Suicide is a leading cause of death in the United States, with rates increasing by >30% between 1999 and 2019. The suicide rate for veterans is 57% greater than the rate for nonveterans (
1). Veterans Crisis Line (VCL) callers have an increased risk of death by suicide compared with the general veteran population, regardless of the reason for their call (
2). In response to the rising suicide rates, the U.S. surgeon general issued a call to action in 2021 that included a recommendation to use caring letters when gaps in care may exist, including after crisis line calls (
3).
Caring letters (a.k.a. caring contacts) is an evidence-based intervention for postacute care that involves sending brief, nondemanding messages of care and concern over a period of 1 year (
4–
7). The caring letters intervention has primarily taken the form of typed postal mail, including letters (
4,
5), flat cards mailed in envelopes (
8–
12), and greeting cards (
13). More recently, messages have been sent via e-mail (
14) and text message (
15–
17). Results of studies examining the efficacy of caring letters have been mixed (
6). Some studies have reported a reduction in suicidal behaviors (
4,
8,
13,
15), but others have not (
18,
19). A meta-analysis indicated a protective effect associated with suicide attempts 1 year after initiation of caring letters (
6). Clinical practice guidelines suggest sending caring letters in addition to usual care after a psychiatric hospitalization for suicidal ideation or suicide attempt (
20). The recommendation from the surgeon general to consider caring letters after crisis calls represents a novel adaptation of the intervention.
To extend the reach of its prevention services, the Department of Veterans Affairs’ (VA’s) VCL developed a caring letters project (
21). VCL services are available 24/7 via telephone, text message, and online chat. The VCL serves veterans, active duty service members, and third parties who have concerns about veterans or active duty service members. Reasons for contacting the VCL vary, but the most common concerns are related to mental health (
22). To adapt caring letters for this new population, the team, which included caring letters experts, drafted messages that were consistent with the evidence base (e.g., nondemanding and caring) and informed by pilot studies (e.g., of veteran preferences) (
23). The team presented the project and messages to two veteran engagement groups and a VCL veteran staff member for feedback. Their recommendations informed changes reflected in the final version. Through a centralized mail service, letters were sent to veterans who called the VCL, identified themselves, and received VA care. Caring letters became part of VCL usual care; the letters were not discussed in calls but were mailed after the call. In the first 12 months, these letters were sent to >100,000 veterans living across the United States and its territories (
21). A formative evaluation indicated that veterans found the intervention helpful by giving them a sense of hope, renewed faith in the VA, and a sense that someone cares (
21).
The project is being examined in an ongoing parent study (
21), a hybrid effectiveness-implementation (
24,
25) type 1 trial, to evaluate the effectiveness of caring letters in this population and to identify barriers and facilitators to implementation of the intervention. For this study, qualitative interviews were conducted with veterans (N=23) receiving caring letters to understand their experiences. This article describes the results of these qualitative interviews.
Methods
Study Design
The parent study was a pre-post evaluation that used a type 1 hybrid effectiveness-implementation approach (
21). Its primary goal was to evaluate the effectiveness of caring letters on suicide attempts. The secondary goal was to evaluate barriers and facilitators to implementation of the intervention. This program evaluation project was reviewed by the authorized program office (the VA’s Suicide Prevention Program) and met criteria for classification as nonresearch, as described in VA policy; therefore, institutional review board approval was not required (
26).
Participants
Starting in June 2020, letters were mailed to all veterans who called the VCL, identified themselves, and used VA services. Veterans were automatically enrolled if their mailing address was present in the VA electronic health record. All others who contacted the VCL (e.g., concerned friends and family members) were excluded, as were veterans who died before mailing began. Eligible veterans were identified weekly. Individuals who called the VCL again within 12 months of their enrollment did not receive additional sets of letters.
Caring Letters Intervention
In this study, caring letters were flat cards mailed in a light-blue envelope, accompanied by a national mental health resource card (images of the cards have been previously published [
21]). They were mailed monthly for 4 consecutive months and every other month thereafter, for a total of 1 year (i.e., months 1, 2, 3, 4, 6, 8, 10, and 12). On the basis of veterans’ feedback, a card was also mailed on Veterans Day (
23). Different message content was created for each time point; each veteran received the same set of nine cards. Veterans were randomly assigned to receive all their cards from either a clinician or a veteran peer signatory, both of whom worked with the VCL.
Recruitment
Veterans who had been mailed all nine letters were sent opt-in letters describing the purpose of the evaluation and inviting them to participate in a telephone interview. Veterans were offered $40 compensation to participate. We oversampled women, who comprised 25% of the recruitment sample. Using simple random sampling stratified by sex, in January 2022, the study team sampled 500 veterans, with a 3:1 male-to-female weighting, who were mailed all nine letters. To avoid overburdening the interview team, opt-in letters were mailed in batches of 50 (men, N=38; women, N=12) every other week until the recruitment goal of 25 participants was met (
27). In total, 250 letters were mailed. Twenty-five veterans volunteered to participate, and 23 completed telephone interviews; the remaining two individuals did not recall receiving the cards and could not participate. Interviews were completed in March 2022.
Data Collection
The qualitative analysis team included a doctoral-level anthropologist team lead (T.H.A.), a research scientist and Army veteran with a background in social work (J.A.W.), and a research scientist and Air Force reservist (N.D.C.). All had experience with qualitative evaluation, health services research, and caring letters. At the beginning of each interview, the interviewer (T.H.A. or J.A.W.) explained the data collection procedures, described the participant’s rights to decline to answer questions or end the interview at any time, and obtained consent to audio-record the interview. Interviews were conducted with a semistructured interview guide (see the online supplement to this article) that included three broad domains of interest informed by the goals of the evaluation: the impact of receiving caring letters, perspectives about caring letters (e.g., likes, dislikes, and preference for signatory), and recommendations for improvement.
Analysis
Audio recordings were transcribed verbatim by a team member experienced in qualitative analysis (N.D.C.). Transcripts were analyzed in two phases. First, two team members (T.H.A., N.D.C.) developed individual templates by using content analysis, and second, the analysis lead (T.H.A.) conducted matrix analysis by using constant comparison (
28). Templates were structured with deductive domains informed by evaluation goals (i.e., perspectives about caring letters). An “other” domain was included to capture unanticipated responses. The team developed inductive categories within each domain, reflecting participants’ perspectives (e.g., categories within the domain of general perspectives included “sense of being cared for” and “liked having resources available”; see the
online supplement) (
29). The lead synthesized data from the 23 individual templates into one participant-by-domain matrix (i.e., the matrix showed each participant’s responses in a single row, organized by domain columns). Using constant comparison, she identified the full range of responses received for each domain (
30). To ensure that data collected from interviews were summarized accurately, the two staff members, who were veterans (J.A.W., N.D.C.) and had conducted or listened to the interviews, confirmed that the templates and matrix matched what they had heard.
Results
Interviews were conducted with 23 veterans. Participants included 16 men and seven women, with a mean age of 53 years (range 25–74). Participants were from 16 U.S. states. They reported their race-ethnicity as non-Hispanic White (N=13), Hispanic White (N=2), and Black/African American (N=8). Interviews were 7–34 minutes long.
Positive Impact
Most participants reported that receiving caring letters had a positive impact on their social or emotional well-being. Participants described feeling appreciated, cared for, supported, connected, encouraged, heard, and seen after receiving the letters. One participant stated, “When you go to the mailbox it feels good to actually have something to open up and be acknowledged and to feel seen and to know that someone is thinking of you” (participant 23). (Participant numbering began at 11 instead of 1 for administrative reasons.) Another participant said, “In the military, . . . we’re used to having to suck it up and move forward. When you guys reach out to us, I feel like somebody is listening” (participant 21).
Some participants described aspects of caring letters they perceived as positive. Participants expressed that they liked the length of the message, timing of the letters, look and feel of letters and envelopes, list of resources, message sentiment, and receiving a special letter on Veterans Day. Two participants appreciated efforts to ensure privacy and confidentiality (e.g., messages did not include personal information from their VCL contact, and letters arrived in a sealed privacy envelope).
Actions Taken
Participants described using the resources from the list provided, for example, to connect with mental health care. Some indicated that the letters helped them stay engaged with existing mental health care. One participant stated, “I used a couple of [the resources]. I saw a psychiatrist, and I got to the VA clinic here” (participant 29). Another said, “The card is a reminder to keep in touch with my therapist, which is nice” (participant 15). Some participants reported keeping at least one of the cards: “I did keep them. I think I put one in a photo album” (participant 29).
Changing Perspectives
Some participants expressed that receiving caring letters shifted their perspective toward seeking help, VA providers, or the VA in general: “It felt nice [to get the letters]. It felt like it was an invitation to call back if I needed to, and I would say if I did need to at that time, I would have felt more comfortable making that phone call [because of the letter]” (participant 22). Another participant commented, “Yeah, it change[d] my perspective. I didn’t know that I was going to be communicated with afterward, so that was good” (participant 15).
Dislikes
Although participants largely noted positive aspects of the letters, some described aspects that detracted from the intervention’s impact. No participant stated disliking the entire intervention. However, one participant reported that the number of letters mailed was excessive. Another noted that the cards all looked the same, and the participant therefore assumed that the cards all said the same thing. Similarly, some participants felt the messages were repetitive and generic. One noted that, after the first card, the messages seemed less personal because “there wasn’t anything different or saying, ‘How are you doing?’ or anything like that. It was just, ‘Hey, we’re concerned still’” (participant 14).
Two participants noted that some veterans might feel as though their confidentiality was compromised by the mailings (e.g., if an unintended recipient opened the envelope): “It did cross my mind that if somebody intercepted this . . . that would be my only concern. It’s such a private matter, and if somebody in my family saw this, and I didn’t already tell them I was going through that, that [would be] a really difficult situation” (participant 22).
Preference for Signatory
Participants often could not remember who had signed the letters they had received. Some incorrectly recalled having received cards signed by both a provider and veteran. Participants who could remember the signatory expressed a range of perspectives regarding which signatory would have been the most impactful:
I would be more comfortable with it coming from doctors. (participant 28)
Veteran to veteran, I think, is a little more meaningful. . . . But also, it felt good that your situation got the attention of a doctor or somebody with a little bit more authority. (participant 23)
It did feel good . . . receiving that from the VA staff, but I would say [that] it did feel a little more personal to receive it from a peer, and I would say it gave me that feeling of being supported, not feeling alone. (participant 22)
Regardless of whether recipients remembered the identity of the signatory, they vividly recalled the impact that receiving caring letters had on them: “I can’t remember his name. I don’t know [whether it was a veteran or clinician]. I just thought it was a friend reaching out to me, and it really didn’t matter. It didn’t matter about nationality or religion. It just mattered that this person reached out to me” (participant 29).
Some participants alluded to unexpected benefits related to the signatory always being the same: “It was the same person [who had signed the other letters], and that right there was nice. There was stability that you don’t necessarily have” (participant 29).
Recommendations and Feedback
Participants provided recommendations for making caring letters more effective. These recommendations sometimes contradicted other recommendations or the evidence supporting the intervention. The recommendations included adding a follow-up call from the VCL, including protocols to assess whether a veteran is receiving appropriate mental health care, and reducing the frequency of letters. One participant said, “I think after that first one, you could send one and then between the 30- to 60-day period send a follow-up just to check in. That first note felt good, but once you get two, three, four, five, six, it begins to seem like marketing mail” (participant 21). Veterans also mentioned that adding calls could increase connection, improve access to care, and ensure privacy.
In contrast, other recommendations were not to change the intervention and to keep the mailing frequency the same: “I think they are pretty good the way it is. It is good they send more than one” (participant 18). Some participants recommended varying the envelope color to indicate that the card messages were different. Participants had conflicting suggestions regarding the resource card. One wanted the resource information printed on the back of the letter to save paper, whereas another stated that it was nice to have the letter and resources separate. A final recommendation was to vary the signatory, having the first card come from a provider and the next from a veteran peer.
Discussion
The results of this evaluation of the caring letters intervention for veterans who had contacted the VCL present the experiences and perspectives of recipients of these letters. Participants discussed the impact of the intervention in terms closely related to the hypothesized mechanisms for caring letters (
31). Many participants stated that caring letters made them feel cared for, supported, connected, and heard. Leading theories of suicide emphasize the negative psychological consequences of isolation and loneliness (
32). Although quantitative results from this evaluation, to be reported in a future article, are needed to examine the intervention’s effectiveness, the increased feelings of social support reported by the participants in this study are expected to decrease suicide risk. Systematic reviews and meta-analyses suggest that caring letters can reduce suicidal behaviors in other populations (
6,
20). Additional research is needed to supplement these preliminary qualitative results.
Many participants found caring letters useful for supporting their mental health needs. Participants described using the information provided to access new mental health services. In the formative evaluation (
21) and this summative evaluation, many participants described saving some of the cards or resources. Some reported a shift in their perspective toward seeking help from the VA. We hypothesized that several mechanisms of the intervention support increased access to care. The quantitative evaluation will examine the impact of caring letters on rates of veterans’ use of VA clinics.
Participants also reported some aspects of the intervention that they disliked, but none disliked the intervention as a whole. This finding is consistent with the low opt-out rate we observed. In the first year, letters were mailed to >102,000 unique veterans; only 36 opted out (
21). Participant dislikes included that the cards looked similar, messages were repetitive or generic, and the number of cards sent was excessive.
It is interesting that veterans often could not recall from whom they had received letters, yet they spoke eloquently about how the letters made them feel. This scenario replicates a preliminary finding in the formative evaluation (
21) and a previous caring letters study with veterans (
11,
12). In both studies, letters were sent from someone the patient had never met. It is possible that the sentiments communicated in such messages are more important than the signatory. In the present study, preferences for the peer or clinician signatory were mixed. Because veterans were randomly assigned to receive cards from one of the two signatories, the quantitative evaluation as part of the parent study will determine differences in outcomes by signatory.
Participants provided a variety of recommendations and varied feedback (e.g., use of several different envelope colors and review of the number of mailings), which will be examined in conjunction with results from the full evaluation. Feedback on privacy and confidentiality was important. The intervention strives to meet all federal and VA privacy requirements, and veteran acceptability is critical. Some participants appreciated the characteristics of the intervention that promoted privacy, but others questioned the unintended consequences. Most of the cards did not mention the VCL, but the first mailing indicated that it came from a clinician or a peer veteran who works with the VCL. This approach was considered important to explain who had mailed the letters and why the veteran received them. To guard the letter recipients’ privacy against unauthorized opening of the letters by others, the intervention was developed with additional privacy protections by stating in the first card that veterans contact the VCL for all kinds of reasons, including simple needs like VA scheduling questions (as opposed to stating that the recipient called about a mental health crisis). Envelopes with privacy features such as opaque paper were used, and veterans have the right to request a confidential communications address for VA mail.
This study had several limitations. Given that this project was conducted with the VCL, only veterans using VA services who called the VCL received the intervention. It is unknown how well these results may generalize to users of other crisis lines. As with other programs using postal mail, recipients were limited to those with a mailing address. The results were limited to veterans who received caring letters and were willing to participate. Veterans who did not like the intervention may have been less willing to participate in an interview.
Conclusions
Participants reported that the caring letters intervention had several positive social and emotional impacts and stated that they felt appreciated, cared for, encouraged, and connected. Some reported using the resources that were sent with the letters, and others stated that their view of the VA had improved. Dislikes included the cards looking alike and the repetitiveness of messages. No clear preference was evident for having either a veteran or a provider as signatory on the message. The results indicate that caring letters delivered after contact with the VCL were well received by participants. Future work will examine the impact of caring letters on outcomes among veterans, such as service utilization and suicidal behavior.