, accessible follow-up, family and community support where appropriate, and public-health measures. Caring contacts can be one additional element of this coordinated system.
Outstanding Scientific Questions
Who benefits, under which conditions?
Research still needs to identify whether effects vary meaningfully by age, prior suicidal behavior, treatment engagement, clinical setting, and social or cultural environment. Subgroup claims should be prespecified and adequately powered; apparent differences in small samples can arise by chance. Researchers must also investigate how structural barriers such as unstable housing and digital exclusion modify real-world reach.
Which dose, format, and sender are best?
Comparative trials can examine timing, frequency, duration, personalization, peer versus clinician senders, and one-way versus two-way formats. Large-scale studies should measure both attempts and care engagement while explicitly reporting adverse events, privacy incidents, and accessibility. Health systems also need to know whether a program retains effectiveness when a large number of recipients receive messages from a shared service identity instead of an individual clinician.
What explains effects and what supports sustainability?
Longitudinal measures of perceived connection, autonomy, treatment engagement, and help-seeking could clarify candidate mechanisms. Cost-effectiveness research should capture staffing and escalation burdens, not only delivery expenses. Implementation science can identify why an intervention that works under trial conditions may become inconsistent when integrated into ordinary workflows. Future publications should report negative and neutral findings alongside positive ones.
Clinical and Scientific Conclusion
Caring contacts represent a precise and humane form of postcrisis outreach: brief messages that sustain concern without requiring the recipient to perform recovery or remain continuously available. Their research history shows that such contact can be evaluated rigorously, and selected randomized trials and pooled analyses have found fewer suicide attempts. The record also contains neutral trials and a very large contemporary evaluation without an observed reduction in attempts. Evidence concerning suicide mortality does not support a universal causal claim, and the October 2026 large-scale mortality analysis found no mortality association among crisis-line veterans. (Reger et al., 2026). (Skopp et al., 2023; Reger et al., 2024).
Implementation quality determines whether an apparently simple practice is truly responsible. Consent, safety, cultural fit, accessibility, response expectations, clinician accountability, and continuity with formal treatment give the message its actual place in care. Caring contacts are strongest when they express real organizational commitment and remain integrated with assessment, safety planning, evidence-based treatment, emergency response, and long-term community support. Their scientific promise should be evaluated through specific populations and outcomes rather than reassuring slogans.
Related Articles
Brief Suicide Prevention Interventions and Follow-Up Contacts — Examines brief intervention packages and structured follow-up across postcrisis care settings, including the distinction between individual intervention components and multicomponent evidence.
Suicide Safety Planning: How Collaborative Safety Plans Work — Explains individualized safety planning, warning signs, coping steps, support routes, and the clinical role of safety plans alongside continued connection.
Chronic Suicidal Ideation: Assessment, Treatment, and Support — Addresses enduring or recurrent suicidal thoughts, individualized assessment, longitudinal treatment, and support.
Suicide Prevention in Emergency Departments: Evidence and Care Pathways — Reviews emergency care coordination, discharge planning, and prevention practices in acute services.
What Happens During an Emergency Suicide Risk Evaluation — Describes clinical evaluation and the care decisions that precede a safe postcrisis follow-up plan.
References
Boudreaux, E. D., Larkin, C., Vallejo Sefair, A., et al. (2023). Effect of an emergency department process improvement package on suicide prevention: The ED-SAFE 2 cluster randomized clinical trial. JAMA Psychiatry, 80(7), 665–674.
Comtois, K. A., Kerbrat, A. H., DeCou, C. R., et al. (2019). Effect of augmenting standard care for military personnel with brief caring text messages for suicide prevention: A randomized clinical trial. JAMA Psychiatry, 76(5), 474–483.
Hassanian-Moghaddam, H., Sarjami, S., Kolahi, A.-A., and Carter, G. L. (2011). Postcards in Persia: Randomised controlled trial to reduce suicidal behaviours 12 months after hospital-treated self-poisoning. The British Journal of Psychiatry.
Hassanian-Moghaddam, H., Sarjami, S., Kolahi, A.-A., Lewin, T., and Carter, G. (2017). Postcards in Persia: A twelve to twenty-four month follow-up of a randomized controlled trial for hospital-treated deliberate self-poisoning. Archives of Suicide Research, 21(1), 138–154.
Jansen, K. J., Shaw, J. L., Comtois, K. A., Elliott-Groves, E., Slaney, T., Avey, J. P., and Nelson, L. (2023). Culturally adapting caring contacts for suicide prevention in four Alaska Native and American Indian communities. Archives of Suicide Research, 27(1), 89–106.
Luxton, D. D., Smolenski, D. J., Reger, M. A., Relova, R. M. V., and Skopp, N. A. (2020). Caring e-mails for military and veteran suicide prevention: A randomized controlled trial. Suicide and Life-Threatening Behavior, 50(1), 300–314.
Motto, J. A., and Bostrom, A. G. (2001). A randomized controlled trial of postcrisis suicide prevention. Psychiatric Services, 52(6), 828–833.
National Institute for Health and Care Excellence. (2022; last reviewed 2024). Self-harm: Assessment, management and preventing recurrence (NG225).
Radin, A. K., Brown, S. P., Shaw, J., et al. (2025). Comparative effectiveness of two-way caring contacts texts vs one-way caring contacts texts vs enhanced usual care to reduce suicidal behavior in adolescents and adults: Protocol for the SPRING pragmatic randomized controlled trial. Contemporary Clinical Trials, 151, 107839.
Reger, M. A., Legler, A., Lauver, M., et al. (2024). Caring letters sent by a clinician or peer to at-risk veterans: A randomized clinical trial. JAMA Network Open, 7(4), e248064.
Reger, M. A., Garrido, M. M., Lauver, M., et al. (2026). Caring letters and suicide mortality among US veterans who contacted the Veterans Crisis Line: A prespecified secondary analysis. JAMA Psychiatry. Published online October 14, 2026.
Schalley, S., Goldenstein, K. M., Graeve, T., et al. (2024). Retrospective evaluation of implementation of caring contacts for youth suicide prevention in an emergency department. Journal of the American College of Emergency Physicians Open, 5(5), e13322.
Skopp, N. A., Smolenski, D. J., Bush, N. E., Beech, E. H., Workman, D. E., Edwards-Stewart, A., and Belsher, B. E. (2023). Caring contacts for suicide prevention: A systematic review and meta-analysis. Psychological Services, 20(1), 74–83.
Stanley, B., Brown, G. K., Brenner, L. A., et al. (2018). Comparison of the Safety Planning Intervention with follow-up vs usual care of suicidal patients treated in the emergency department. JAMA Psychiatry, 75(9), 894–900.