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.
Smoking rates are significantly higher among young people experiencing homelessness than in the general population. Despite a willingness to quit, homeless youth have little success in doing so on their own, and existing cessation resources tailored to this population are lacking. Homeless youth generally enjoy the camaraderie and peer support that group-based programs offer, but continuous in-person support during a quit attempt can be prohibitively expensive.
This study aimed to assess the feasibility and acceptability of an automated text messaging intervention (TMI) as an adjunct to group-based cessation counseling and provision of nicotine patches to help homeless youth quit smoking. This paper outlines the lessons learned from the implementation of the TMI intervention.
Homeless youth smokers aged 18 to 25 years who were interested in quitting (n=77) were recruited from drop-in centers serving homeless youth in the Los Angeles area. In this pilot randomized controlled trial, all participants received a group-based cessation counseling session and nicotine patches, with 52% (40/77) randomly assigned to receive 6 weeks of text messages to provide additional support for their quit attempt. Participants received text messages on their own phone rather than receiving a study-issued phone for the TMI. We analyzed baseline and follow-up survey data as well as back-end data from the messaging platform to gauge the acceptability and feasibility of the TMI among the 40 participants who received it.
Participants had widespread (smart)phone ownership—16.4% (36/219) were ineligible for study participation because they did not have a phone that could receive text messages. Participants experienced interruptions in their phone use (eg, 44% [16/36] changed phone numbers during the follow-up period) but reported being able to receive the majority of messages. These survey results were corroborated by back-end data (from the program used to administer the TMI) showing a message delivery rate of about 95%. Participant feedback points to the importance of carefully crafting text messages, which led to high (typically above 70%) approval of most text messaging components of the intervention. Qualitative feedback indicated that participants enjoyed the group counseling session that preceded the TMI and suggested including more such group elements into the intervention.
The TMI was well accepted and feasible to support smoking cessation among homeless youth. Given high rates of smartphone ownership, the next generation of phone-based smoking cessation interventions for this population should consider using approaches beyond text messages and focus on finding ways to develop effective approaches to include group interaction using remote implementation. Given overall resource constraints and in particular the exigencies of the currently ongoing COVID-19 epidemic, phone-based interventions are a promising approach to support homeless youth, a population urgently in need of effective smoking cessation interventions.
ClinicalTrials.gov NCT03874585; https://clinicaltrials.gov/ct2/show/NCT03874585
RR2-10.1186/s13722-020-00187-6
National data indicate that 19% of people aged 18 to 25 years in the United States are current (past 30 day) cigarette smokers [
This paper is based on our previous research that focused specifically on young homeless smokers and supported the need for cessation programs adapted to this population. We found high willingness to quit smoking among this population but little success in doing so on their own or when using existing available cessation resources [
We also found that homeless youth who are interested in quitting enjoy the camaraderie and peer support that group-based programs offer [
In this paper, we report on our experiences implementing the first TMI targeted at smoking cessation among homeless youth as a part of a pilot evaluation that found promising results of the TMI on smoking reduction. We lay out the lessons learned in designing appropriate and effective text messages and implementing the technical aspects of the intervention and report on the feedback from the participants regarding different aspects of the intervention. We conclude with a discussion of what worked well and what could benefit from improvement with the hope that our lessons can be of value to other, much-needed TMIs for this particularly vulnerable population.
The findings reported here are from a pilot study of 77 current smokers aged 18 to 25 years who desired to quit and were recruited from 3 drop-in centers serving homeless youth in the Los Angeles area. The unit of analysis was the individual, but individuals were assigned to groups (standard care alone vs TMI adjunct) based on the drop-in center where they were seeking services during recruitment hours. The intervention was carried out in a cluster cross-over randomized controlled design such that each drop-in center alternated between serving as an intervention or control site by phase across the field period. After formative work with 26 participants [
Study flow.
The project team generated 174 text messages to send to all TMI participants (see previous work [
It was important to ensure that the content of the text messages reflected the unique circumstances of young people experiencing homelessness. In addition, to make messages more effective we used recent insights from behavioral economics in the design of the messages [
Given that TMIs have not been previously implemented for behavior change among young homeless smokers, there is little guidance from the existing literature regarding the frequency and timing of text messages for this population. Therefore, we conducted several focus groups with a total of 18 homeless smokers and elicited usability testing feedback with a separate sample of 10 homeless smokers recruited from the drop-in centers [
The results in the following section come from two surveys. The first was a baseline survey that was administered when participants were first recruited for the TMI before the counseling session. This survey asked respondents for information such as demographics, phone use, smoking behavior at baseline, and other substance use. Three months after the baseline survey (and hence about 6 weeks after the end of the intervention phase), all participants received a follow-up survey that asked for information similar to that at baseline; those in the TMI condition were asked additional questions on their experience with the TMI. The baseline survey was administered in a group setting immediately prior to the group counseling session. We used self-administered paper-pencil forms, which was the most feasible option in our field setting. The follow-up surveys were administered either in person or via phone interview. Survey response forms were then scanned and checked for accuracy. While staff were available to assist participants in completing the surveys, no one required such assistance. Participants received $20 for the baseline survey and $40 for the follow-up survey. The baseline survey had 77 participants while the follow-up survey had 66 participants due to attrition.
In
Participant cellphone characteristics at baseline.
Variables | Standard condition (answered yes; n=37), n (%) | TMIa condition (answered yes; n=40), n (%) |
It is my own phone | 32 (86) | 34 (85) |
It is a phone owned by a friend, partner, or relative | 5 (13) | 6 (15) |
It is a phone obtained through a benefits program | 13 (32) | 21 (52) |
It is a smartphone | 31 (84) | 29 (73) |
I have unlimited minutes on this phone | 23 (62) | 31 (78) |
I have unlimited texts on this phone | 25 (68) | 32 (80) |
I have unlimited data on this phone | 15 (40) | 22 (55) |
I can view webpages on this phone because I have a data plan | 27 (73) | 27 (67) |
I can only view webpages on this phone when I’m connected to the Wi-Fi | 15 (40) | 20 (50) |
aTMI: text messaging–based intervention.
As can be seen in
Despite the frequent occurrence of phone loss and switching of numbers, 81% (29/36) of TMI participants reported not having problems receiving the texts. Further, 72% (26/36) had no trouble accessing the hyperlinks provided in some of the text messages, allowing them to access supplemental information (see
Cell phone and number retention at follow-up (n=66).
Intervention delivery metrics.
Metric | Never, n (%) | Sometimes, n (%) | Often, n (%) | Always, n (%) |
Trouble keeping phone battery charged | 10 (28) | 17 (47) | 6 (17) | 3 (8) |
Ran out of messages and couldn’t get messages | 29 (80) | 3 (8) | 2 (6) | 2 (6) |
Trouble accessing hyperlinks on phone | 26 (72) | 4 (11) | 2 (6) | 4 (11) |
Did not have cellphone reception | 24 (66) | 10 (28) | 1 (3) | 1 (3) |
Intervention acceptability metrics.
Metric | Strongly disagree, n (%) | Somewhat disagree, n (%) | Somewhat agree, n (%) | Strongly agree, n (%) | |||||
|
|||||||||
|
helped when I was experiencing a craving | 4 (11) | 6 (17) | 17 (49) | 8 (23) | ||||
|
helped when I was having trouble staying quit | 3 (9) | 9 (27) | 14 (41) | 8 (23) | ||||
|
helped keep me motivated to quit | 3 (9) | 6 (17) | 11 (31) | 15 (43) | ||||
|
provided me information that I could use | 6 (17) | 3 (9) | 4 (11) | 22 (63) | ||||
|
|||||||||
|
being able to text a keyword for extra support | 5 (14) | 4 (11) | 7 (20) | 19 (53) | ||||
|
the memes and emojis that were sent | 4 (11) | 4 (11) | 10 (29) | 17 (49) | ||||
|
that the text messages were personalized | 4 (12) | 2 (6) | 11 (32) | 17 (50) | ||||
|
the tone of the text messages | 6 (18) | 4 (12) | 12 (35) | 12 (35) | ||||
|
when the messages asked me to respond | 6 (18) | 5 (14) | 14 (41) | 9 (27) | ||||
I received text messages late because my phone wasn’t charged, or I had bad reception | 14 (41) | 5 (14) | 9 (27) | 5 (14) | |||||
Hyperlinks to websites provided information I could use | 5 (14) | 3 (9) | 10 (29) | 16 (47) | |||||
I would recommend the CRUSH IT! program to a friend who is trying to quit smoking | 4 (12) | 1 (3) | 7 (19) | 23 (64) |
In this paper, we discussed the development of a smoking cessation TMI for young homeless smokers and reported lessons learned regarding intervention components such as development of the messages, (smart)phone ownership, technical challenges with the intervention, and feasibility (such as whether messages were received) and acceptability (ie, feedback about what intervention aspects the participants liked).
Regarding message development, we found that conducting focus groups with young homeless smokers and pilot testing with participants before rolling out the intervention were crucial steps to successfully tailoring the content, tone, and frequency of messages to the target population. Participants preferred messages that provided information specific to their current living situation, were light in tone rather than being preachy, and contained fun elements such as emojis and memes. We then programmed the messages developed with the participants’ input using a web-based text messaging platform that allowed for automatized sending of messages and categorization of participant responses for subsequent sending of appropriate follow-up messages, which minimized human resources required for implementation and guaranteed that messages were sent out with a high success rate.
As would be expected based on other research showing widespread cell phone ownership among individuals experiencing homelessness [
Last, similar to prior research finding high acceptability among young people experiencing homelessness for using text messaging for daily data collection [
Our study has significant strengths such as being the first TMI for smoking cessation support for homeless youth and implementation at 3 sites in the Los Angeles area, which allowed for testing in different contexts. However, it also had limitations. For example, results may not generalize to homeless youth in other geographic areas or to those younger than age 18 years. Due to the study being a pilot, we were intentionally not powered to evaluate statistical differences between the standard and TMI conditions. Future studies should test the promising results of this pilot in a fully powered trial ideally in several different locations to test for generalizability of our results. In addition, it is a limitation that we did not collect qualitative data (eg, debriefing interviews) with the homeless youth who used the TMI at the end of the study, which might have identified additional strengths and weaknesses of this approach that could inform future research efforts in this area. Clearly, our study contributes to a growing evidence base that TMIs for smoking cessation can be effectively implemented even for highly transient and resource-constrained populations such as homeless youth, but that adaptation to their specific needs (including conducting appropriate formative work) is needed to render the TMI acceptable.
In terms of future directions, one specific recommendation we got from participant feedback is that they really enjoyed the group cessation counseling session prior to receiving the TMI and wanted more group elements incorporated into the intervention. Going forward, future smoking cessation interventions for this population should consider using approaches making full use of smartphone capabilities, including virtual approaches to leverage group interaction not requiring in-person meetings such as using social media chatrooms. Given their low costs and low requirements for human resources (particularly given the currently ongoing COVID-19 epidemic), (smart)phone-based interventions are a promising approach to support homeless youth, a population urgently in need of effective smoking cessation interventions.
In conclusion, we find that most young homeless smokers have cell phones that allow mHealth interventions, with many being in possession of smartphones that typically have unlimited minutes and texts. Given that almost half also have unlimited data plans, it seems that in the near future internet-based interventions requiring smartphones will also be feasible in this population. In line with widespread concern in the literature, we find that homeless youth reported frequent occurrence of phone loss and switching of numbers. However, despite these difficulties the majority of participants reported not having problems receiving the study texts, and engagement with the different intervention components was generally high. We hope that the lessons derived from this pilot intervention serve as useful inputs for future mHealth studies for this population in need of smoking cessation interventions.
mobile health
odds ratio
text messaging–based interaction
This research was supported by funds from Tobacco-Related Disease Research Program of the University of California (grant number 27IP-0051; principal investigator JT). The authors wish to thank the drop-in centers that allowed us to conduct this research, Isabel Leamon, Alice Kim, and Sarika Bharil for their assistance with enrollment and program delivery, and Rick Garvey and the RAND Survey Research Group for their assistance in conducting the follow-up data collection. Our thanks to the study participants who so generously volunteered their time and insight to allow this study to happen.
None declared.