This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mhealth and uhealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Youth homelessness is a substantial issue, and many youths experiencing homelessness have mental health issues as both a cause and consequence of homelessness. These youths face many barriers to receiving traditional mental health services, and as a result, only a few youths experiencing homelessness receive any form of mental health care.
This project aimed to develop and determine the feasibility and acceptability of engaging young adults (ie, individuals aged 18-24 years) experiencing homelessness in a remotely delivered mental health intervention. This intervention provided brief emotional support and coping skills, drawing from cognitive behavioral principles as an introduction into psychosocial support. The intervention was piloted in a homeless shelter network.
A total of 35 young adults experiencing homelessness participated in a single-arm feasibility pilot trial. Participants received a mobile phone, a service and data plan, and 1 month of support from a coach consisting of up to 3 brief phone sessions, text messaging, and mobile mental health apps. We evaluated feasibility by looking at completion of sessions as well as the overall program and acceptability with satisfaction ratings. We also collected clinical symptoms at baseline and the end of the 1-month support period. We used validity items to identify participants who might be responding inappropriately and thus only report satisfaction ratings and clinical outcomes from valid responses.
Most participants (20/35, 57%) completed all 3 of their phone sessions, with an average of 2.09 sessions (SD 1.22) completed by each participant. Participants sent an average of 15.06 text messages (SD 12.62) and received an average of 19.34 messages (SD 12.70). We found higher rates of satisfaction among the participants with valid responses, with 100% (23/23) of such participants indicating that they would recommend participation to someone else and 52% (12/23) reporting that they were
This study demonstrated that it was feasible to engage homeless young adults in mental health services in this technology-based intervention with high rates of satisfaction. We did not find changes in clinical outcomes; however, we had a small sample size and a brief intervention. Technology might be an important avenue to reach young adults experiencing homelessness, but additional work could explore proper interventions to deliver with such a platform.
ClinicalTrials.gov NCT03620682; https://clinicaltrials.gov/ct2/show/NCT03620682
Youth homelessness is a substantial issue. In 2018, 36,361 youths, that is, those younger than 25 years, experienced homelessness on a single night [
As such, traditional mental health services may be ineffective at reaching and addressing the needs of this population. One complement to traditional mental health services that could help overcome some of these barriers would be technology-based treatment options [
Thus, technology-based resources have the potential to be useful for homeless youth. A first question is what types of resources might yield themselves to technology-based interventions in this population. A few studies have suggested that cognitive behavioral therapy (CBT) delivered in shelters can lead to significant decreases in depression and other mental health problems and improvements in self-efficacy [
Notably, CBT principles are amenable to delivery both in very brief formats [
In this study, we evaluated the feasibility, acceptability, and preliminary benefits of a multicomponent mobile phone–based intervention for homeless young adults. Although youth refers to those younger than 25 years, we only engaged those in the age group of 18 to 24 years who were receiving services from a homeless shelter network focused on this age group. Feasibility was determined based on the young adults’ engagement with various components of the program. Acceptability was determined based on satisfaction ratings of the program as a whole as well as its different components. Finally, preliminary benefits were explored by examining changes in symptoms of depression, PTSD, and emotion regulation. To do so, we conducted a single-arm pilot trial of our multicomponent intervention recruiting participants from an urban homeless shelter network.
Participants were recruited from January 2016 to November 2017 from a homeless shelter network located in Chicago, IL, United States. Young adults responded to flyers distributed throughout the shelters or were referred to the study by their case manager. Potential participants were screened at the shelter by a member of the study staff. If eligible and interested, the participant went through the informed consent process and then filled out a series of baseline assessments on an iPad. These assessments collected information about demographics, trauma history, experience with technology and with the mental health care system, and current psychological symptoms. After completion of the surveys, they were provided with a mobile phone with an activated data plan, phone case, and headphones. Participants were then shown how to use the 3 study apps and were given tips on conserving data usage and using the phone responsibly and safely in an urban space including how best to secure their phones both physically (eg, keeping it safely concealed in certain spaces) and digitally (eg, setting up a password).
The eligibility criteria included (1) aged between 18 and 24 years, (2) English-speaking, and (3) homeless as defined by lacking “a fixed, regular, and adequate nighttime residence,” and sleeping in a Chicago-based shelter for at least 4 nights of the past week. In the initial phase of the study, we had the additional inclusion criterion that participants must have had a self-reported history of physical or sexual abuse before the age of 17 years. This criterion of childhood abuse was removed a year and a half into the study (corresponding to 24 participants) to allow us to reach more young adults who could benefit from the program. The exclusion criteria included (1) cognitive impairment that would prevent them from understanding or fully engaging in study procedures, (2) involvement in any psychotherapy or legal proceedings that would impact participation, and (3) significant suicidal ideation or behaviors or alcohol or substance dependence in the past 6 months. The substance dependence exclusion criterion was removed a year and a half into the study (corresponding to 28 participants) as long as participants were able to engage in the study procedures.
This pilot trial was approved by the Rush University Medical Center institutional review board. People interested in participating completed an initial intake interview and, if eligible to continue, were offered participation in the pilot trial. We referred to the developed intervention program as the Stepping Stone Project, which was a name collaboratively developed with young adults in focus groups. More information about the other substantive findings of these focus groups are reported elsewhere [
Three apps were preinstalled on all mobile phones before distribution to study participants. All participants were asked to use the phones provided by the study and were provided help transferring over their contacts and other phone data if necessary. The service plans were linked to the phones provided to the participants to reduce the likelihood that participants would use multiple phones during the study period. The preinstalled apps included Pocket Helper, which is a daily survey app developed specifically for this study based on focus group input from homeless young adults from the same shelter network [
Pocket Helper is an app that consists of a daily survey and a daily coping skills–focused tip. Data entered into the app were displayed as feedback to the participant and sent to a coach dashboard.
Pocket Helper was also accompanied by a coach dashboard (
Screenshots of the Pocket Helper app.
Coach dashboard for the Pocket Helper app.
IntelliCare is a modular treatment suite consisting of 13 mini-apps, each focused on a singular behavior change technique drawn from CBT and positive psychology [
All participants were eligible to receive up to three 30-min sessions with the coach over the course of 1 month; these sessions were conducted over the phone. The first session was either scheduled during the onboarding procedure conducted in the shelter, or the coach contacted the participant within 24 hours of onboarding to provide a brief introduction into the schedule and protocol for sessions. Subsequent sessions were scheduled either during the first session or via call, text, or email. These sessions were designed to help provide coping skills to participants in line with CBT principles. We used a modular manualized format in which we outlined the general structure of the 3-session format and identified specific skills and strategies that could be provided based on a participant’s needs and preferences.
The session format was as follows: session 1 consisted of orientation and identification of goals, problems, and resources; session 2 consisted of a check-in on progress and a dedicated focus on a specific topic or skill; and session 3 consisted of reviewing progress and discussing steps for moving forward. The skills and strategies outlined in the manual included the following: psychoeducation, problem solving, mindfulness, relaxation, emotion regulation, imagery rehearsal, sleep hygiene, distress tolerance, interpersonal effectiveness, and safety planning. Session content drew from principles of cognitive behavioral approaches; the session structure was not based on any particular treatment but drew from our experience in designing and delivering mobile-based mental health treatments. Participants were also instructed that they could text message the coach at times outside of these scheduled sessions or set up a brief check-in call of approximately 10 to 15 min. For our initial participants, these consisted of
Sessions, phone, and text messaging support were provided by clinical psychology postdoctoral fellows (AKR and CLD). These fellows underwent weekly supervision where they provided updates on their current participants, the activities in the phone sessions, and any outreach made via text messages. The content of all text messages was transcribed and stored, and coaches completed a form outlining their activities during the phone sessions to indicate the skills and strategies that were covered.
Participants completed a series of self-report assessments during enrollment and again at the conclusion of the 1-month intervention.
Demographic information was collected using a 20-item questionnaire developed by the study team. This was administered only at the baseline session. Variables assessed included age, race and ethnicity, gender, sexual orientation, educational, employment, homelessness status, and history of head injuries.
Current technology usage habits were assessed using a 10-item Technology Questionnaire. This was administered only at the baseline session. This questionnaire was created by the study team and asked individuals to respond to items indicating the types of devices used (ie, desktop computer, laptop computer, tablet, and mobile phone), and how often they have access to a mobile phone, phone reception, and Wi-Fi. Items also asked how often they used texting, email, and mobile apps, and what incentives and barriers there were to using telemental health resources.
Current depressive symptoms were assessed using the 9-item Patient Health Questionnaire (PHQ-9) [
Adeptness in identifying and regulating emotions was assessed using the 36-item Difficulties in Emotion Regulation Scale (DERS) [
Current symptoms of PTSD were assessed using the 20-item PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders-5 (PCL-5) [
Lifetime experience with mental health and psychiatric treatment were assessed using a 19-item Treatment Questionnaire, developed by the study team. Individuals responded to questions asking if they had ever, or if they currently, received one-on-one therapy, attended support groups, used self-help resources (books, websites, and mobile phone apps), or taken medication for psychological problems.
Participants were asked to report whether or not they had been exposed to 12 different types of traumatic events during the 1-month study period using a modified version of the Traumatic Events Questionnaire [
Feedback on and satisfaction with various aspects of the study (ie, coaching sessions, daily tips, daily surveys, and IntelliCare apps) were assessed using a 20-item Feedback Questionnaire created by the study team. This was administered only at the endpoint session. Satisfaction with the different aspects of the study was indicated on a 5-point Likert scale. Open-text responses were also provided for participants to describe what they liked most and what they liked least about the coaching sessions, office hours (or contact with the coach outside of scheduled sessions), and the Pocket Helper app. Participants were also asked to give suggestions to improve the study.
Due to the challenges associated with remote data collection, we instituted several procedures to ensure data quality including attempts to reduce missing data by having research staff follow up with participants, incentivizing completion of questionnaires, and offering to set up in-person meetings to promote survey completion. We also attempted to evaluate the validity of the completed self-report measures. Included in both the baseline and follow-up survey were 4 validity questions. These validity question items were placed at the end of a survey and asked the participant to respond with a certain answer. For example, a validity question embedded in the PCL-5 stated “For this question, please select the answer ‘A little bit’.” If an individual completed more than 1 of these incorrectly, their data were considered to be invalid. For clinical outcomes, which were based on self-report questionnaires, we analyzed data from the 22 participants who completed the study and provided valid responses to both the pre- and postassessment. For satisfaction data, which were completed only at the postassessment, we reported on data from all 26 participants with a valid postsurvey. Because participants received the mobile phones, apps, and phone sessions regardless of providing self-report data, we included all participants’ reports of their use of the apps, sessions, and text messaging to demonstrate feasibility of working in this population.
A total of 35 participants consented and were enrolled in the field trial.
Participant flow through recruitment and intervention.
Clinical characteristics of sample at baseline and endpoint (1 month).
Outcome | Baseline, mean (SD) | Endpoint, mean (SD) | Probable disorder at baseline, n (%) | Probable disorder at endpoint, n (%) |
Depressiona | 11.2 (8.0) | 10.1 (8.2) | 10 (46) | 10 (50) |
Post-traumatic stress disorderb | 32.4 (23.8) | 28.2 (23.1) | 11 (50) | 9 (42) |
Emotion regulation | 88.9 (30.6) | 87.0 (34.6) | No clinical cutoff exists | No clinical cutoff exists |
aClinical cutoff for probable depression ≥ 10.
bClinical cutoff for probable post-traumatic stress disorder ≥ 33.
We also assessed participants’ access to and use of various technologies at baseline including computers, mobile devices, and the internet. Most of our participants had mobile phones (25/35, 71%) before receiving one as part of participation in this study, although the type of mobile phone varied considerably. Participants also had access to a variety of other devices including desktops (24/35, 69%), laptops (21/35, 60%), and tablets (19/35, 54%). Similarly, most participants reported having cellular access (20/35, 57%) and access to Wi-Fi for at least half of the waking day (21/35, 60%) although it is worth noting that 20% of our participants (7/35) noted they did not have Wi-Fi access for any part of the day. Use of text messaging and email appeared to be similarly high with 63% (22/35) endorsing texting at least a few times per day and 60% (21/35) endorsing checking email at least a few times per day. Thus, although internet access and use is not ubiquitous, it is quite high.
Few of our participants reported either current or past experience with mental health treatments, with only 17% (6/35) engaged in current and 37% (13/35) engaged in past individual therapy, 3% (1/35) engaged in current and 31% (11/35) engaged in past group therapy, and 6% (2/35) receiving current and 23% (8/35) receiving past medication. Thus, according to self-report, this is a fairly treatment-naïve population.
Most participants (20/35, 57%) completed all 3 of their phone sessions with an average of 2.09 sessions (SD 1.22) completed by each participant. The complete breakdown of the number of sessions completed by participants is provided in
As mentioned previously, we also had our coaches track session content using a reporting form with codes drawing from the topics included in our manual. A total of 73 sessions were provided, and 66 had codes. Of the sessions with codes, 17 had multiple codes with 16 including 2 codes and 1 including 3 codes resulting in 84 total codes across all sessions. The sessions most commonly addressed interpersonal issues (16/84, 19%) or stress management (15/84, 18%). Other common topics included goal setting (9/84, 11%), emotion regulation (8/84, 10%), and family conflict (6/84, 7%). Consistent with our skills-focused approach, very few sessions covered exclusively psychoeducation (1/84, 1%).
Satisfaction with the intervention was high, with 100% (23/23) of participants indicating that they would recommend the study to someone else and 52% (12/23) reporting that they were
Almost half of participants found the skills they learned in session to be beneficial (11/23, 48%), and almost as many report that they regularly used the skills (10/23, 43%). The intervention length was deemed appropriate by most participants with 56% (12/23) indicating that the 1-month intervention was
We logged all issues that occurred with regard to mobile phones provided to participants. Mobile phone loss (through damage, theft, or other loss) was an anticipated event in this study, and we budgeted mobile phones such that we would be able to replace phones of participants if problems occurred. Overall, we replaced 4 mobile phones during the study period. Two of these phones were replaced because of theft, and 2 phones were replaced because of issues with the device. Issues with the device were a result of our mobile service provider switching carriers, which resulted in complications with access to service. This represents a phone replacement rate of 11% (4/35) and a theft rate of 6% (2/35). The cause of phone loss was obtained through self-report, with the exception of replacements because of issues with mobile service, which was validated independently by our research team through communication with the mobile service provider.
In addition to mobile phone loss, we also closely monitored the use of mobile phone data. All participants received a service plan consisting of unlimited calls and text messages and a data plan consisting of 5 GB of data per month. Although we encountered some issues with participants exceeding this data limit, most participants were able to adopt strategies to allow them to remain within this data cap including monitoring data use on their device and switching to Wi-Fi when secure networks were available. During a period where data caps were temporarily suspended, we did have 1 participant use over 100 GB of data mostly because of streaming videos and music, which demonstrates the tendency to use devices as entertainment devices, which can result in large data demands.
Participants experienced very little change on clinical outcomes with small effect sizes for symptoms of depression,
This study is the first attempt to deploy and evaluate a mobile phone–based intervention to address the mental health needs of young adults experiencing homelessness. Our findings demonstrated promising results regarding the feasibility and acceptability of such an intervention. Most participants engaged with the program and reported they were satisfied with their participation, although less than half reported that they found the program helpful. Our goal in this study was to establish whether it would be possible to consider digital interventions as a means to bridge individuals until they are ready for care or act as adjunctive elements to more traditional treatment.
Although this pilot feasibility trial was neither designed nor powered to determine the efficacy of the intervention on clinical outcomes, our exploratory results suggest several points for consideration in future studies. For example, most participants experienced only slight improvements in symptoms of depression, PTSD, and emotion regulation as indicated by small effect sizes quantifying the change. We found larger benefits in PTSD symptoms for those who did not experience traumatic events during the course of our 1-month intervention, but those participants also had slight increases in symptoms of depression and emotion dysregulation. Overall, it seems that our intervention was an attractive way to engage young adults experiencing homelessness, but further work might need to consider proximal outcomes that correspond to long-term outcomes of interest and the most effective interventions to deploy within this style of engagement.
In terms of feasibility, we found that engagement with the program tended to be bimodal with most participants completing either 0 or 3 of the offered phone sessions, although nearly 3 times as many participants completed 3 sessions than 0 sessions. This tends to be quite a different pattern from engagement in technology-based interventions more generally where most people who start do not continue with the intervention after an initial use [
In terms of acceptability, we found high levels of satisfaction with the program as a whole as well as many of the individual components. Specifically, the daily tips and Pocket Helper app was viewed positively by the young adults. Interestingly, this was the part of the program that was developed specifically for this population, based on our formative work with residents of the same shelter network [
Findings for satisfaction and acceptability, however, did not correspond to statistically or clinically significant changes on clinical outcomes. Although these findings might be due to our small sample size, it is also worth considering if other factors were at play. For example, perhaps the participants appreciated the attention and support but did not have enough opportunities or structure in their day-to-day lives to implement the skills learned. Another possibility is that the clinical outcomes overlooked other improvements in well-being, such as sense of connection or validation, that might have been worth exploring. Future work might consider more flexibility in the intervention both in terms of content and dosing as well as exploring alternative means of understanding impact. It is worth considering if the dose of the intervention negatively affected the potential to find benefits as three 30-min sessions over a 1-month period is a relatively light intervention, especially in the context of the stressors and challenges faced by homeless young adults. It might be the case that a longer intervention or an intervention of similar length but targeted at times of high need might be more beneficial for homeless young adults. For example, young adults could be monitored through daily surveys as used in our program, and phone support could be initiated when users face significant challenges such as traumatic events or interpersonal stressors or have needs related to work or school placement. Identifying a treatment strategy that will lead to clinical benefits while remaining feasible for youth is an important direction for future research.
Our findings are worth contextualizing with regard to the limitations of our design and methods and the formative stage of this work to inform the delivery of technology-based interventions to meet the needs of individuals experiencing homelessness. First, we focused on quantitative data to assess satisfaction and feasibility, and qualitative data could have helped elaborate some of the satisfaction and feasibility issues, particularly the potentially conflicting results on high satisfaction and low helpfulness. Second, although we had fairly broad eligibility criteria, we did allow for some iteration in these criteria during the study. As noted, we removed the criterion related to substance or alcohol dependence but only after 5 potential participants had already been screened out because of those criteria. Of those 5 participants, however, 3 were also receiving concurrent counseling that would also have made them ineligible for this study. Third, it is impossible to disentangle the impact of the clinical intervention (eg, the apps, phone sessions, and text messaging) from receiving a mobile phone with data service. Access to a mobile device with internet connectivity may be an intervention in and of itself. Some research shows that mobile devices may be used by individuals to help cope with stressful situations and when used appropriately can improve psychological and physiological functioning [
Although our satisfaction ratings might be viewed as somewhat mixed given the positive reviews of the program as a whole, the lower ratings of the coach support and office hours, and the lowest ratings of the IntelliCare apps, it is also worth comparing these ratings with satisfaction with other mental health services, especially among young adults experiencing homelessness. Indeed, past work has demonstrated that homeless young adults have low rates of satisfaction with mental health services and that issues of mistrust must be overcome in engaging these young adults successfully in treatment. Our overall rate of 57% of participants completing all of their sessions compares favorably with outpatient psychotherapy generally [
We again note that the impact on clinical symptoms was small and somewhat disappointing in light of the high levels of engagement and positive evaluation of the program we received from the participants. Future work should also assess more proximal outcomes that might mediate subsequent symptom change. Judging from the topics covered in our sessions, outcomes worth assessing would be interpersonal functioning and social support, emotion regulation and coping, and goal setting and problem solving. It must be noted that many of our participants reported using the skills they learned in the sessions and through the apps, but it would be worth evaluating this. Unfortunately, the impact of our intervention on clinical symptoms was small to moderate, with the exception of symptoms of PTSD for individuals who did not experience traumatic events or depression symptoms for those individuals who did experience traumatic events during the intervention period. As noted previously, it is possible that our intervention did not provide a proper dose of treatment. Another possibility is that the intervention itself, but not the delivery platform, may have been a poor fit for our population. Although we used CBT-based coping strategies based on past work that indicated benefits of cognitive behavioral treatments for individuals experiencing homelessness [
This pilot trial of a mobile phone–based mental health intervention for homeless young adults provides mixed support for the promises of such an intervention moving forward. On one hand, we found high rates of engagement and satisfaction, which is noteworthy especially among a population that typically has low rates of trust and satisfaction with mental health services. On the other hand, we found only small benefits in symptoms of mental health issues such as depression, PTSD, and emotion regulation. Given this was a pilot study focused on feasibility and acceptability, we caution against over interpretation of these findings. Nevertheless, we would remiss if we did not note that the extent of changes in clinical symptoms were smaller than we had hoped. We think future work could consider more proximal outcomes to clinical symptom change, as well as other outcomes linked to intervention engagement such as interest in mental health resources, mental health literacy, or mental health stigma. In addition, future research should consider whether intervention content or dose could be modified to be more appropriate for the platform we developed. As such, we believe the major contribution of this study is that it reveals important lessons about engaging homeless young adults using mobile technology. These lessons include the benefits of consistent outreaches (ie, tips and text messaging) as well as the success of providing mobile phones and remote interventions without significant safety issues or loss of devices. We hope further studies continue to explore best ways to create engaging and impactful mental health interventions for homeless young adults and believe this study demonstrated potential ways that technology could play an important role in improving mental health care access for this population.
cognitive behavioral therapy
Difficulties in Emotion Regulation Scale
PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders-5
9-item Patient Health Questionnaire
post-traumatic stress disorder
Funding for this project was provided by pilot funding from Help for Children, Hedge Funds Care (AKZ and NSK: Multiple Principal Investigators), and mobile phones were donated by Sparrow, Mobile for All. SMS was supported by a career development award from the National Institute of Mental Health (K08-MH102336) and is an investigator with the Implementation Research Institute, at the Washington University in St Louis; through an award from the National Institute of Mental Health (R25-MH08091607) and the Department of Veterans Affairs, Health Services Research & Development Service, Quality Enhancement Research Initiative. AKZ is supported by a career development award from the National Institute of Mental Health (K23-MH103394). NSK is supported by the National Center for Advancing Translational Sciences (UL1-TR002398), the National Institute on Drug Abuse (R01-DA041071), and the Cynthia Oudejans Harris, MD, Endowment Fund at Rush University Medical Center. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official view of the National Institutes of Health. We would like to thank Chris Karr and the Center for Behavioral Intervention Technologies for programming Pocket Helper as well as the previously developed apps made use of in this study.
None declared.