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.
Interventions to reduce alcohol use typically include several elements, such as information on the risks of alcohol consumption, planning for sensible drinking, and training of protective behavioral strategies. However, the effectiveness of these individual intervention elements within comprehensive programs has not been addressed so far, but it could provide valuable insights for the development of future interventions. Just-in-time interventions provided via mobile devices are intended to help people make healthy decisions in the moment and thus could influence health behavior.
The aim of this study was to test the proximal effects of a mobile phone–delivered, just-in-time planning intervention to reduce alcohol use in adolescents who reported recent binge drinking. The efficacy of this individual intervention element was tested within a comprehensive intervention program to reduce problem drinking in adolescents.
The study had an AB/BA crossover design, in which participants were randomly allocated to (1) a group receiving the planning intervention (A) in period 1 and assessment only (B) in period 2 or (2) a group receiving assessment only (B) in period 1 and the planning intervention (A) in period 2. The planning intervention included a text message to choose one of two predetermined if-then plans to practice sensible drinking with friends or when going out and a prompt to visualize the chosen plan. There was a washout period of at least 1 week between period 1 and period 2.
Out of 633 program participants who recently binge drank, 136 (21.5%) were receptive in both periods of time and provided data on the proximal outcome, which was the number of alcoholic drinks consumed with friends or when going out. After the planning intervention, the number of alcoholic drinks consumed was approximately one standard drink lower compared with the finding without the intervention (P=.01).
A mobile phone–delivered, just-in-time, if-then planning intervention to practice sensible drinking with friends or when going out is effective in reducing alcohol consumption among adolescents who report recent binge drinking. Based on the relatively low percentage of participants with self-reported receptivity for the planning intervention, measures to increase the population impact of similar planning interventions should be implemented and tested in future trials.
ISRCTN Registry ISRCTN52150713; http://www.isrctn.com/ISRCTN52150713
Alcohol use is a major cause of disease burden in most countries worldwide and is among the 10 leading risk factors in all Central European countries [
Internationally recognized indicators of problem drinking are (1) average daily consumption of more than two standard drinks for men and one standard drink for women [
Interventions, including personalized normative feedback and drinking reduction strategies, as major intervention elements show small short-term effects on the reduction of binge drinking prevalence in young people [
In this study, we tested the proximal effects of a mobile phone–delivered, just-in-time, if-then planning intervention to reduce alcohol use in adolescents who reported recent binge drinking. We hypothesized that the number of alcoholic drinks consumed with friends or when going out would be lower during the just-in-time planning intervention as compared with assessment only.
The study aimed to determine the proximal effects of a just-in-time planning intervention to reduce alcohol use in adolescents who reported recent binge drinking. The study was registered at Current Controlled Trials International Standard Randomized Controlled Trials Number (ISRCTN 52150713, assigned June 2, 2017). The study protocol was approved by the ethics committee of the Philosophical Faculty at the University of Zurich, Switzerland (date of approval: April 18, 2017). The trial was executed in compliance with the Declaration of Helsinki.
Within a randomized controlled AB/BA crossover design, each participant received the planning intervention (A) and assessment only (B) in a randomized order. The trial was conducted in Switzerland, and participants were recruited between June 2017 and July 2018. Participants were recruited in vocational and upper secondary schools and participated in a comprehensive mobile phone–based intervention program to reduce problem drinking with a duration of 3 months. The inclusion criteria were (1) ownership of a mobile phone, (2) recent binge drinking, (3) alcohol consumption in the evening/night with friends or when going out, and (4) available data on preferred if-then plans. The just-in-time planning intervention was based on effective implementation intention and action planning interventions [
Vocational and upper secondary schools in the Swiss cantons of Zurich and Berne were invited to participate in the comprehensive mobile phone–based intervention program named
The participants for this study to test the proximal effects of the mobile phone–delivered, just-in-time, alcohol planning intervention were selected automatically by a computer algorithm within the
The computer algorithm selected recent binge drinking adolescents with alcohol consumption in the evening or night when going out or when being with friends and with available data on preferred if-then plans. A total of 633 (44.6%) of the 1419
The assessment questions and the text messages of the planning intervention are depicted in
The data necessary for the provision of the just-in-time planning intervention were assessed within the baseline assessment. This assessment was also performed during the school lesson immediately after the online screening and informed consent procedure for the comprehensive intervention program (mentioned above). It included (1) selection of the typical drinking day in the course of the week when going out or when being with friends, (2) selection of the typical drinking time when going out or when being with friends, and (3) selection of two out of nine favorite if-then plans providing strategies to drink little or no alcohol with friends or when going out. One reason to select only two out of nine if-then plans was that these plans were also presented within SMS text messages, which are restricted in their length and number of characters. Another reason was not to overstrain the participant within this situation. Regarding the typical drinking days, we did not include Monday, Tuesday, and Wednesday, as previous studies [
As the entire intervention program had a total duration of 12 weeks and we considered a washout period of 1 week as appropriate between period 1 and 2, assessments of the state of receptivity and potential provision of the subsequent planning intervention were possible six times during the intervention (in weeks 1, 3, 5, 7, 9, and 11). In order to obtain a maximum sample size for this crossover trial, the state of receptivity was assessed in as many weeks as possible until a participant was receptive twice, that is, there were up to six chances for the period 1 assessment, and the remaining assessments after responding to the period 1 assessment were for period 2. After being receptive twice, the participants no longer received this state of the receptivity assessment. The time interval for responding to the state of the receptivity assessment was 6 hours. Participants who did not respond within this time period and those who indicated that they did not meet with friends or go out did not receive the subsequent messages of the planning intervention and the outcome assessment.
Flow of study participants.
Assessments and planning intervention from participants' perspective.
The estimation of the effect size was based on the results of a controlled study testing the effectiveness of implementation intentions to reduce alcohol use in a sample of the general population [
The online baseline assessment included the following variables: sex, age, immigration background, tobacco smoking, the Alcohol Use Disorders Identification Test (AUDIT) [
Tobacco smoking was assessed using the following question: “Do you currently smoke cigarettes or have you smoked in the past?” The response options were as follows: (1) I smoke cigarettes daily; (2) I smoke cigarettes occasionally but not daily; (3) I smoked cigarettes in the past, but I do not smoke anymore; and (4) I have never smoked cigarettes or have smoked less than 100 cigarettes throughout my life. In the analysis, we combined categories (1) and (2) as smokers and categories (3) and (4) as nonsmokers. Alcohol use was assessed through the consumption items of the AUDIT (AUDIT-C) [
The primary outcome of this planning intervention study was the number of alcoholic drinks with friends or when going out on the day of the intervention. This proximal outcome was assessed 24 hours after assessment of the state of receptivity, that is, at 5 pm on the day following the individual indicated drinking day. The secondary outcome of this planning intervention study was binge drinking, which was defined as drinking at least five drinks in males and four drinks in females with friends or when going out on the individual indicated drinking day.
We initially examined the data of the primary outcome, which was based on the self-reported number of alcoholic drinks entered as free text. Based on a visual inspection of the distributions and the recommendations of Osborne and Overbay [
We used chi-square tests for categorical variables and
Intervention effects of the planning intervention were tested following the recommendations of Wellek and Blettner [
Following the latter recommendation and to rule out that the treatment effects were confounded by time effects or a carryover effect, as the washout phase was not long enough, we calculated the sum of the number of drinks consumed in the two periods for each subject and compared it across the two sequence groups by a
To test for the effects of the alcohol planning intervention, we used a
This analyzed group did not differ from participants who could not be assessed (n=497) with respect to the following baseline characteristics: age, immigration background, tobacco smoking status, AUDIT-C, typical drinking day, and typical drinking time on the indicated drinking day. However, the nonreceptive participants or nonresponders were more likely male (51.5% vs 40.8%, χ21=4.9;
Characteristics of the analyzed study sample (n=136) and the two sequence groups (AB/BA) are shown in
Baseline characteristics of the study sample.
Variable | Sequence: intervention–control (AB) (n=66) | Sequence: control–intervention (BA) (n=70) | Total (n=136) | |
|
|
|
|
|
|
Male | 35 (53.0%) | 35 (50.0%) | 70 (51.5%) |
|
Female | 31 (47.0%) | 35 (50.0%) | 66 (48.5%) |
Age, mean (SD) | 16.9 (1.0) | 17.2 (1.3) | 17.1 (1.1) | |
|
|
|
|
|
|
No immigration background | 30 (45.5%) | 42 (60.0%) | 72 (52.9%) |
|
One or both parents born outside Switzerland | 36 (54.5%) | 28 (40.0%) | 64 (47.1%) |
|
|
|
|
|
|
Upper secondary school | 20 (30.3%) | 24 (34.3%) | 44 (32.4%) |
|
Vocational school | 46 (69.7%) | 46 (65.7%) | 92 (67.6%) |
|
|
|
|
|
|
Daily or occasional cigarette smoking | 31 (47.0%) | 34 (48.6%) | 65 (47.8%) |
|
Nonsmoking | 35 (53.0%) | 36 (51.4%) | 71 (52.2%) |
AUDIT-Ca, mean (SD) | 6.2 (1.6) | 6.7 (1.7) | 6.4 (1.7) | |
|
|
|
|
|
|
Thursday | 1 (1.5%) | 0 (0%) | 1 (0.7%) |
|
Friday | 28 (42.4%) | 27 (38.6%) | 55 (40.4%) |
|
Saturday | 37 (56.1%) | 43 (61.4%) | 80 (58.8%) |
|
Sunday | 0 (0%) | 0 (0%) | 0 (0%) |
|
|
|||
|
Morning or afternoon | 0 (0%) | 0 (0%) | 0 (0%) |
|
Between 6 pm and 8 pm | 4 (6.1%) | 5 (7.1%) | 9 (6.6%) |
|
Between 8 pm and 10 pm | 32 (48.5%) | 40 (57.1%) | 72 (52.9%) |
|
After 10 pm | 30 (45.5%) | 25 (35.7%) | 55 (40.4%) |
aAUDIT-C: consumption items of the Alcohol Use Disorders Identification Test
To rule out that the treatment effects were confounded by time effects or carryover effects, we calculated the sum of the number of drinks consumed in the two periods for each subject and compared it across the two sequence groups by a
As shown in
Concerning the secondary outcome (binge drinking on the previous day with friends or when going out), in the AB and BA sequence groups, the prevalences were 30% (20/66) and 36% (25/70), respectively, after receiving the planning intervention and 35% (23/66) and 36% (25/70), respectively, without the intervention. A chi-square test comparing the number of participants who did not change between period 1 and period 2, who reported binge drinking in period 2 but not in period 1, and who reported binge drinking in period 1 but not in period 2, showed no statistically significant difference between the two sequence groups (
Effects of the alcohol planning intervention.
Variable | Sequence: intervention–control (AB) (n=66) | Sequence: control–intervention (BA) (n=70) | Test value | |||||
Period 1 (A) | Period 2 (B) | Within-subject difference (period 2−period 1) | Period 1 (B) | Period 2 (A) | Within-subject difference (period 2−period 1) | |||
Number of alcoholic drinks on the previous day with friends or when going out, mean (SD) | 2.79 (3.09) | 3.68 (3.62) | 0.89 (3.71) | 4.21 (3.67) | 3.43 (3.85) | −0.79 (4.69) | 2.31a | .01 |
Binge drinking on the previous day with friends or when going out, n (%) | 20 (30%) | 23 (35%) | 3 (5%) | 25 (36%) | 25 (36%) | 0 (0.0%) | 1.34b | .25 |
a
bChi-square test for the comparison of binge drinking change in period 2 compared with period 1 between the condition sequences.
This study aimed to test the proximal effects of a just-in-time planning intervention for reducing alcohol use in adolescents who reported recent binge drinking. To the best of our knowledge, this is the first study to test the effects of a just-in-time intervention for reducing alcohol consumption. The study revealed the following three main results, which are discussed below: (1) It is feasible to test the proximal effects of single intervention elements like implementation intentions within a comprehensive multicomponent intervention program if the program is delivered via a mobile phone and has a minimum duration; (2) An if-then alcohol planning intervention is effective to reduce the mean alcohol use on typical drinking days in young people by about one standard drink; (3) A large proportion of adolescents is not receptive to the just-in-time planning intervention, when the state of receptivity is assessed via text messaging and a reply within a limited response time is required for triggering the intervention.
This study underlines that the proximal effects of specific intervention elements like just-in-time implementation intentions could be tested in a randomized controlled crossover design within a comprehensive intervention program, if some requirements are met as follows: (1) The intervention program is provided automatically (eg, via a mobile phone that allows server-triggered just-in-time interventions); (2) The duration of the program is long enough to assess receptivity for the intervention several times in order to have at least two time points for comparison; and (3) The intervention element is presented after a long enough time (wash-out period) from the other elements or contents of the program in order to reduce confounding. Another requirement, which was hardly met in previous trials on just-in-time interventions, is an adequate sample size to have enough statistical power [
The result that a single if-then planning intervention is effective to reduce alcohol consumption is in line with previous findings derived from laboratory research [
Although the just-in-time delivery of the alcohol planning intervention might be partly responsible for its effectiveness, it remains unclear whether the population impact (number of participants reached multiplied by effectiveness [
Several limitations of this study should be mentioned. First, the number of alcoholic drinks consumed was self-reported, and in contrast to the baseline assessment, no examples of standard drinks were provided within the text messaging–based assessment of the proximal outcome. Second, although the crossover design applied has several advantages (eg, it avoids problems of comparability of the study and control groups because each participant is his/her own control and the required sample size is low), carryover effects might have confounded part of the intervention effects, although the respective finding was not relevant. Third, the sample analyzed within this study systematically differed from all participants randomized, with respect to sex and type of school, which limits the external validity of the results. Fourth, we solely used baseline data for tailoring the intervention time. The possibility of updating or adapting the intervention time periodically might increase intervention effectiveness.
In conclusion, this study shows that just-in-time interventions could be tested and implemented in the area of addiction and that digitally provided alcohol planning interventions could reduce alcohol use in adolescents who report recent binge drinking. Future studies should focus on increasing the reach and outcome of just-in-time alcohol planning interventions by testing other delivery formats or by sensor-triggered intervention delivery, which, for example, dynamically monitors a participant’s context and provides support when high-risk environments, such as areas with many nightlife locations and social situations, are sensed [
CONSORT-eHEALTH checklist (V 1.6.1).
Alcohol Use Disorders Identification Test
consumption items of the Alcohol Use Disorders Identification Test
Funding for this project was provided by the Swiss Federal Office of Public Health (grant no. 16.929761).
TK is affiliated with the Center for Digital Health Interventions (www.c4dhi.org), a joint initiative of the Department of Management, Technology and Economics at ETH Zurich and the Institute of Technology Management at the University of St. Gallen, which is funded in part by the Swiss health insurer CSS. TK is also the cofounder of Pathmate Technologies, a university spinoff company that creates and delivers digital clinical pathways and has used the open source MobileCoach platform for that purpose. The funding institution did not influence the design and conduct of the study; the management, analysis or interpretation of data; or the preparation, review or approval of the manuscript.