_____________ Ande Nesmith, Ph.D., LISW, is an Associate Professor, School of Social Work, University of St. Thomas, St. Paul, MN 55105. Copyright © 2018 Author, Vol. 18 No. 4 (Fall 2018), 1147-1164, DOI: 10.18060/21590 This work is licensed under a Creative Commons Attribution 4.0 International License. Reaching Young People Through Texting-Based Crisis Counseling: Process, Benefits, and Challenges Ande Nesmith Abstract: Texting-based crisis intervention counseling reaches young people who suffer from mental health issues at high rates yet hesitate to seek help. As a new interface, it is neither well-researched nor well-understood. This study examined 49 randomly selected text counseling transcripts and key informant interviews with two counselors to identify unique characteristics of the text counseling process and learn texter reactions to the sessions. Texters presented problems that were similar to those reported in voice-based hotlines. Texters valued the privacy and flexibility of texting that permitted them to receive help immediately regardless of their location. Counselors reported that they must be brief and direct with questions and avoid assigning emphasis to words. The written format required that both parties must be explicit and clear to convey their messages accurately. Both texters and counselors suggested that the texting option might lead young people to seek help that they might otherwise avoid. Recommendations include specialized training on strategies to assess and connect with texters using only the written word and research to develop best practices for texting-based crisis intervention services. Keywords: Crisis intervention; technology; at-risk youth; suicide prevention; texting counseling services Crisis intervention services with a texting option is an innovative approach to reaching at-risk young people. Texting-based hotlines are surfacing in large metro areas at runaway and homeless youth programs across the country (e.g., Safe Place (n.d.) in Chicago; Teen Line (n.d.) in Los Angeles; The Bridge for Youth (n.d.) in Minneapolis), yet little to no published information exists about their efficacy, intervention approaches, or even how much they are used. Recently, the national Crisis Text Line, Inc. (Crisis Trends, 2016), made its aggregate data available online, but by and large, information about most text lines appears in program annual reports rather than peer-review research. This study aimed to address that gap by examining how texters convey whether or not their needs were addressed during the session, as well as techniques the counselors used to assess and connect with service users. Texting-based crisis lines may appeal to young people because texting is currently the primary mode of communication among this age group (Lenhart, Smith, Anderson, Duggan, & Perrin, 2015). Because youth are comfortable with texting, these services may reach a cohort that is at high risk of suicide and other mental health problems (Centers for Disease Control [CDC], 2014). Preliminary evidence suggests that many young people do not choose between voice versus text crisis lines when it comes to highly sensitive and painful topics. Rather, they would rather go without help than talk in person or on the phone (Evans, Davidson, & Sicafuse, 2013; Gibson, Cartwright, Kerrisk, Campbell, & Seymour, 2016). It is therefore important to understand texting-specific counseling strategies for https://creativecommons.org/licenses/by/4.0/ ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1148 providing crisis intervention and whether service users find the intervention useful in this new interface. Literature Review The goals of crisis intervention broadly are to assess, deescalate, establish a safety plan, provide support, and offer referrals, and have demonstrated positive impacts in voice calls (Kitchingman, Wilson, Caputi, Woodward, & Hunt, 2015; Ramchand, Jaycox, Ebener, Gilbert & Barnes-Proby, 2017; Stevens & Ellerbrock,1995). In texting format, the goals of crisis intervention are no different, but there is some evidence to suggest the population reached through texting draws more adolescents, as will be discussed here. Texting-based crisis counseling has the potential to reach an underserved population that is in urgent need of help. The most prevalent chronic conditions found in pediatric health care are mental health disorders (Santor, Poulin, LeBlanc, & Kusumakar, 2007; Saunders, Resnick, Hoberman & Blum, 1994). Untreated mood disorders such as depression are directly related to suicidal behavior which, in turn, is the third leading cause of death among young people aged 10 to 24 (CDC, 2014). A fraction of adolescents with depressive disorders seek professional help; in fact, the more severe the depressive symptoms, the less likely they are to pursue help (Sawyer, Sawyer, & La Greca, 2012). This is reflected in voice-based crisis line use; in some studies, adolescents and young adults comprise as little as 2% of voice hotline users (Gould, Greenberg, Munfakh, Kleinman, & Lubell, 2006) and 4% in the state affiliate of the National Suicide Prevention Hotline used for this study (Canvas Health, 2011). Adolescence is a time in which social connection and peer perceptions are of heightened importance, making stigma over mental health concerns particularly powerful. A study of high school student perceptions about mental health services found that stigma was the most-named barrier to accessing school-based services (Bowers, Manion, Papadopoulos, & Gauvreau, 2013). Other barriers include: fear of family reactions, concerns about the expense of care, and lack of adequate services especially for trauma (Bowers et al., 2013; Lipmann, 2010). There is currently no research study or central location that tracks how many programs offer a texting option for crisis services. The few programs that do provide such services report positive findings and a rapidly growing volume of contacts from young people (Kaufman, 2014). The national Crisis Text Line (Crisis Trends, 2016) boasts messages numbering in the millions in the few years since its inception in 2013. The program used for the current study, TXT4Life, began in 2011 and reported a six-fold increase of use in its second year of operation, which included a dramatic rise in the percentage of adolescents and young adults seeking help compared with voice services (TXT4Life, 2013). TXT4Life has since closed, but there continues to be texting-based crisis lines both locally and nationally (Crisis Text Line, 2018; SAMHSA, n.d.) There is a growing body of research pointing to the benefits of socially interactive technologies which may lead to more innovative use of it in mental health services. Pew Research found that 88% of teens use texting to connect with their friends - with 55% texting friends daily - and 83% reporting that social media made them feel closer to their https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Kitchingman,+Taneile+A./$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Kitchingman,+Taneile+A./$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Caputi,+Peter/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Caputi,+Peter/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Hunt,+Tara/$N?accountid=14756 Nesmith/ REACHING YOUNG PEOPLE 1149 friends (Lenhart et al., 2015). As more people of all ages have embraced texting, it has become a more widely accepted means of communication. According to a 2014 Gallup poll (Newport, 2014), texting is now the most common form of communication among Americans through age 50. Electronic communication in mental health services has increased over the last decade. The first crisis line in the U.S. to accept text messages was launched in 2010 in Nevada (Evans et al., 2013). This line yielded immediate use among high schoolers, with more than half using it more than once. A systematic review of “telemedicine” found that providers used a variety of interfaces to conduct mental health interventions, including email, Facebook, video conferencing, and online groups (Martin et al., 2011). On the whole, while providers did not view these means as equivalently effective as face-to-face, they reported that they were adequate or almost as good. Moreover, they observed significant improvements in their patients. Other research provides examples of online venues to conduct longer-term counseling, including group therapy, and integrating cell phones into in-person therapy sessions rather than battling youth to put their phones away (Dubus, 2015; Koltz & Tarabochia, 2014). The appeal of texting may be explained in part by the Online Disinhibition Effect. Suler (2004) examined this phenomenon to explain online behavior that is discordant with one’s in-person behavior, where positive expressions are more candid and negative ones may be more aggressive. One factor linked to this effect is dissociative anonymity in which the texter has control over nearly all information, especially affect, that is revealed about him- or herself. That control may generate a sense of safety when discussing painful topics. Another factor is the asynchronicity of the exchange which permits texters to delay revealing their immediate reactions, allowing them to “suspend time” while considering a response (Suler, 2004). As an emerging mode of crisis service provision, texting-based counseling is not well understood. To some extent, this is an artifact of rapidly changing technology. Research on technology and mental health services in the past ten years centered primarily on computer- based communication rather than mobile device communications, including email, online groups, chat rooms, or video-conferencing (Krysinska & DeLeo, 2007; Martin et al., 2011). However, receiving crisis services via a mobile device that allows one to access help in any location at any time is arguably different from computer-based services. After an extensive search of available research across several scholarly databases, only four studies surfaced that investigated some aspect of texting-based crisis intervention and one that assessed chat sessions (Evans et al., 2013; Gibson et al., 2016; Haxel, 2013; Mokkenstorm et al., 2017). Evans et al. (2013) conducted a pilot study of a texting helpline, using focus groups with 113 high school youth and post-intervention demographic and service data from the crisis counselors. Most (59%) of the youth accessing the services were not in crisis nor suicidal, but rather sought information or support. About a quarter (26%) of their texters closed the session without resolution, most often by not responding to further texts. The counselors provided very few referrals, in part because their texters did not want outside help. They also observed that youth reported they “preferred a text-based line specifically ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1150 because they would not be identifiable by voice cues and no one could overhear their conversations,” possible indicators of the Online Disinhibition Effect (Evans et al., 2013, p. 479). The texters also reported that the anonymity made it easier to disclose more details about their problems. Haxell (2013) examined text transcripts and interviewed counselors in a New Zealand program. While the service users had a choice between voice and text-based crisis lines, “20% of the helpline user screens would be filled with messages asking, if not demanding, that counselling be provided by text” (p. 149). The reasons users presented for choosing text-based services support the theory behind the Online Disinhibition effect: texting was less scary than talking, it was hard to talk when crying, or the texter was shy or struggled with talking. Gibson et al. (2016) interviewed adolescents in New Zealand who used at least one form of mental health services via in-person, phone, or texting. The formats that were not in-person were attractive to youth because they could “bypass adult control and use these privately and retain their sense of autonomy” (Gibson et al. 2016, p. 1063). The participants who used text services indicated that they preferred texting over voice calls because it was a medium of communication they were used to and because it made them feel more confident about their communication. Chat sessions are similar to texting, though they originated for use on a stationary computer rather than a mobile application, which offers less flexibility in use than texting. Mokkenstorm and colleagues (2017) compared chat sessions in a Netherlands program to voice hotline data from the U.S. The chat users were younger, had more mental health problems, and were more likely to be suicidal than hotline users. While up to 49% had elevated moods at the end of the session relative to the beginning, more often there was no change. Given the risks associated with young people who grapple with mental health issues and the limited use of crisis hotlines among this same group, this is an important gap in service provision worthy of further investigation. The scant research available from the four studies on texting-based crisis services suggests that youth may be interested in, or even prefer, the texting or chat interface and the sense of anonymity may be an important factor. While only one study examined outcomes (Martin et al., 2011), it is notable that it found some indications of improved mood. With so little research on the use of texting in crisis intervention services, there is much to be gained from continued examination of this modality. Currently, this is not published data demonstrating how crisis counseling unfolds in a texting context or to what extent it is helpful or that users are satisfied with it. It is especially critical to understand the best counseling strategies in this context and to what extent users find these useful. The Present Study While crisis intervention programs routinely track statistics on presenting problems and categories of counselor actions, the analysis here contributes a perspective not well- studied nor understood - that of the texter. The anonymous nature of crisis lines makes it extraordinarily difficult to learn if the counseling session was helpful as there is not an Nesmith/ REACHING YOUNG PEOPLE 1151 avenue for gathering follow-up information; therefore this study sought clues within the session transcripts. This qualitative study collaborated with a human services agency, Canvas Health, to answer two questions: 1) In what ways do those using texting crisis services communicate whether or not their needs have been addressed, at least for the short- term?; and 2) How does the texting interface affect the counseling approach to assessing and connecting with service users? Partnering Agency and Texting Crisis Program Canvas Health is a comprehensive integrated human service center that provides a wide range of mental and physical health services in several Minnesota locations. This project worked with one of the agency’s programs, TXT4Life, which offers texting-based suicide prevention counseling services. Their crisis line services aim to provide short-term help which can include (but is not limited to) establishing a safety plan to prevent self-harm, directing texters to longer-term local sources of support, or reducing the intensity of their emotional state. TXT4Life serves the entire state, though is accessible as a toll-free number to anyone in the U.S., and uses HIPAA-compliant technology to offer confidential services with a trained counselor. Various outcomes are tracked based on provided services and concrete action steps. Text-based counseling quickly became popular. In 2014 when data for this study was collected, the program received 5,658 text calls, a figure that jumped to 9,968 texts one year later (TXT4Life, 2015). At the time, the agency was funded to provide texting services through midnight and again at 8:00am in the morning. Those who texted during closed hours were offered the opportunity to use the 24-hour voice hotline. If they did not choose that alternative, their texting attempt was tracked and a counselor contacted them in the morning to offer support if they still wanted it. The agency now offers full 24-hour support for their texting line. In 2014, the program assessed all the text sessions of any length including those with non-crisis needs. The data revealed that 76% of the sessions ended with positive action steps, the largest share of which included personal care actions (18%) and internal coping strategies (17%; TXT4Life, 2015). In an effort to exclude sessions in which the texter only sought referral information, the current study selected sessions lasting at least 30 minutes to capture individuals with more complicated or serious concerns than the full spectrum of texter contacts. Therefore, the data in this study likely reflect a narrower troubled group than the total population of texters in this program. While the agency did not report information about how many text sessions included only information or referral requests, they did report that 70% included supportive listening, which likely referred to the types of sessions included in this study. Methods This study is primarily qualitative with some descriptive statistics, using counseling session transcripts to understand how texters expressed their needs and how both counselors and texters communicated via texting. The sample size and analytic methods are in keeping with qualitative approaches (Corbin & Strauss, 2008; Creswell, Hanson, Clark, & Morales, 2007). ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1152 Sampling and Data Collection Sampling Frame. Text sessions were drawn from 2014 data. To capture sessions that were most likely to include a counseling exchange, only text sessions lasting 30 minutes or longer were included. This duration was selected based on the agency’s annual data which showed that the conversation density (number of words) in a text session was comparable to about one-sixth that of a voice session of the same length (TXT4Life, 2015). For example, a typical 10-minute voice session would take about an hour in texting to exchange a similar amount of content. Therefore, text sessions shorter than 30 minutes were similar to voice sessions less than five minutes and were likely to be requests for information. Sampling. Forty-eight text crisis counseling session transcripts were randomly selected. To ensure a representative sample throughout the course of the year, four randomly selected dates from each month in 2014 were used to select counseling sessions. For privacy reasons, the program supervisor extracted the text transcripts from each random date and redacted all identifying information prior to analysis. Key Informant Interviews. Because there is so little published information about the counseling process using a texting interface, key informant in-person interviews with the program’s two counseling supervisors were used to provide additional context to the transcript findings. These interviews were not intended in any way to serve as a broadly representative voice on counselor experiences. Rather, they augment the main analysis of counseling transcripts by presenting the rationale for certain techniques used and to offer insight into how the texting format influenced the counseling process in this particular program. These were the only two staff at the agency who had experience with both voice and text lines and could provide a unique perspective about them. As supervisors, they monitored counseling sessions for both types of crisis lines and trained new counselors for the text line. Analysis Demographic information was gathered from the text sessions to the extent possible though it was optional, and most declined. The entire transcript from each session was examined, with particular attention to their presenting problems and the texter language and tone as the sessions closed. Ascertaining outcomes is difficult with an anonymous short-term service. Without directly-asked questions in a follow-up survey, it is more an assessment of hints and clues that may reveal whether users felt their needs were addressed, at least for the short-term. The counselor approaches to connect with the texter and the texter reactions were examined to help understand how those interactions unfold in a texting interface. To accomplish this, the analysis used an iterative process of reading, identifying patterns, negative case analysis (examples where a theme emerged differently), and constant comparison of themes (Corbin & Strauss, 2008; Creswell et al., 2007; Miles, Huberman, & Saldana, 2014). Counselor interviews provided information about counseling and intervention techniques used that were specific to the texting format. Drawing from a mixed methods approach and using an explanatory sequential design (Haight & Bidwell, 2016), the Nesmith/ REACHING YOUNG PEOPLE 1153 specific counseling supervisor interview questions were developed after the text sessions were analyzed to elucidate features in need of further clarification. With only two interviews and direct questions, the analysis did not involve an in-depth analysis of patterns and themes. Rather the findings from those interviews provide information about the intervention strategies used and observations about the impact of text counseling. Ethical Research with Human Subjects The research protocol for this study was approved by the principal investigator’s university institutional review board in advance of any data collection. An anonymous service by design, in most cases the staff also did not know the identity of the texter. Texters were not required to provide names or any demographic information. In the instances where there was such data in the transcripts, the supervisor fully redacted all identifying information in the transcripts prior to making the data accessible for analysis. Findings The findings present data from both the text transcripts and the counseling supervisor interviews. For each section, the findings from the transcript analysis are presented, followed by counseling supervisor perspectives on that topic. Texter quotes are presented as received; any spelling errors or abbreviations reflect how they appeared in the original transcript. Obtained Sample Forty-nine randomly selected transcripts representing all 12 months of 2014 were used. One texter initiated contact later in the same day and was included in the sample, increasing it from 48 to 49. The sample therefore represents up to 48 unique individuals, though given anonymity, there may be repeat texters who were not identified as such. Of those who reported gender, 29 were female, eight male, and one transgender. Two-thirds opted not to provide their age. Of those who did, the vast majority were between 12 and 24 years old, averaging 18.5 years. There was one outlier who was 44 years old. Most of the texters did not opt to report their ethnicity. Of the ten who did, two identified as African American, one as Hispanic, and the remaining were Caucasian. Seventeen different counselors facilitated the text sessions that were reviewed, averaging three text sessions per counselor. Presenting Problems The texters sought help in three broad areas: 1) their current emotional state reported as part of a pre-existing mental health condition causing chronic bouts of depression, anxiety, or suicidality, 2) a romantic interest yielding rejection, conflict, or abuse, or 3) family and peer relations, typically in regard to feeling isolated, unwanted, or unloved by family, or bullying by peers. Disinhibition with Texting Both the text transcripts and the counseling supervisors revealed some indication of disinhibition linked to the anonymity of the texting interface. ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1154 Texters. There was some evidence to suggest that texters would elect not to receive crisis services if they were not offered via texting. Nearly a third (29.2%) of the texters originally texted during the night when the text line was not available. While they were given the option to use the voice line to receive immediate help, none chose that option. Instead they waited until morning to text again or accepted support when a counselor texted them. As texters began sharing their reasons for seeking help, some explained that they chose texting because it was too difficult to talk to others in-person or over the phone about their concerns. Below is an excerpt from a young person sharing for the first time her longstanding problem with self-harm and eating disorders. Texter: Well, I want to get help somehow but I'm too scared to tell my parents or any one at school. Counselor: Thank you for reaching out. I know it takes a lot of courage to speak up. Could you tell what you're going through? Maybe talking about it might help. Texter: I've been self-harming and I've had an eating disorder for years and I feel like it'll never go away. Counselor: I'm sorry to hear that. It must be very hard to pretend and hide this from your family. How have you been coping? Texter: I don't know, it's kinda up and down I guess. Counselor: Are you receiving any services for your eating disorder and depression? Texter: No. I've never really told anyone about it. Another texter explained that she had not talked to anyone about her suicidal ideations and that she could not use voice services. Counselor: who do you have that provides support? Texter: i don't really talk to anyone about this so no one knows Texter: i was looking for suicide hotlines online but since i can't really speak over the phone i decided text message would be best The texters sought help from a wide range of locations and situations. Although not all texters shared their location, many did. While some were at home, especially during evening sessions, others mentioned that they were in class, in the school restroom, in the school cafeteria, at work, or on a walk. Counseling Supervisors. The counseling supervisors also observed that the privacy of texting ameliorated inhibitions. The anonymity really helps. People are honest very quickly, in our experience. As opposed to face-to-face or the [voice-based] calls who are less so. Nesmith/ REACHING YOUNG PEOPLE 1155 One of the supervisors discussed the relevance in situations warranting mandated reports. The counselor explained that youth who were otherwise too afraid to report on their abuse and name their perpetrator in-person or by voice, were able to do it via texting. Kids, you know, a 14 year old…who’s been a survivor of abuse – I think that it would be very hard for them to make the report on the phone or in person…like at school those counselors might know their parents or know their uncle or might know the perpetrator. Whereas with us…you don’t have to look at us. You don’t have to hear us…so we’re telling them we’re mandated reporters upfront, making it their choice, then they’re willing to give us more information than they would otherwise. Then we’ll get, “here’s my last thing, here’s the person in question. This is perpetrator.” After we started doing this, we experienced a spike over the last year in mandated reports due to kids sharing enough information to report. There were not examples of such abuse reports in the obtained sample of transcripts. However this comment was included here because, if it is reflective of a real phenomenon, it is an important issue worthy of future exploration. Text Session Closings Text statements were examined at or near the end of the sessions to identify two indicators of outcomes that might shed light on whether or not their needs were addressed: 1) if they reached a plan that they were willing to try, and 2) if their remarks offered any indication of whether or not they found the session to be helpful. The nature of the data is biased toward an absence of positive findings because the only way to know with certainty that a texter found it useful is if they explicitly volunteered that information. Texters who offered no remarks about the services may have found them helpful, but simply did not state it; the converse could be true as well. Four themes from the close of the conversations surfaced (see list below) . These were categorized by the texter remarks indicating the following: whether or not a plan for safety was established, or in cases where safety was not a concern, if at least a tentative plan for the presenting problem was achieved, and who initiated the closure. We also examined whether or not it was a jointly reached plan. 1. Positive: (65%) a) the texter specifically indicated some satisfaction with the session toward addressing their needs, a plan was reached, and the session was jointly closed by texter and counselor; or b) the texter language was cautiously optimistic, a plan was reached, and the counselor initiated the close. 2. Not in danger/no indication about whether their needs were addressed: (17%) The texter was not currently in danger. The texter did not provide language suggesting satisfaction or dissatisfaction with session. The counselor initiated the close. No plan regarding the initial presenting problem was created. 3. Abrupt: (14%) The texter initiated the close abruptly without explanation. 4. Vague: (4%) The texter initiated the close, offering a vague or generic reason. Or the texter’s language in response to a plan was noncommittal. ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1156 Positive Closings. Nearly two-thirds (65%) of the sessions ended positively. Texters were not asked to report their satisfaction with the sessions in regard to meeting their needs, so it is not known definitively how many found the services helpful. This section describes both those who explicitly expressed satisfaction and those who might have found the services helpful but revealed it in indirect ways. Of all the types of closures, a positive collaborative end to the text session was the most common. In these cases, there was some resolution with either a short-term plan to get through the day or night or a more extended plan, jointly reached by texter and counselor.  I’m sure [going for a run] will help me feel better about things. Thanks for listening. I will text in tomorrow if I need to.  You've convinced me to reach out more for help, thank you!  Thank you for your help. I feel better having a plan for what I'm going to do  I'm glad I reached out to you im beginning to feel a lil better I've calmed down  Thank you I feel a lot more comfortable after talking to you…  Thank you so so so much  Ok thank you! And yes, thank you for talking to me! A texter who reaches out in crisis can have a positive outcome without necessarily sounding upbeat. Some of the texters agreed to a plan using language that was cautiously optimistic. In the exchange below, the texter agreed to accept several referrals. The texter’s phrasing, “I think that will work” suggests that he or she was at least willing to try the plan. After this remark, the “okay” comments exude neither a positive or negative sentiment. They demonstrate that the texter was paying attention and accepted the referral information, though there was a significant lapse before acknowledging them, a common occurrence in the texting venue. Counselor: So I found several counselors who specialize in self-harming. They accept Sliding Scale so you may still be charged but in a less amount. Will that work for you? Counselor: [10 minutes later] Are you still there? Texter: Yes, i am still here. I think that will work. Counselor: Great! Here are the two counselor's name and address. Texter: okay Counselor: Also, here is a clinic that seems provide free services Texter: okay Counselor: I will let you go now. I hope you have a good night! :) Not in danger, no indication if needs were addressed. Seventeen percent of the sessions fit this category. These were often texters seeking to talk about their concerns, but they indicated no specific need beyond support. Therefore unlike other sessions, counselors did not attempt to close the conversation with an offer of service referrals. In these sessions, Nesmith/ REACHING YOUNG PEOPLE 1157 the texters also did not offer any information whether they found the session to be helpful or unhelpful. Abrupt Ending. In the sessions classified as having an abrupt ending (14% of sessions), the texter initiated the close suddenly without a plan and without any indication of whether or not the session was helpful. I have gym class so i have to go I have to go. My husband is home Other examples of abrupt endings occurred when there was a long lapse in texter response. Sometimes the counselor would ask if the texter still wanted to talk or if this was still a good time to talk and the texter replied “no.” Other times they simply disappeared. One used the “STOP” option offered to all texters to immediately close the session. While it is never known with certainty, some abrupt endings may be due to external factors that compel a texter to close quickly, such as someone walking in the room, or as one young texter reported, they were texting at school between classes and class was about to start. Others may not want to continue the session and use these statements as a way to close the conversation. Vague/Ambiguous. The closings coded as vague were the smallest category (4%). Here, the texter initiated the close, offering a generic reason, or the language in response to help was noncommittal. In these cases, the texter responses were increasingly delayed. They most often stated that they were tired and wanted to sleep, regardless of the time of day. I think I might take a nap for a while. Im feeling very sleepy can we talk tomorrow? In these cases, the texter typically did not respond to specific counselor suggestions, or did not respond to counselor efforts to identify possible plans. Counseling Strategies – Counseling Supervisor Perspectives The counseling supervisors were asked to explain the program’s intervention strategies, challenges, and strengths when using the texting interface. Both observed that assessment and connection, each critical to the process, unfolded differently with text than with voice. Assessment must occur quickly in crisis intervention and was challenging without sound-based cues. On the phone you’re picking up different cues. Background noises, inflection, their tone, are they emotional? … you’re assessing based on the non-verbal cues you’re getting. Are they outside or inside? Are they pacing? Are they crying? With text we don’t get any of that information. Therefore, counselors had to explicitly ask questions they might otherwise detect in voice calls. The supervisors explained that new counselors were tempted to attach emotions or emphasis to the typed words and highlighted the importance of careful training. ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1158 What happens especially with new staff who don’t have any experience in the field is that they want to project some emotional situation to the text…They want to emphasize words…but you only have the words on the screen. That’s it. That’s all you have. You cannot under any circumstances put any weight on any of those words because you can’t infer inflection. Unless they capitalize it. Because there’s no bold. There’s no italics. There’s not anything else. The supervisors offered some insight into the unique effort and tactics required to cultivate a connection with texters with only the written word. Both emphasized that they must be more direct and specific in their language than with voice conversations. One explained, Everything that I want to convey, I have to convey through words…I have to make a more conscious effort to connect. Whereas on the call side you can connect just through your tone of voice, being there, active listening and they know someone is listening. On the text side, you have to prove that you are listening…I have to kind of verbalize my active listening a lot…and we have to actively sound empathetic, because by nature, text messages do not sound empathetic. Both counseling supervisors also explained that it was necessary to provide a single question or statement per text and wait for a response before moving on to another question or else risk that the texts become off-sequence, making it difficult to know which question the texter was responding to. In the following excerpt, the counselor made explicit empathic statements that could perhaps be delivered through vocal prompts (e.g., “mmhm”) in a voice call. The counselor kept the comments brief, one idea per text, to keep the texter engaged. Short counselor texts also mirrored the typically short texter messages. The vast majority of texter messages across all texters were single sentences or phrases. Near the end of this excerpt, the counselor touched on an important precipitating event that the texter started to deny but then revealed in the pursuant exchange. Texter: I just want to talk. No I'm not in danger Counselor: What did you want to talk about? Texter: I'm falling apart and I feel so worthless Counselor: That sounds overwhelming. Counselor: I want to check in with you and ask; are you thinking about suicide? Texter: I think about it every day.. I won't do anything though Counselor: It must be hard to feel that way all the time. I'm glad to hear that you won't do anything though. Counselor: Why do you feel that you want to die? Texter: I have people I would hurt... I feel like shit. I feel unwanted and worth less. My boyfriend never helps the situation Counselor: I'm sorry to hear your boyfriend hasn't been helpful -Is there anyone who is helpful? Nesmith/ REACHING YOUNG PEOPLE 1159 Texter: Not really. I feel like everyone hates me Counselor: Did something happen to make you feel this way? Texter: No… maybe idk* *idk is an abbreviation for “I don’t know” Despite the challenges of texting-based counseling, the supervisors both highlighted the value of reaching people who might not otherwise seek help. One noted that when new staff were trained in both crisis line formats, they commonly asked if text lines are as effective as voice calls. My answer to that is, no, but that’s not the point. People are using this service who would not otherwise use it. Like, they are not looking at it like, “I’m going to talk to my school counselor or text TXT4Life.” They’re looking at it only like, “I’m going to text” because they’re looking for something more anonymous, no face- to-face or voice. They want the privacy piece and impersonal as possible…The point is we’re getting people to reach out that wouldn’t normally reach out. The question therefore, according to the supervisors, was not so much one of whether or not texting-based crisis counseling is equivalent to voice-based hotlines, but rather whether or not the texter would receive any help at all without the text service. Limitations The sampling frame in this study excluded sessions shorter than 30 minutes, thereby eliminating texters seeking only referrals and those who were not in crisis but needed a brief check-in, or other more straightforward needs. As such, this study sample was disproportionately comprised of those with more complex or more serious problems. It is important, therefore, to use caution when comparing these findings to those typically presented by other text lines that report on all text sessions. The anonymous nature of crisis lines limits access to information like demographics and outcomes post-session, both important to fully understanding the scope and needs of text line users. The key informant interviews can only speak to the techniques and approaches specific to the collaborating agency and should not be interpreted more broadly. Discussion This study highlights benefits and challenges of texting-based crisis lines, as well as nuances of the counseling process. Because it is an emerging medium for service delivery, most of the available information is found in program reports that are not peer-reviewed research and primarily provide descriptive statistics about users and services delivered. Very little empirical information is available on texter reactions to counseling sessions or counseling strategies with texting. There is growing support for the value of texting-based crisis counseling. The texting interface is popular, especially among young users. This is evidenced in part by the growing volume of text calls. The program saw a six-fold increase in the year following the initial year of operation and has seen dramatic monthly increases in text volume since ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1160 then (Canvas Health, 2015). In 2014, they received 5,658 texts and one year later in 2015, this figured jumped to nearly 10,000 texts (Canvas Health, 2015). The largest such service in the U.S., the Crisis Text Line, reports vast usage of their line since its inception in 2013 (Crisis Trends, 2016). This sheer usage suggests that young people view this as a viable modality to receive needed support. The counselors in this study argued that without texting, some youth in crisis may not otherwise seek help, a point also raised by several of the system users. The overall agency statistics and those from other research reveal that voice hotlines are underutilized by adolescents and young adults, whereas the text users disproportionately represent this same age group (Gould et al., 2006; Canvas Health, 2011; TXT4Life, 2015). Other research supports this phenomenon as well. For example, Evans et al. (2013) reported that their text users rejected “outside help,” instead requesting more text options and that youth believed the text line would “help youth who might not otherwise seek help” (p. 480). Gibson and colleagues (2016) also arrived at this conclusion, reporting that adolescents using texting services explicitly stated their preference for this mode over all others. The findings here offer some support of the Online Disinhibition Effect (Suler, 2004) as an explanation for the popularity of texting services. Some of the texters in this study reported that it was the texting-based format that encouraged them to use the service, often sharing their struggles for the first time via text. This is consistent with Haxel’s (2013) findings as well. Both this study and Haxel’s work suggest that texting creates room for dissociative anonymity (Suler, 2004) in which the texter has control over information and verbal expression, allowing them to communicate more candidly than in-person, if at all. About two-thirds (65%) of the texters in this study gave some indication of a positive outcome. This is a larger share than Mokkenstorm et al. (2017) found (49%) with chat sessions, though the comparison should be viewed with caution as the measures were different. This still leaves a third who did not explicitly suggest a positive outcome. These findings are tempered by the scope of anonymous crisis counseling. While counselors could do their best to reach a shared agreement of a safety plan, just as with any counseling situation, not all texters are ready for this, even for the short-term to get through the day. Others, not in immediate danger, may have chronic presenting problems that cannot be reasonably resolved, even temporarily, during a text session. A higher rate of positive closings (76%) were reported when including non-crisis calls seeking information or other quick responses, all excluded from this analysis (TXT4Life, 2015). This is in keeping with at least one other texting program (Evans et al., 2013). One of the greatest challenges for texting-based crisis counseling is assessing effectiveness. This is extraordinarily difficult in the context of anonymous crisis intervention, particularly determining impact after the session has ended. There is inherent bias in regard to who spontaneously provides feedback at a later date (Mokkenstorm et al., 2017). Post-session surveys also raise problematic issues with non-response, as those with some of the worst outcomes such as suicide, will, by definition, not be reachable. Yet, mid- session may not be the time that a texter thinks about saying “thank you” or sharing that it is making a difference. In fact, the nature of texting tends to omit this type of exchange. Therefore, the absence of affirmative statements toward usefulness does not mean that a Nesmith/ REACHING YOUNG PEOPLE 1161 session was unhelpful, only that it was not expressed at the time. Nevertheless, there was some evidence of user satisfaction in the text transcripts; the majority of the texters provided some indication that the services were useful. The flexibility of texting is an important factor. Unlike online chats or other telecommunication services, texters may reach out for help in any setting while retaining privacy from others in close proximity, freeing them to seek help at almost any time rather than waiting for an opportunity and space to have a private phone call. In fact, the texters in this study sought help while at work, during school, meals, and in the presence of friends and family, a feature also observed in Haxell’s (2013) study. The benefit of immediacy is critical; texters do not have to delay seeking help until there is a private time or space. There are many similarities but also important distinctions between voice and text crisis counseling. The goals of crisis intervention - to assess, deescalate, establish a safety plan, provide support, and offer referrals - are the same whether the mode is texting or voice. Just as with voice calls, counselors aim to connect with texters, reflect their feelings, gather information, and help them arrive at a short-term plan when necessary. The texters in this sample raised the same types and range of presenting problems as in voice-based calls. Nevertheless, several critical differences between the two modes are important to understand. These differences lie in the way that texters share information, the flow of the sessions, and knowledge and techniques that counselors must have in order to engage effectively with the texters. Without tone of voice or other sound-based cues, a counselor cannot know texter emotions until they are stated. In voice calls, counselors attend to cues from the caller’s tone, rapidity of speech, sounds of crying, and background noises. They also use their own vocal tone to influence interactions. Without those conversational attributes in texting, questions must be framed more carefully and directly. Exchanges must be brief, attending to a single idea per text, lest the exchange becomes off-sync and leads to confusion about which question or comment the response refers to. The counseling supervisors who provided context for this study emphasized that interpretation of texts must be drawn exclusively from typed words, avoiding the temptation to assign emotions to written statements when none have been expressed to avoid arriving at an inaccurate conclusion or assumption. Ideally, a training manual would be created and made available for programs offering texting as a new service. It is possible that programs that have been in operation and have developed effective techniques specific to texting already have such manuals that could be modified for more general use. Specialized training is important for any new programs or counselors new to this venue for crisis counseling. Training should be comprehensive, but some key components should include assessing by asking critical questions directly, practicing not placing emphasis on certain words, and establishing a connection quickly through brief empathic statements delivered one at a time. The written nature of the texting interface has some advantages for supervision and training. Examples of past sessions can be easily reviewed to demonstrate counseling strategies that work as well as those that are less effective. Real-time supervision can be provided with a supervisor reading texts side-by-side with a new counselor, talking them ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1162 through sessions. With asynchronicity allowing for short delays, counselor and supervisor can discuss next steps together during the session. Finally, research must catch up to the pace of technological changes in crisis counseling. Very little is known about best practices, about the most effective ways to establish a connection with texters, or about what the system users value the most in texting-based crisis services. More research is needed to better assess outcomes and to examine the efficacy of various communication and intervention strategies. The young demographic of text line users represents the most vulnerable and most underserved in crisis counseling (Santor et al., 2007; Sawyer et al., 2012). Text-based crisis counseling is cutting edge and germane in this age of electronic communication and shows promise for young people suffering from mental health and other socio-emotional problems who are not ready for voice-based or in-person services. References Bowers, H., Manion, I., Papadopoulos, D., & Gauvreau E. (2013). Child and Adolescent Mental Health, 18(3), 165-170. doi: https://doi.org/10.1111/j.1475- 3588.2012.00673.x The Bridge for Youth. (n.d.). TEXT4 help. Text SAFE. Minneapolis, MN. Retrieved from http://www.bridgeforyouth.org/contact/ Canvas Health. (2011). Annual report. Minneapolis, Minnesota: Canvas Health. Centers for Disease Control. (2014). Suicide prevention: Youth suicide. Retrieved from www.cdc.gov/violenceprevention/pub/youth_suicide.html Corbin, J., & Strauss, A. (2008). Basics of qualitative research. Thousand Oaks, CA: Sage. Creswell, J. W., Hanson, W. E., Clark, V. L. P., & Morales, A. (2007). Qualitative research designs: Selection and implementation. The Counseling Psychologist, 35(2), 236-264. doi: https://doi.org/10.1177/0011000006287390 Crisis Text Line. (2018). How does crisis text line work? [Text HOME to 741741]. Retrieved from https://www.crisistextline.org/ Crisis Trends. (2016, December). Crisis Text Line. Retrieved from www.crisistrends.org Dubus, N. (2015). Texting: The third client in the room. Clinical Social Work Journal, 43, 209-214. doi: https://doi.org/10.1007/s10615-014-0504-3 Evans, W. P., Davidson, L., & Sicafuse, L. (2013). Someone to listen: Increasing youth help-seeking behavior through a text-based crisis line for youth. Journal of Community Psychology, 41(4), 471-487. doi: https://doi.org/10.1002/jcop.21551 Gibson, K., Cartwright, C., Kerrisk, K., Campbell, J., & Seymour, F. (2016). What young people want: A qualitative study of adolescents’ priorities for engagement across psychological services. Journal of Child and Families Studies, 25, 1057-1065. doi: https://doi.org/10.1007/s10826-015-0292-6 https://doi.org/10.1111/j.1475-3588.2012.00673.x https://doi.org/10.1111/j.1475-3588.2012.00673.x http://www.bridgeforyouth.org/contact/ http://www.cdc.gov/violenceprevention/pub/youth_suicide.html https://doi.org/10.1177/0011000006287390 https://www.crisistextline.org/ http://www.crisistrends.org/ https://doi.org/10.1007/s10615-014-0504-3 https://doi.org/10.1002/jcop.21551 https://doi.org/10.1007/s10826-015-0292-6 Nesmith/ REACHING YOUNG PEOPLE 1163 Gould, M. S., Greenberg, T., Munfakh, J. L., Kleinman, M., & Lubell, K. (2006). Teenagers’ Attitudes about seeking help from telephone crisis services. Suicide and Life-Threatening Behavior, 36(6), 601-613. doi: https://doi.org/10.1521/suli.2006.36.6.601 Haight, W. L., & Bidwell, L. N. (2016). Mixed methods research for social work: Integrating methodologies to strengthen practice and policy. Chicago, IL: Lyceum. Haxell, A. J. (2013). On becoming textually active at Youthline, New Zealand. British Journal of Guidance & Counselling, 43(1), 144-155. doi: https://doi.org/10.1080/03069885.2014.922163 Kaufman, L. (2014, February 4). In texting era, crisis hotlines put help at youths’ fingertips. New York Times. Retrieved from http://www.nytimes.com Kitchingman, T. A., Wilson, C. J., Caputi, P., Woodward, A., & Hunt, T. (2015). Development and evaluation of the Telephone Crisis Support Skills Scale. Crisis: The Journal of Crisis Intervention and Suicide Prevention, 36(6), 407-415. doi: https://doi.org/10.1027/0227-5910/a000341 Koltz, R. L., & Tarabochia, D. S. (2014). Technology: Using it as a means to creatively connect with adolescents. Journal of Creativity in Mental Health, 9, 380-398. doi: https://doi.org/10.1080/15401383.2014.899481 Krysinska, K. E., & DeLeo, D. (2007). Telecommunication and suicide prevention: Hopes and challenges for the new century. OMEGA, 55(3), 237-253. doi: https://doi.org/10.2190/OM.55.3.e Lenhart, A., Smith, A., Anderson, M., Duggan, M., & Perrin, A. (2015, August). Teens, Technology and Friendships. Pew Research Center. http://assets.pewresearch.org/wp- content/uploads/sites/14/2015/08/Teens-and-Friendships-FINAL2.pdf Lipmann, A. D. P. (2010). The current status and impact of US national policy on the help-seeking behavior of adolescents. Children and Youth Services Review, 32, 822- 828. doi: https://doi.org/10.1016/j.childyouth.2010.01.020 Martin, S., Sutcliffe, P., Griffiths, F., Sturt, J., Powell, J., Adams, A., & Dale, J. (2011). Effectiveness and impact of networked communication interventions in young people with mental health conditions: A systematic review. Patient Education and Counseling, 85, e108-e119. doi: https://doi.org/10.1016/j.pec.2010.11.014 Miles, M. B., Huberman, M. A., & Saldana, J. (2014). Qualitative data analysis: A methods sourcebook. Los Angeles: Sage. Mokkenstorm, J. K., Eikelenboom, M., Huisman, A., Wiebenga, J., Gilessen, R., Kerkhof, J. F. M., & Smit, J. H. (2017). Evaluation of the 113 online suicide prevention crisis chat service: Outcomes, helper behaviors and comparison to telephone hotlines. Suicide and Life-Threatening Behavior, 47(3), 282-296. doi: https://doi.org/10.1111/sltb.12286 https://doi.org/10.1521/suli.2006.36.6.601 https://doi.org/10.1080/03069885.2014.922163 http://www.nytimes.com/ https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Kitchingman,+Taneile+A./$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Wilson,+Coralie+J./$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Caputi,+Peter/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Woodward,+Alan/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Hunt,+Tara/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/pubidlinkhandler/sng/pubtitle/Crisis:+The+Journal+of+Crisis+Intervention+and+Suicide+Prevention/$N/60904/DocView/1749218385/fulltextwithgraphics/210452BCB8E048AEPQ/8?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/pubidlinkhandler/sng/pubtitle/Crisis:+The+Journal+of+Crisis+Intervention+and+Suicide+Prevention/$N/60904/DocView/1749218385/fulltextwithgraphics/210452BCB8E048AEPQ/8?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexingvolumeissuelinkhandler/60904/Crisis:+The+Journal+of+Crisis+Intervention+and+Suicide+Prevention/02015Y01Y01$232015$3b++Vol.+36+$286$29/36/6?accountid=14756 https://doi.org/10.1080/15401383.2014.899481 https://doi.org/10.2190/OM.55.3.e http://assets.pewresearch.org/wp-content/uploads/sites/14/2015/08/Teens-and-Friendships-FINAL2.pdf http://assets.pewresearch.org/wp-content/uploads/sites/14/2015/08/Teens-and-Friendships-FINAL2.pdf https://doi.org/10.1016/j.childyouth.2010.01.020 https://doi.org/10.1016/j.pec.2010.11.014 https://doi.org/10.1111/sltb.12286 ADVANCES IN SOCIAL WORK, Fall 2018, 18(4) 1164 Newport, F. (2014, November). The new era of communication among Americans. Retrieved from http://www.gallup.com/poll/179288/new-era-communication- americans.aspx Ramchand, R.; Jaycox, L.; Ebener, P; Gilbert, M; Barnes-Proby, D., & Goutam, P. (2017). Characteristics and proximal outcomes of calls made to suicide crisis hotlines in California: Variability across centers, Crisis: The Journal of Crisis Intervention and Suicide Prevention, 38(1), 26-35. doi: https://doi.org/10.1027/0227- 5910/a000401 Safe Place. (n.d.). Txt 4 Help. Chicago, IL. Retrieved from http://nationalsafeplace.org/safe-place-teens/ Santor, D., Poulin, C., LeBlanc, J., & Kusumakar, V. (2007). Facilitating help seeking behavior and referrals for mental health difficulties in school aged boys and girls: A school-based intervention. Journal of Youth & Adolescence, 36(6), 741-752. doi: doi: https://doi.org/10.1007/s10964-006-9092-z Saunders, S. M., Resnick, M. D., Hoberman, H. M., & Blum, R. W. (1994). Formal help- seeking behavior of adolescents identifying themselves as having mental health problems. Journal of the American Academy of Child & Adolescent Psychiatry, 33(5), 718-728. doi: https://doi.org/10.1097/00004583-199406000-00013 Sawyer, M. G., Sawyer, A. C. P., & La Greca, A. M. (2012). Help-seeking by young people with depressive disorders. Prevention Researcher, 19(4), 11-14. Stevens, B.A. & Ellerbrock, L.S. (1995). Crisis intervention: An opportunity to change. ERIC Digest EDO-CG-95-34 . Retrieved from https://www.counseling.org/resources/library/ERIC%20Digests/95-034.pdf Substance Abuse and Mental Health Services Administration. (n.d.). National suicide prevention lifeline [1-800-273-8255]. Retrieved from https://suicidepreventionlifeline.org/ Suler, J. (2004). The online disinhibition effect. Cyberpsychology & Behavior, 7(3), 321- 325. doi: https://doi.org/10.1089/1094931041291295 Teen Line. (n.d.). Teen Line. Text TEEN. Los Angeles, CA. Retrieved from https://teenlineonline.org/ TXT4Life. (2013). Annual progress report for 2012. Minneapolis, MN: Canvas Health. TXT4Life. (2015). Statewide expansion utilization report. Minneapolis, MN: Canvas Health. Author note: Address correspondence to: Ande Nesmith, Ph.D., LISW, is an Associate Profesor, School of Social Work Mail SCB 201, University of St. Thomas, St. Paul, MN 55105. nesm3326@stthomas.edu http://www.gallup.com/poll/179288/new-era-communication-americans.aspx http://www.gallup.com/poll/179288/new-era-communication-americans.aspx https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Jaycox,+Lisa/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Ebener,+Pat/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Gilbert,+Mary+Lou/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexinglinkhandler/sng/au/Barnes-Proby,+Dionne/$N?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/pubidlinkhandler/sng/pubtitle/Crisis:+The+Journal+of+Crisis+Intervention+and+Suicide+Prevention/$N/60904/DocView/1800417791/fulltextwithgraphics/210452BCB8E048AEPQ/14?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/pubidlinkhandler/sng/pubtitle/Crisis:+The+Journal+of+Crisis+Intervention+and+Suicide+Prevention/$N/60904/DocView/1800417791/fulltextwithgraphics/210452BCB8E048AEPQ/14?accountid=14756 https://search-proquest-com.ezproxy.stthomas.edu/indexingvolumeissuelinkhandler/60904/Crisis:+The+Journal+of+Crisis+Intervention+and+Suicide+Prevention/02017Y01Y01$232017$3b++Vol.+38+$281$29/38/1?accountid=14756 https://doi.org/10.1027/0227-5910/a000401 https://doi.org/10.1027/0227-5910/a000401 http://nationalsafeplace.org/safe-place-teens/ https://doi.org/10.1007/s10964-006-9092-z https://doi.org/10.1097/00004583-199406000-00013 https://www.counseling.org/resources/library/ERIC%20Digests/95-034.pdf https://www.counseling.org/resources/library/ERIC%20Digests/95-034.pdf https://suicidepreventionlifeline.org/ https://doi.org/10.1089/1094931041291295 https://teenlineonline.org/ mailto:nesm3326@stthomas.edu