Identifying helpful text message strategies to reduce alcohol use among young adults
Alcohol use often increases through young adulthood, which for some can have immediate and long-term consequences. Young adults, however, are harder to engage in treatment than their older counterparts. Innovative, “low-touch” strategies like texting may help reach more young adults that might not otherwise seek treatment. This randomized controlled trial examined several text message interventions designed to reduce alcohol use in non-treatment seeking young adults who presented to an emergency department.
During young adulthood (e.g., 18-25 years of age), many individuals change their living situation, begin gainful employment, and enroll in post-secondary schooling. Alcohol use typically peaks during this life stage and binge drinking—consuming four or more drinks per occasion for women and five or more drinks for men—is highest among young adults, with 29% (9.8 million) reporting binge drinking in the last month. The heightened prevalence of heavy alcohol use during this life stage and the associated risks and consequences highlights the importance of developing impactful intervention strategies.
Despite elevated rates of alcohol harms and alcohol use disorder in this age group, young adults are less likely than middle and older aged adults to seek treatment. Linking non-treatment seeking young adults with engaging, accessible interventions could reduce alcohol-related harms, yet, identifying those young adults with potentially problematic drinking can be difficult given they are not actively looking for treatment. Consequently, conducting universal screening among young adults in clinical settings where they are likely to be (e.g., primary care and emergency departments for alcohol related accidents and injuries) may be an efficient way to reach more of them. Once identified, however, the challenge still remains about how best to engage this age group in some kind of helpful intervention that can make a difference.
Young adults most commonly use text messaging when communicating. Text messaging, specifically, has been used in alcohol research as a stand-alone intervention and in tandem with other digital media (e.g., apps, websites). These interventions can modestly reduce alcohol use compared with no, or basic, health information, but the power of texting interventions lies primarily in their scalability—a small impact on a large scale can have a big overall effect at a population level. Yet, we know little about which types of intervention ingredients are most important. Identifying the active ingredients of intervention effects would enhance the efficiency of these types of interventions. It would also be helpful to know if active ingredients work as well alone or in combination with other ingredients. To help identify more effective and efficient alcohol use interventions for this difficult to engage age-group, this study assessed the impact of five different behavioral change strategies delivered in a text messaging format.
HOW WAS THIS STUDY CONDUCTED?
The research team conducted a randomized controlled trial that tested the effects of five different 12-week text messaging strategies on reducing binge alcohol use. Measures of alcohol use were obtained for the past 30 days before the intervention (i.e., baseline), at a 3-month follow-up (i.e., immediately following the 12-week intervention), and at a 6-month follow-up (i.e., 3 months after intervention completion). Due to small sample sizes resulting from COVID-related recruitment challenges, each strategy was only compared to a control strategy instead of each of the other text message strategies, which was the original plan. The study was primarily interested in frequency of binge drinking, but they also include odds of binge drinking, number of drinks per drinking occasion, and drinking consequences as additional outcomes.
Participants were young adults who presented to one of four emergency departments for any reason in western Pennsylvania between August 2017 and April 2021. Young adults aged 18–25 were eligible if they were medically stable and had an AUDIT-C screening score of ≥ 3 women or ≥4 for men, which indicates higher risk for hazardous drinking or active alcohol use disorder. The AUDIT-C is a 3-item abridged version of the 10-item Alcohol Use Disorders Identification Test (AUDIT) that focuses on items that assess alcohol consumption resulting in a possible score range from 0–12, with higher scores indicating a higher probability of hazardous drinking and alcohol use disorder. Young adults were not eligible if they had previously received treatment for an alcohol or other drug problem; were currently receiving psychiatric treatment; or did not own a cellphone. Thus, the sample was non-treatment seeking young adults who reported high risk levels of alcohol use.
In total, 1,131 young adults were recruited and randomized to 1 of 5 conditions. All participants completed a baseline assessment and a 2-week onboarding process during which they were texted on the 2 days a week they identified as days they most typically drank alcohol. The text was sent at 3pm on these days and asked the participant to report drinking plans and desire to get drunk. If the participant did not respond within two hours, each text was sent once more. Only participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of the 2-week period were randomized to 1 of the 5 conditions.
The interventions were delivered via automated text messaging for 12 weeks. Each condition focused on a unique component or combination of components of behavioral change techniques.
Pre-event messages were sent at 3pm on the 2 days per week that an individual reported drinking the most alcohol. Post-event messages were sent at 12pm on the days following the 2 days per week an individual reporting dinking the most.
If no response was given within 2 hours of the initial text, the message was resent once after which the response window closed.
The five conditions were:
1) TRACK: this sought to promote self-monitoring. It served as the control intervention and was identical to the 2-week onboarding period. It asked participants if they planned to drink and, if so, their desire to drunk on a scale from 0 (not at all) to 8 (completely) followed by a post-event alcohol consumption assessment asking how many drinks they consumed the previous day. All other conditions described below also included these elements.
2) PLAN: this focused on challenging drinking expectancies, or what a person thought they would get out of drinking. It added a component by providing feedback based on a participant’s response to their plans to drink and desire to get drunk. For example, if a young adult reported that they desired to get drunk they would receive the text: “We understand how you feel. Remember that the excitement induced by the first couple drinks often goes away after that.”
3) USE: this aimed to transform a person’s post-drinking evaluation of how positive the drinking event was. It included post-event feedback based on the number of drinks reported. For example, if a young adult reported binge drinking, they would receive the text: “You reported a binge drinking episode yesterday. This is more alcohol than 80% of young adults like you consumes when drinking.”
4) GOAL: this sought to enhance goal self-efficacy by asking participants if they’d “commit to a goal to drink less than [X] drinks today?” and to rate on a scale from 0 (not at all) to 8 (completely) how confident they were that they could achieve that goal. Feedback was also provided when participants selected a goal (e.g., “You report low confidence in meeting this goal. Trust yourself. You know more than you think you do—Dr. Spock”). Participants in the GOAL condition also received post-event feedback on their success or failure of goal attainment (e.g., “You met your drinking limit goal yesterday. Right on!”).
5) COMBO: in this condition the young adults received all the components.
At the 3-month follow-up immediately after the intervention and again at the 6-month follow-up, all participants completed the full assessment asking participants about their past month alcohol use and associated consequences. The primary outcome of interest was the number of binge drinking days in the past month—defined as drinking 4 or more drinks for women and 5 or more drinks for men within 2 hours. Binge drinking (yes/no), drinks per drinking occasion, and alcohol-related consequences were additional outcomes.
Of the 1,131 participants randomized to conditions, just over half (54%) identified as White/Caucasian and just over two-thirds as female (69%), with a mean age of 22. A little over 1/3 (36%) were enrolled in college, and the average AUDIT-C score was 6 – above cut-offs for probable DSM 5 moderate or severe alcohol use disorder. They averaged 3 binge drinking days in the month before treatment. About half (52%) reported daily or almost daily tobacco use, and 60%, 5%, 18% reported lifetime use of cannabis, opioids, and “other drugs” respectively. The mean trait impulsivity score was 42 out of a possible 80, with higher scores representing higher impulsivity.
WHAT DID THIS STUDY FIND?
COMBO, GOAL, and USE interventions were linked with reduced binge drinking.
Figure 1. TRACK is the group to which the other study conditions are being compared to.
At the 3-month follow-up, which occurred immediately after the intervention period, participants in the COMBO, GOAL, and USE conditions reported a decrease in the mean binge drinking days by 0.7, 0.4, and 0.4, respectively. After controlling for other factors, COMBO, GOAL, and USE were also associated with reductions in binge drinking days, any binge drinking, drinks per occasion, and alcohol consequences compared to the TRACK control intervention. Participants in the COMBO, GOAL, and USE interventions reported about 1 less binge drinking day per month, an average of 8% lower prevalence of any binge drinking days, 1 less standard drink per drinking occasion, and 3 fewer alcohol consequences. These reductions were maintained at the 6-month follow-up. While PLAN was no better than TRACK control intervention at the 3-month follow-up, it was associated with more favorable binge drinking and fewer alcohol consequences at the 6-month follow-up.
Figure 2. TRACK is the group to which the other study conditions are being compared to.
WHAT ARE THE IMPLICATIONS OF THE STUDY FINDINGS?
Researchers in this study tested the relative impact of four different text messaging interventions compared to a control text messaging condition where participants just tracked the number of drinks they consumed. There were not enough participants to compare the four intervention strategies to each other directly. Yet, the findings highlight the potential benefit of text messaging interventions to reduce hazardous alcohol use to at least some degree among young adults.
It should also be remembered however, that analyses only included participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of an initial 2-week texting engagement period. This may mean that the results pertain only to those young adults who may be particularly motivated to engage in this type of intervention approach or otherwise motivated to change their alcohol use.
The GOAL, USE, and COMBO text messaging interventions seem to be helpful across time (i.e., 3-month follow-up and 6-month follow-up) and for different outcomes (e.g., binge drinking, alcohol consequences). These findings suggest text message interventions that both facilitate alcohol related thinking before a drinking event and offer feedback after the drinking event are equally useful. The modest benefits associated with these text messaging interventions could potentially confer major health impacts if scaled, which seems quite feasible given they are automated and require few resources to implement and are easy to access. Text messaging interventions, especially for young adults, could be helpful and cost effective, but more research is needed across different young adult samples to confirm the robustness of these initial findings and also determine which young adults, in particular, may or may not be responsive to such approaches.
More research is needed also to identify and confirm the most effective text messaging interventions for harmful alcohol use among young adults. An additional consideration is the fit between intervention and young adult. The young adults in this study were non-treatment seeking and reported 3 binge drinking days per month. They were not, in general, individuals with severe alcohol use disorder. They were identified and recruited for the study because they showed up to the emergency department, disclosed substance use, and reported some impulsivity. This was a group of young adults who occasionally do risky things. Thus, findings from this study may be applicable to a large proportion of young adults, not just those typically considered “needing” treatment. Based on findings from this study, text message interventions could be helpful in reducing alcohol use among young adults engaging in hazardous drinking, especially those not seeking treatment.
Only participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of the 2-week period were randomized to 1 of the 5 conditions. Any generalizations should be made with caution as this means that only young adults that might have been particularly interested in changing their alcohol use and well suited to this approach participated.
The reason for emergency department visits among young adults were not reported, which limits the generalizability of the findings because a young adult presenting to the emergency department for an alcohol-related issue is likely to be qualitatively different from one who is being admitted for something else such as for a sports injury.
Participant recruitment overlapped with the height of the COVID-19 pandemic. COVID-19, social restrictions, and related stress likely influenced alcohol use and technological behaviors among young adults in unmeasured ways, which may have influenced results. Thus, the ability to generalize to young adults post COVID-19 is unknown. However, the ubiquity of both cellphone and alcohol use among this population, before and after COVID-19, suggest text messaging interventions could still be impactful.
BOTTOM LINE
Young adults that initially responded to at least 50% of texts and stated that they wished to continue participating in a study at the end of a 2-week texting induction period were randomized to 1 of 5 conditions where they received texts asking about drinking plans and were provided with drinking-related feedback after a drinking event 2 days per week. Overall, the young adults had less frequent and lower odds of binge drinking, fewer drinks per drinking occasion, and fewer alcohol consequences at the 3-month follow-up immediately after the intervention period and at the 6-month follow-up. Findings also suggest text message interventions that both facilitate alcohol related thinking before a drinking event and offer feedback after the drinking event may be equally useful. The modest benefits associated with these text messaging interventions could have major health impact if scaled, which is feasible because they require few resources and are easily accessible. However, more research is needed to continue teasing out the optimal active ingredients of text messaging interventions and also which types of these texting interventions may be most helpful for which specific subgroups of young adults. The findings do suggest that text message interventions for alcohol use among young adults with hazardous drinking could be somewhat helpful, unlikely to hurt, and cost-effective at a population level.
For individuals and families seeking recovery: The use of automated text messages to address alcohol behaviors and intentions represents an accessible and potentially helpful intervention for young adults. This study’s findings suggest that long-term feedback via text messages can help reduce alcohol use and its related consequences. However, further research is needed to compare the effectiveness of different types of texting interventions and identify optimal timing and frequency of messages. Also, more research is needed to discover which young adults in particular may respond particularly well or poorly to these kinds of approaches. In general, the results suggest that providing long-term feedback regarding alcohol behaviors and intentions is likely to help and unlikely to hinder positive changes in alcohol use.
For treatment professionals and treatment systems: Mobile health interventions are an increasingly accessible and low-resource intervention that may be especially useful to help engage young adults with harmful and hazardous drinking. This study’s findings suggest that text messages providing drinking-related feedback only 2 days per week were associated with less binge drinking, lower odds of binge drinking, fewer drinks per drinking occasion, and fewer alcohol consequences. Although the study cannot pinpoint the exact messaging that is most helpful, it does appear that reflecting on alcohol use before and after drinking is helping. More work is needed to say for certain which approached work best for different people.
For scientists: Mobile technology offers new ways to reach and engage with young adults not seeking treatment. Although findings from this study suggest that promoting reflection before a drinking event and also providing alcohol-related feedback after a drinking event via text messages could help young adults reduce their alcohol use, more research is needed to establish what behavior change strategies work best via texting and for whom. Furthermore, future studies should help to explore how mobile technology can be both a stand-alone intervention as well as a component of face-to-face engagement with treatment and continuing care.
For policy makers: This study found that text messages providing alcohol-related feedback were associated with reductions in binge drinking days, odds of binge drinking, drinks per drinking occasion, and alcohol consequences. Policies and funding that help implement texting interventions in primary care and treatment settings may help reduce alcohol-related harm among young adults and their communities.
During young adulthood (e.g., 18-25 years of age), many individuals change their living situation, begin gainful employment, and enroll in post-secondary schooling. Alcohol use typically peaks during this life stage and binge drinking—consuming four or more drinks per occasion for women and five or more drinks for men—is highest among young adults, with 29% (9.8 million) reporting binge drinking in the last month. The heightened prevalence of heavy alcohol use during this life stage and the associated risks and consequences highlights the importance of developing impactful intervention strategies.
Despite elevated rates of alcohol harms and alcohol use disorder in this age group, young adults are less likely than middle and older aged adults to seek treatment. Linking non-treatment seeking young adults with engaging, accessible interventions could reduce alcohol-related harms, yet, identifying those young adults with potentially problematic drinking can be difficult given they are not actively looking for treatment. Consequently, conducting universal screening among young adults in clinical settings where they are likely to be (e.g., primary care and emergency departments for alcohol related accidents and injuries) may be an efficient way to reach more of them. Once identified, however, the challenge still remains about how best to engage this age group in some kind of helpful intervention that can make a difference.
Young adults most commonly use text messaging when communicating. Text messaging, specifically, has been used in alcohol research as a stand-alone intervention and in tandem with other digital media (e.g., apps, websites). These interventions can modestly reduce alcohol use compared with no, or basic, health information, but the power of texting interventions lies primarily in their scalability—a small impact on a large scale can have a big overall effect at a population level. Yet, we know little about which types of intervention ingredients are most important. Identifying the active ingredients of intervention effects would enhance the efficiency of these types of interventions. It would also be helpful to know if active ingredients work as well alone or in combination with other ingredients. To help identify more effective and efficient alcohol use interventions for this difficult to engage age-group, this study assessed the impact of five different behavioral change strategies delivered in a text messaging format.
HOW WAS THIS STUDY CONDUCTED?
The research team conducted a randomized controlled trial that tested the effects of five different 12-week text messaging strategies on reducing binge alcohol use. Measures of alcohol use were obtained for the past 30 days before the intervention (i.e., baseline), at a 3-month follow-up (i.e., immediately following the 12-week intervention), and at a 6-month follow-up (i.e., 3 months after intervention completion). Due to small sample sizes resulting from COVID-related recruitment challenges, each strategy was only compared to a control strategy instead of each of the other text message strategies, which was the original plan. The study was primarily interested in frequency of binge drinking, but they also include odds of binge drinking, number of drinks per drinking occasion, and drinking consequences as additional outcomes.
Participants were young adults who presented to one of four emergency departments for any reason in western Pennsylvania between August 2017 and April 2021. Young adults aged 18–25 were eligible if they were medically stable and had an AUDIT-C screening score of ≥ 3 women or ≥4 for men, which indicates higher risk for hazardous drinking or active alcohol use disorder. The AUDIT-C is a 3-item abridged version of the 10-item Alcohol Use Disorders Identification Test (AUDIT) that focuses on items that assess alcohol consumption resulting in a possible score range from 0–12, with higher scores indicating a higher probability of hazardous drinking and alcohol use disorder. Young adults were not eligible if they had previously received treatment for an alcohol or other drug problem; were currently receiving psychiatric treatment; or did not own a cellphone. Thus, the sample was non-treatment seeking young adults who reported high risk levels of alcohol use.
In total, 1,131 young adults were recruited and randomized to 1 of 5 conditions. All participants completed a baseline assessment and a 2-week onboarding process during which they were texted on the 2 days a week they identified as days they most typically drank alcohol. The text was sent at 3pm on these days and asked the participant to report drinking plans and desire to get drunk. If the participant did not respond within two hours, each text was sent once more. Only participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of the 2-week period were randomized to 1 of the 5 conditions.
The interventions were delivered via automated text messaging for 12 weeks. Each condition focused on a unique component or combination of components of behavioral change techniques.
Pre-event messages were sent at 3pm on the 2 days per week that an individual reported drinking the most alcohol. Post-event messages were sent at 12pm on the days following the 2 days per week an individual reporting dinking the most.
If no response was given within 2 hours of the initial text, the message was resent once after which the response window closed.
The five conditions were:
1) TRACK: this sought to promote self-monitoring. It served as the control intervention and was identical to the 2-week onboarding period. It asked participants if they planned to drink and, if so, their desire to drunk on a scale from 0 (not at all) to 8 (completely) followed by a post-event alcohol consumption assessment asking how many drinks they consumed the previous day. All other conditions described below also included these elements.
2) PLAN: this focused on challenging drinking expectancies, or what a person thought they would get out of drinking. It added a component by providing feedback based on a participant’s response to their plans to drink and desire to get drunk. For example, if a young adult reported that they desired to get drunk they would receive the text: “We understand how you feel. Remember that the excitement induced by the first couple drinks often goes away after that.”
3) USE: this aimed to transform a person’s post-drinking evaluation of how positive the drinking event was. It included post-event feedback based on the number of drinks reported. For example, if a young adult reported binge drinking, they would receive the text: “You reported a binge drinking episode yesterday. This is more alcohol than 80% of young adults like you consumes when drinking.”
4) GOAL: this sought to enhance goal self-efficacy by asking participants if they’d “commit to a goal to drink less than [X] drinks today?” and to rate on a scale from 0 (not at all) to 8 (completely) how confident they were that they could achieve that goal. Feedback was also provided when participants selected a goal (e.g., “You report low confidence in meeting this goal. Trust yourself. You know more than you think you do—Dr. Spock”). Participants in the GOAL condition also received post-event feedback on their success or failure of goal attainment (e.g., “You met your drinking limit goal yesterday. Right on!”).
5) COMBO: in this condition the young adults received all the components.
At the 3-month follow-up immediately after the intervention and again at the 6-month follow-up, all participants completed the full assessment asking participants about their past month alcohol use and associated consequences. The primary outcome of interest was the number of binge drinking days in the past month—defined as drinking 4 or more drinks for women and 5 or more drinks for men within 2 hours. Binge drinking (yes/no), drinks per drinking occasion, and alcohol-related consequences were additional outcomes.
Of the 1,131 participants randomized to conditions, just over half (54%) identified as White/Caucasian and just over two-thirds as female (69%), with a mean age of 22. A little over 1/3 (36%) were enrolled in college, and the average AUDIT-C score was 6 – above cut-offs for probable DSM 5 moderate or severe alcohol use disorder. They averaged 3 binge drinking days in the month before treatment. About half (52%) reported daily or almost daily tobacco use, and 60%, 5%, 18% reported lifetime use of cannabis, opioids, and “other drugs” respectively. The mean trait impulsivity score was 42 out of a possible 80, with higher scores representing higher impulsivity.
WHAT DID THIS STUDY FIND?
COMBO, GOAL, and USE interventions were linked with reduced binge drinking.
Figure 1. TRACK is the group to which the other study conditions are being compared to.
At the 3-month follow-up, which occurred immediately after the intervention period, participants in the COMBO, GOAL, and USE conditions reported a decrease in the mean binge drinking days by 0.7, 0.4, and 0.4, respectively. After controlling for other factors, COMBO, GOAL, and USE were also associated with reductions in binge drinking days, any binge drinking, drinks per occasion, and alcohol consequences compared to the TRACK control intervention. Participants in the COMBO, GOAL, and USE interventions reported about 1 less binge drinking day per month, an average of 8% lower prevalence of any binge drinking days, 1 less standard drink per drinking occasion, and 3 fewer alcohol consequences. These reductions were maintained at the 6-month follow-up. While PLAN was no better than TRACK control intervention at the 3-month follow-up, it was associated with more favorable binge drinking and fewer alcohol consequences at the 6-month follow-up.
Figure 2. TRACK is the group to which the other study conditions are being compared to.
WHAT ARE THE IMPLICATIONS OF THE STUDY FINDINGS?
Researchers in this study tested the relative impact of four different text messaging interventions compared to a control text messaging condition where participants just tracked the number of drinks they consumed. There were not enough participants to compare the four intervention strategies to each other directly. Yet, the findings highlight the potential benefit of text messaging interventions to reduce hazardous alcohol use to at least some degree among young adults.
It should also be remembered however, that analyses only included participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of an initial 2-week texting engagement period. This may mean that the results pertain only to those young adults who may be particularly motivated to engage in this type of intervention approach or otherwise motivated to change their alcohol use.
The GOAL, USE, and COMBO text messaging interventions seem to be helpful across time (i.e., 3-month follow-up and 6-month follow-up) and for different outcomes (e.g., binge drinking, alcohol consequences). These findings suggest text message interventions that both facilitate alcohol related thinking before a drinking event and offer feedback after the drinking event are equally useful. The modest benefits associated with these text messaging interventions could potentially confer major health impacts if scaled, which seems quite feasible given they are automated and require few resources to implement and are easy to access. Text messaging interventions, especially for young adults, could be helpful and cost effective, but more research is needed across different young adult samples to confirm the robustness of these initial findings and also determine which young adults, in particular, may or may not be responsive to such approaches.
More research is needed also to identify and confirm the most effective text messaging interventions for harmful alcohol use among young adults. An additional consideration is the fit between intervention and young adult. The young adults in this study were non-treatment seeking and reported 3 binge drinking days per month. They were not, in general, individuals with severe alcohol use disorder. They were identified and recruited for the study because they showed up to the emergency department, disclosed substance use, and reported some impulsivity. This was a group of young adults who occasionally do risky things. Thus, findings from this study may be applicable to a large proportion of young adults, not just those typically considered “needing” treatment. Based on findings from this study, text message interventions could be helpful in reducing alcohol use among young adults engaging in hazardous drinking, especially those not seeking treatment.
Only participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of the 2-week period were randomized to 1 of the 5 conditions. Any generalizations should be made with caution as this means that only young adults that might have been particularly interested in changing their alcohol use and well suited to this approach participated.
The reason for emergency department visits among young adults were not reported, which limits the generalizability of the findings because a young adult presenting to the emergency department for an alcohol-related issue is likely to be qualitatively different from one who is being admitted for something else such as for a sports injury.
Participant recruitment overlapped with the height of the COVID-19 pandemic. COVID-19, social restrictions, and related stress likely influenced alcohol use and technological behaviors among young adults in unmeasured ways, which may have influenced results. Thus, the ability to generalize to young adults post COVID-19 is unknown. However, the ubiquity of both cellphone and alcohol use among this population, before and after COVID-19, suggest text messaging interventions could still be impactful.
BOTTOM LINE
Young adults that initially responded to at least 50% of texts and stated that they wished to continue participating in a study at the end of a 2-week texting induction period were randomized to 1 of 5 conditions where they received texts asking about drinking plans and were provided with drinking-related feedback after a drinking event 2 days per week. Overall, the young adults had less frequent and lower odds of binge drinking, fewer drinks per drinking occasion, and fewer alcohol consequences at the 3-month follow-up immediately after the intervention period and at the 6-month follow-up. Findings also suggest text message interventions that both facilitate alcohol related thinking before a drinking event and offer feedback after the drinking event may be equally useful. The modest benefits associated with these text messaging interventions could have major health impact if scaled, which is feasible because they require few resources and are easily accessible. However, more research is needed to continue teasing out the optimal active ingredients of text messaging interventions and also which types of these texting interventions may be most helpful for which specific subgroups of young adults. The findings do suggest that text message interventions for alcohol use among young adults with hazardous drinking could be somewhat helpful, unlikely to hurt, and cost-effective at a population level.
For individuals and families seeking recovery: The use of automated text messages to address alcohol behaviors and intentions represents an accessible and potentially helpful intervention for young adults. This study’s findings suggest that long-term feedback via text messages can help reduce alcohol use and its related consequences. However, further research is needed to compare the effectiveness of different types of texting interventions and identify optimal timing and frequency of messages. Also, more research is needed to discover which young adults in particular may respond particularly well or poorly to these kinds of approaches. In general, the results suggest that providing long-term feedback regarding alcohol behaviors and intentions is likely to help and unlikely to hinder positive changes in alcohol use.
For treatment professionals and treatment systems: Mobile health interventions are an increasingly accessible and low-resource intervention that may be especially useful to help engage young adults with harmful and hazardous drinking. This study’s findings suggest that text messages providing drinking-related feedback only 2 days per week were associated with less binge drinking, lower odds of binge drinking, fewer drinks per drinking occasion, and fewer alcohol consequences. Although the study cannot pinpoint the exact messaging that is most helpful, it does appear that reflecting on alcohol use before and after drinking is helping. More work is needed to say for certain which approached work best for different people.
For scientists: Mobile technology offers new ways to reach and engage with young adults not seeking treatment. Although findings from this study suggest that promoting reflection before a drinking event and also providing alcohol-related feedback after a drinking event via text messages could help young adults reduce their alcohol use, more research is needed to establish what behavior change strategies work best via texting and for whom. Furthermore, future studies should help to explore how mobile technology can be both a stand-alone intervention as well as a component of face-to-face engagement with treatment and continuing care.
For policy makers: This study found that text messages providing alcohol-related feedback were associated with reductions in binge drinking days, odds of binge drinking, drinks per drinking occasion, and alcohol consequences. Policies and funding that help implement texting interventions in primary care and treatment settings may help reduce alcohol-related harm among young adults and their communities.
During young adulthood (e.g., 18-25 years of age), many individuals change their living situation, begin gainful employment, and enroll in post-secondary schooling. Alcohol use typically peaks during this life stage and binge drinking—consuming four or more drinks per occasion for women and five or more drinks for men—is highest among young adults, with 29% (9.8 million) reporting binge drinking in the last month. The heightened prevalence of heavy alcohol use during this life stage and the associated risks and consequences highlights the importance of developing impactful intervention strategies.
Despite elevated rates of alcohol harms and alcohol use disorder in this age group, young adults are less likely than middle and older aged adults to seek treatment. Linking non-treatment seeking young adults with engaging, accessible interventions could reduce alcohol-related harms, yet, identifying those young adults with potentially problematic drinking can be difficult given they are not actively looking for treatment. Consequently, conducting universal screening among young adults in clinical settings where they are likely to be (e.g., primary care and emergency departments for alcohol related accidents and injuries) may be an efficient way to reach more of them. Once identified, however, the challenge still remains about how best to engage this age group in some kind of helpful intervention that can make a difference.
Young adults most commonly use text messaging when communicating. Text messaging, specifically, has been used in alcohol research as a stand-alone intervention and in tandem with other digital media (e.g., apps, websites). These interventions can modestly reduce alcohol use compared with no, or basic, health information, but the power of texting interventions lies primarily in their scalability—a small impact on a large scale can have a big overall effect at a population level. Yet, we know little about which types of intervention ingredients are most important. Identifying the active ingredients of intervention effects would enhance the efficiency of these types of interventions. It would also be helpful to know if active ingredients work as well alone or in combination with other ingredients. To help identify more effective and efficient alcohol use interventions for this difficult to engage age-group, this study assessed the impact of five different behavioral change strategies delivered in a text messaging format.
HOW WAS THIS STUDY CONDUCTED?
The research team conducted a randomized controlled trial that tested the effects of five different 12-week text messaging strategies on reducing binge alcohol use. Measures of alcohol use were obtained for the past 30 days before the intervention (i.e., baseline), at a 3-month follow-up (i.e., immediately following the 12-week intervention), and at a 6-month follow-up (i.e., 3 months after intervention completion). Due to small sample sizes resulting from COVID-related recruitment challenges, each strategy was only compared to a control strategy instead of each of the other text message strategies, which was the original plan. The study was primarily interested in frequency of binge drinking, but they also include odds of binge drinking, number of drinks per drinking occasion, and drinking consequences as additional outcomes.
Participants were young adults who presented to one of four emergency departments for any reason in western Pennsylvania between August 2017 and April 2021. Young adults aged 18–25 were eligible if they were medically stable and had an AUDIT-C screening score of ≥ 3 women or ≥4 for men, which indicates higher risk for hazardous drinking or active alcohol use disorder. The AUDIT-C is a 3-item abridged version of the 10-item Alcohol Use Disorders Identification Test (AUDIT) that focuses on items that assess alcohol consumption resulting in a possible score range from 0–12, with higher scores indicating a higher probability of hazardous drinking and alcohol use disorder. Young adults were not eligible if they had previously received treatment for an alcohol or other drug problem; were currently receiving psychiatric treatment; or did not own a cellphone. Thus, the sample was non-treatment seeking young adults who reported high risk levels of alcohol use.
In total, 1,131 young adults were recruited and randomized to 1 of 5 conditions. All participants completed a baseline assessment and a 2-week onboarding process during which they were texted on the 2 days a week they identified as days they most typically drank alcohol. The text was sent at 3pm on these days and asked the participant to report drinking plans and desire to get drunk. If the participant did not respond within two hours, each text was sent once more. Only participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of the 2-week period were randomized to 1 of the 5 conditions.
The interventions were delivered via automated text messaging for 12 weeks. Each condition focused on a unique component or combination of components of behavioral change techniques.
Pre-event messages were sent at 3pm on the 2 days per week that an individual reported drinking the most alcohol. Post-event messages were sent at 12pm on the days following the 2 days per week an individual reporting dinking the most.
If no response was given within 2 hours of the initial text, the message was resent once after which the response window closed.
The five conditions were:
1) TRACK: this sought to promote self-monitoring. It served as the control intervention and was identical to the 2-week onboarding period. It asked participants if they planned to drink and, if so, their desire to drunk on a scale from 0 (not at all) to 8 (completely) followed by a post-event alcohol consumption assessment asking how many drinks they consumed the previous day. All other conditions described below also included these elements.
2) PLAN: this focused on challenging drinking expectancies, or what a person thought they would get out of drinking. It added a component by providing feedback based on a participant’s response to their plans to drink and desire to get drunk. For example, if a young adult reported that they desired to get drunk they would receive the text: “We understand how you feel. Remember that the excitement induced by the first couple drinks often goes away after that.”
3) USE: this aimed to transform a person’s post-drinking evaluation of how positive the drinking event was. It included post-event feedback based on the number of drinks reported. For example, if a young adult reported binge drinking, they would receive the text: “You reported a binge drinking episode yesterday. This is more alcohol than 80% of young adults like you consumes when drinking.”
4) GOAL: this sought to enhance goal self-efficacy by asking participants if they’d “commit to a goal to drink less than [X] drinks today?” and to rate on a scale from 0 (not at all) to 8 (completely) how confident they were that they could achieve that goal. Feedback was also provided when participants selected a goal (e.g., “You report low confidence in meeting this goal. Trust yourself. You know more than you think you do—Dr. Spock”). Participants in the GOAL condition also received post-event feedback on their success or failure of goal attainment (e.g., “You met your drinking limit goal yesterday. Right on!”).
5) COMBO: in this condition the young adults received all the components.
At the 3-month follow-up immediately after the intervention and again at the 6-month follow-up, all participants completed the full assessment asking participants about their past month alcohol use and associated consequences. The primary outcome of interest was the number of binge drinking days in the past month—defined as drinking 4 or more drinks for women and 5 or more drinks for men within 2 hours. Binge drinking (yes/no), drinks per drinking occasion, and alcohol-related consequences were additional outcomes.
Of the 1,131 participants randomized to conditions, just over half (54%) identified as White/Caucasian and just over two-thirds as female (69%), with a mean age of 22. A little over 1/3 (36%) were enrolled in college, and the average AUDIT-C score was 6 – above cut-offs for probable DSM 5 moderate or severe alcohol use disorder. They averaged 3 binge drinking days in the month before treatment. About half (52%) reported daily or almost daily tobacco use, and 60%, 5%, 18% reported lifetime use of cannabis, opioids, and “other drugs” respectively. The mean trait impulsivity score was 42 out of a possible 80, with higher scores representing higher impulsivity.
WHAT DID THIS STUDY FIND?
COMBO, GOAL, and USE interventions were linked with reduced binge drinking.
Figure 1. TRACK is the group to which the other study conditions are being compared to.
At the 3-month follow-up, which occurred immediately after the intervention period, participants in the COMBO, GOAL, and USE conditions reported a decrease in the mean binge drinking days by 0.7, 0.4, and 0.4, respectively. After controlling for other factors, COMBO, GOAL, and USE were also associated with reductions in binge drinking days, any binge drinking, drinks per occasion, and alcohol consequences compared to the TRACK control intervention. Participants in the COMBO, GOAL, and USE interventions reported about 1 less binge drinking day per month, an average of 8% lower prevalence of any binge drinking days, 1 less standard drink per drinking occasion, and 3 fewer alcohol consequences. These reductions were maintained at the 6-month follow-up. While PLAN was no better than TRACK control intervention at the 3-month follow-up, it was associated with more favorable binge drinking and fewer alcohol consequences at the 6-month follow-up.
Figure 2. TRACK is the group to which the other study conditions are being compared to.
WHAT ARE THE IMPLICATIONS OF THE STUDY FINDINGS?
Researchers in this study tested the relative impact of four different text messaging interventions compared to a control text messaging condition where participants just tracked the number of drinks they consumed. There were not enough participants to compare the four intervention strategies to each other directly. Yet, the findings highlight the potential benefit of text messaging interventions to reduce hazardous alcohol use to at least some degree among young adults.
It should also be remembered however, that analyses only included participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of an initial 2-week texting engagement period. This may mean that the results pertain only to those young adults who may be particularly motivated to engage in this type of intervention approach or otherwise motivated to change their alcohol use.
The GOAL, USE, and COMBO text messaging interventions seem to be helpful across time (i.e., 3-month follow-up and 6-month follow-up) and for different outcomes (e.g., binge drinking, alcohol consequences). These findings suggest text message interventions that both facilitate alcohol related thinking before a drinking event and offer feedback after the drinking event are equally useful. The modest benefits associated with these text messaging interventions could potentially confer major health impacts if scaled, which seems quite feasible given they are automated and require few resources to implement and are easy to access. Text messaging interventions, especially for young adults, could be helpful and cost effective, but more research is needed across different young adult samples to confirm the robustness of these initial findings and also determine which young adults, in particular, may or may not be responsive to such approaches.
More research is needed also to identify and confirm the most effective text messaging interventions for harmful alcohol use among young adults. An additional consideration is the fit between intervention and young adult. The young adults in this study were non-treatment seeking and reported 3 binge drinking days per month. They were not, in general, individuals with severe alcohol use disorder. They were identified and recruited for the study because they showed up to the emergency department, disclosed substance use, and reported some impulsivity. This was a group of young adults who occasionally do risky things. Thus, findings from this study may be applicable to a large proportion of young adults, not just those typically considered “needing” treatment. Based on findings from this study, text message interventions could be helpful in reducing alcohol use among young adults engaging in hazardous drinking, especially those not seeking treatment.
Only participants that responded to at least 50% of texts and responded that they wished to continue participating at the end of the 2-week period were randomized to 1 of the 5 conditions. Any generalizations should be made with caution as this means that only young adults that might have been particularly interested in changing their alcohol use and well suited to this approach participated.
The reason for emergency department visits among young adults were not reported, which limits the generalizability of the findings because a young adult presenting to the emergency department for an alcohol-related issue is likely to be qualitatively different from one who is being admitted for something else such as for a sports injury.
Participant recruitment overlapped with the height of the COVID-19 pandemic. COVID-19, social restrictions, and related stress likely influenced alcohol use and technological behaviors among young adults in unmeasured ways, which may have influenced results. Thus, the ability to generalize to young adults post COVID-19 is unknown. However, the ubiquity of both cellphone and alcohol use among this population, before and after COVID-19, suggest text messaging interventions could still be impactful.
BOTTOM LINE
Young adults that initially responded to at least 50% of texts and stated that they wished to continue participating in a study at the end of a 2-week texting induction period were randomized to 1 of 5 conditions where they received texts asking about drinking plans and were provided with drinking-related feedback after a drinking event 2 days per week. Overall, the young adults had less frequent and lower odds of binge drinking, fewer drinks per drinking occasion, and fewer alcohol consequences at the 3-month follow-up immediately after the intervention period and at the 6-month follow-up. Findings also suggest text message interventions that both facilitate alcohol related thinking before a drinking event and offer feedback after the drinking event may be equally useful. The modest benefits associated with these text messaging interventions could have major health impact if scaled, which is feasible because they require few resources and are easily accessible. However, more research is needed to continue teasing out the optimal active ingredients of text messaging interventions and also which types of these texting interventions may be most helpful for which specific subgroups of young adults. The findings do suggest that text message interventions for alcohol use among young adults with hazardous drinking could be somewhat helpful, unlikely to hurt, and cost-effective at a population level.
For individuals and families seeking recovery: The use of automated text messages to address alcohol behaviors and intentions represents an accessible and potentially helpful intervention for young adults. This study’s findings suggest that long-term feedback via text messages can help reduce alcohol use and its related consequences. However, further research is needed to compare the effectiveness of different types of texting interventions and identify optimal timing and frequency of messages. Also, more research is needed to discover which young adults in particular may respond particularly well or poorly to these kinds of approaches. In general, the results suggest that providing long-term feedback regarding alcohol behaviors and intentions is likely to help and unlikely to hinder positive changes in alcohol use.
For treatment professionals and treatment systems: Mobile health interventions are an increasingly accessible and low-resource intervention that may be especially useful to help engage young adults with harmful and hazardous drinking. This study’s findings suggest that text messages providing drinking-related feedback only 2 days per week were associated with less binge drinking, lower odds of binge drinking, fewer drinks per drinking occasion, and fewer alcohol consequences. Although the study cannot pinpoint the exact messaging that is most helpful, it does appear that reflecting on alcohol use before and after drinking is helping. More work is needed to say for certain which approached work best for different people.
For scientists: Mobile technology offers new ways to reach and engage with young adults not seeking treatment. Although findings from this study suggest that promoting reflection before a drinking event and also providing alcohol-related feedback after a drinking event via text messages could help young adults reduce their alcohol use, more research is needed to establish what behavior change strategies work best via texting and for whom. Furthermore, future studies should help to explore how mobile technology can be both a stand-alone intervention as well as a component of face-to-face engagement with treatment and continuing care.
For policy makers: This study found that text messages providing alcohol-related feedback were associated with reductions in binge drinking days, odds of binge drinking, drinks per drinking occasion, and alcohol consequences. Policies and funding that help implement texting interventions in primary care and treatment settings may help reduce alcohol-related harm among young adults and their communities.
