Heavy drinking is responsible for major health and social problems. Brief interventions have been shown to be effective, but there have been difficulties in reaching those who might benefit from them. Pilot studies have indicated that a Web-based intervention is likely to be acceptable to heavy drinkers and may produce some health benefits. However, there are few data on how many people might use such a program, the patterns of use, and potential benefits.
The aim was to examine the demographic characteristics of users of a free, Web-based, 6-week intervention for heavy drinkers and to describe the methods by which users identified the site, the pattern of site use and attrition, the characteristics associated with completing the program, and the self-reported impact on alcohol-related outcomes.
Cohort study. Visitors to the Web site were offered screening with the Fast Alcohol Screening Test, and those scoring above the cutoff for risky drinking were invited to register with the program. Demographic information was collected routinely at registration, and questionnaires were completed at the end of weeks 1 and 6. The outcome measures assessed dependency (Short Alcohol Dependency Data Questionnaire), harms (modified Alcohol Problems Questionnaire), and mental health (Clinical Outcomes in Routine Evaluation–Outcome Measure).
The records of 10000 users were analyzed. The mean age was 37.4 years, 51.1% were female, 37.5% were single, and 42.4% lived with children. The majority were White British, lived in the United Kingdom, and reported occupations from the higher socioeconomic strata. Over 70% connected to the Down Your Drink (Down Your Drink) site from another Internet-based resource, whereas only 5.8% heard about the site from a health or other professional. Much of the Web site use (40%) was outside normal working hours. Attrition from the program was high, with only 16.5% of registrants completing the whole 6 weeks. For those who completed the program, and the final outcome measures, measures of dependency, alcohol-related problems, and mental health symptoms were all reduced at week 6.
The Web-based intervention was highly used, and those who stayed with the program showed significant reductions in self-reported indicators of dependency, alcohol-related problems, and mental health symptoms; however, this association cannot be assumed to be causal. Programs of this type may have the potential to reach large numbers of heavy drinkers who might not otherwise seek help. There are significant methodological challenges and further research is needed to fully evaluate such interventions.
Excess alcohol consumption, and the harm caused by it, is a major public health concern throughout the developed world [
Brief interventions seek to change views of the personal acceptability of excessive drinking and to encourage self-directed behavior change. They can be delivered by practitioners or as self-help materials. There is a substantial body of evidence demonstrating that brief interventions for individuals at risk can have significant impact on reducing alcohol consumption and, in some cases, alcohol-related harm when delivered both in primary and secondary health care settings [
Until recently, self-help materials were almost exclusively paper-based. However, the Internet has triggered a growth in more interactive self-help materials. It is thought that this interactivity is likely to enhance the potential for behavior change [
One exception is the Alcohol Help Centre, which provided online personal feedback to users of an online eHealth service [
Down Your Drink is a well-established, comprehensive, freely available, interactive Web-based treatment program for people with alcohol problems. An initial pilot study demonstrated the feasibility of the approach [
The aims of the present study were to describe the patterns of use and self-reported effectiveness among users of Down Your Drink. The study set out to describe the demographic characteristics of users, the methods employed to identify and access Down Your Drink, the patterns of use, the demographic and clinical characteristics associated with completing the 6-week program, and self-reported changes in alcohol-related outcomes associated with use of Down Your Drink.
This was a pragmatic cohort study of the first 10000 people who registered to use the site, after the end of the pilot phase (ie, after September 2003).
The study was approved by the Camden and Islington Local Research Ethics Committee.
Down Your Drink was developed with support from the Alcohol Education and Research Council as a Web-based interactive program of brief interventions to reduce alcohol consumption in heavy drinkers and is hosted on a single dedicated Web site [
The animated home page invited visitors to assess their level of drinking by taking the Fast Alcohol Screening Test (FAST) [
Down Your Drink home page
The Web site content was structured as a set of six consecutive intervention modules designed to be accessed by the registrants at weekly intervals. Once a given module had been completed, access to the subsequent module was barred for 7 days. The 6-week program was based on the stages of change model [
Sample Down Your Drink page
An off-line advertising campaign was run in September 2001 to coincide with the launch of the Web site and the early part of the pilot study. All off-line advertising ceased at the end of 2001 (Multimedia Appendix 2). The site was also registered with the Yahoo search engine and was listed in the information pages of a number of UK-based health-related periodicals. By the time this study started (September 2003), there had been no off-line advertising for this Web site for over 18 months. All visitors to the Down Your Drink site were invited to complete an initial assessment using the FAST [
In order to participate, users were invited to read the policy on confidentiality, complete the consent procedure, and then choose a username and password. Those who agreed to register were required to submit information about their age, gender, marital status, family composition, ethnicity, occupation, country of residence, and how they had learned about the site.
The primary outcome measure was the 14-item Short Alcohol Dependency Questionnaire (SADD), which measures dependency on alcohol [
All the data for the study were collected automatically on a live database located on a secure dedicated Web site. Access to this database was by password only and was restricted to members of the research team. This database was initially launched simultaneously with the Down Your Drink Web site in October 2001, but was withdrawn at the end of the pilot phase for re-development. The upgraded database was subsequently re-launched in September 2003, and the data presented here are from the first 10000 users after the launch of the revised database.
Usage data were collected automatically on the Down Your Drink Web server and analyzed by a Web server log file analysis program (Webalizer, version 2.01). These data were stored separately from the Down Your Drink registrant database. The usage data reported here are from a subset of the total participants in the study collected between January and December 2004.
Data for the first 10000 users were extracted from the live database on March 1, 2006, and transferred into SPSS, version 12.0.1 for Windows via Microsoft Access, where all data related to individual Web site users were linked using the unique identifier created by the Down Your Drink site.
These data were then subjected to frequency analysis and paired
It took just over 27 months, from the launch date on September 24, 2003, to January 3, 2006, to complete recruitment of 10000 users.
The self-reported demographic characteristics of the sample are shown in
The great majority (83.9%) lived in the United Kingdom, and 9.3% reported living in other English-speaking countries (United States, Canada, Australia, and New Zealand). Over 100 countries of residence were given by the remaining 6.7% of users.
Demographic characteristics of Down Your Drink users
|
||
|
|
|
|
37.44 | 9.84 |
|
|
|
Male | 4891 | 48.9 |
Female | 5109 | 51.1 |
|
||
Single | 3754 | 37.5 |
Married or living with partner | 6246 | 62.5 |
|
||
Yes | 4244 | 42.4 |
No | 5756 | 57.6 |
|
||
Managerial/professional | 2579 | 25.8 |
Self-employed | 862 | 8.6 |
Administrative/secretarial | 854 | 8.5 |
Information technology | 770 | 7.7 |
Academic | 461 | 4.6 |
Housewife/househusband | 431 | 4.3 |
Unemployed | 353 | 3.5 |
All other | 3690 | 36.9 |
|
||
White British | 8185 | 81.9 |
White other | 911 | 9.1 |
White Irish | 527 | 5.3 |
Asian | 135 | 1.4 |
Mixed | 93 | 0.9 |
Black | 69 | 0.7 |
Other | 81 | 0.8 |
|
||
United Kingdom | 8385 | 83.9 |
United States | 554 | 5.5 |
Ireland | 172 | 1.7 |
Australia | 111 | 1.1 |
Canada | 104 | 1.0 |
Other | 674 | 6.7 |
Nearly three quarters of registrants (n = 7167, 71.7%) visited the Down Your Drink site from another Internet-based resource. Most of these connected via a link from another Web site (n = 4156, 41.6%) or from a search engine (n = 2900, 29.0%), whereas a smaller group responded to a banner advert on another site (n = 111, 1.0%). Relatively few registrants had been directed to Down Your Drink from the health service (n = 583, 5.8%).
The daily pattern of Down Your Drink use between January and December 2004 (represented by number of “hits” per hour) is shown in
Mean hourly usage of Down Your Drink (Jan to Dec 2004)
There was a high attrition rate from the program with only 1654 (16.5%) of the original 10000 registrants completing the 6-week program (
Number of users completing each week of the program. Percentages are the proportion of those registered, for example, 16.5% of all users that registered with Down Your Drink had completed the 6-week program by the time of the data extraction (March 1, 2006).
|
|
|
||||
|
|
|
|
|
|
|
Registered | 10000 | 100.0 | 4891 | 100 | 5109 | 100 |
Completed week 1 | 8933 | 89.3 | 4302 | 88.0 | 4631 | 90.6 |
Completed week 2 | 4020 | 40.2 | 1916 | 39.2 | 2104 | 41.2 |
Completed week 3 | 3006 | 30.1 | 1403 | 28.7 | 1603 | 31.4 |
Completed week 4 | 2411 | 24.1 | 1128 | 23.1 | 1283 | 25.1 |
Completed week 5 | 1928 | 19.3 | 887 | 18.1 | 1041 | 20.4 |
Completed week 6 | 1654 | 16.5 | 770 | 15.7 | 884 | 17.3 |
In order to determine whether there were demographic or clinical characteristics that were associated with completion of the 6-week program, we compared the characteristics of those who completed the sixth week with those who had only completed the first week. We chose first week completers as the reference, as we did not have data on the outcome measures from those who had dropped out before completing the first week. We defined anyone who had done some part of week 6 as a completer, whether or not they filled in the questionnaires at the end of week 6. It can be seen that female users, users who were married or living with a partner, and users without children were more likely to complete the program than men, single users, or users with children (
Comparison of baseline demographic and clinical characteristics of those who completed the six week program (completers) with those who only completed the first week of the program (starters)
|
|
|
|
|
|
|
|
|
0.04 | ||
Male | 4302 (48) | 770 (47) | |
Female | 4631 (52) | 884 (53) | |
|
< .001 | ||
Married or living with partner | 5646 (63) | 1125 (68) | |
Single | 3287 (37) | 529 (32) | |
|
.002 | ||
With children | 3808 (43) | 646 (39) | |
Without children | 5125 (57) | 1008 (61) | |
|
|
|
|
Years | 37.6 (9.8) | 38.9 (9.6) | |
|
|
|
|
SADD* | 12.31 (6.09) | 11.52 (5.24) | < .001 |
Abbreviated APQ† | 7.38 (5.01) | 6.83 (4.66) | < .001 |
Core functioning | 1.38 (0.82) | 1.35 (0.74) | .17 |
Core problem | 1.59 (0.92) | 1.59 (0.81) | .93 |
Core well-being | 1.62 (1.00) | 1.57 (0.92) | .06 |
Core risk | 0.41 (0.60) | 0.34 (0.50) | < .001 |
*Short Alcohol Dependency Data Questionnaire
†Alcohol Problems Questionnaire
Of the program completers, 57% also fully completed the outcome questionnaires at the end of the 6-week program. The mean scores for the SADD and modified APQ were significantly lower for these people at week 6 than week 1, indicating that alcohol dependency and alcohol-related harm were significantly reduced at the end of the program (
Change in clinical outcomes between week 1 and week 6 in users who completed the 6-week program
|
|
|
|
|
|
||
|
|||
Men (n = 421) | 11.51 (5.17) | 7.65 (4.51) | < .001 |
Women (n = 520) | 11.58 (5.35) | 7.64 (5.04) | < .001 |
|
|||
Men (n = 421) | 7.18 (4.74) | 3.43 (3.90) | < .001 |
Women (n = 520) | 6.61 (4.35) | 3.05 (3.66) | < .001 |
|
|||
|
|||
Functioning | 1.32 (0.73) | 0.87 (0.69) | < .001 |
Problem | 1.51 (0.81) | 0.94 (0.74) | < .001 |
Well-being | 1.41 (0.90) | 0.88 (0.84) | < .001 |
Risk | 0.31 (0.46) | 0.14 (0.33) | < .001 |
|
|||
Functioning | 1.37 (0.75) | 0.85 (0.72) | < .001 |
Problem | 1.66 (0.80) | 0.99 (0.80) | < .001 |
Well-being | 1.70 (0.91) | 1.00 (0.89) | < .001 |
Risk | 0.38 (0.55) | 0.18 (0.44) | < .001 |
This large pragmatic cohort study of users of a freely available online program for nondependent drinkers at risk of harm from alcohol use suggests that a small but significant (16.5% or 1 in 6) proportion of users will complete the 6-week program. Those that completed the program and provided outcome measures reported clinically significant benefit, with reduction in mean dependency and harm from alcohol, and improved mental health. Women, users with a partner, and users without children were more likely to complete the program.
Our findings add to the growing body of literature suggesting that behavior change can be achieved through online interventions [
This online service, in common with other studies [
This was a pragmatic cohort study aimed at exploring the usage patterns of a freely available online intervention. All data were self-reported, and we have no objective confirmation of any of the data reported here. Some users may have provided false or inaccurate data. However, there is no particular reason why registrants should have systematically lied at registration, and, given the large sample size (10000 registrants), the demographic data probably provide a reasonable description of the characteristics of users. Users who completed the 6-week course showed considerable motivation and commitment. The outcome measures were an integral part of the intervention, and users completed them entirely for their own benefit. It is difficult to know how truthful they were in completing these questionnaires, but there is some evidence to support the suggestion that responses to online questionnaires are comparable to traditional paper-based versions [
As this was an uncontrolled study, the data can only suggest an association between use of the program and an improvement in health outcomes. This study cannot determine whether this association was causal.
If confirmed, these data suggest that an online intervention aimed at nondependent heavy drinkers can make a useful addition to the public health armamentarium. Although only 16% of those who registered completed the course, in public health terms this still represents a significant number of people who could benefit. The advantages of such an intervention are that the costs are unaffected by the number of users and that the intervention can be used at home, at any time of day or night, unlike more traditional services. Future work should explore whether adapting the program, including increasing the options for flexible use, removing the requirement to work through the elements sequentially, and introducing more personalized feedback, can improve “stickiness” or adherence to the program, and whether the association between use of the program and improved outcomes is causal. We are currently undertaking a randomized controlled trial to explore both these questions.
The development of Down Your Drink was supported by grants from the Alcohol Education and Research Council and the Department of Health.
None declared.
Screenshots of the Down Your Drink program
Recruitment leaflet sent to health care professionals
Participant information sheet
Alcohol Problems Questionnaire
Clinical Outcomes in Routine Evaluation-Outcome Measure
Fast Alcohol Screening Test
Greenwich mean time
Short Alcohol Dependency Questionnaire