Web-based screening and brief intervention for the ... - CiteSeerX

Preventive Medicine 39 (2004) 969 – 975 www.elsevier.com/locate/ypmed

Web-based screening and brief intervention for the spectrum of $ alcohol problems Richard Saitz, M.D., M.P.H., a,b,d,* Eric D. Helmuth, M.A., c Susan E. Aromaa, M.S., c Anara Guard, M.S., c Marc Belanger, M.A., c and David L. Rosenbloom, Ph.D. c,d a

Clinical Addiction Research and Education (CARE) Unit, Section of General Internal Medicine, Department of Medicine, Boston University School of Medicine and Boston Medical Center, Boston, MA 02118, USA b Department of Epidemiology, Boston University School of Public Health, Boston, MA 02118, USA c Join Together, Department of Social and Behavioral Sciences, Boston University School of Public Health, Boston, MA 02118, USA d Center to Prevent Alcohol-related Problems Among Young People, Boston University School of Public Health, Boston, MA 02118, USA Available online 2 June 2004

Abstract Context. Many persons who drink excessively remain unidentified and do not receive interventions. Screening and intervention using the World Wide Web could make such services more accessible and therefore more widely used. Objective. To evaluate the use of a novel alcohol screening and brief intervention Web site. Design. A Web site was developed, posted, and its use was evaluated. We analyzed a sample of visitors who completed alcohol screening over a 14-month period to describe their alcohol use, and their use of portions of the Web site that provide information and referral resources. Setting. The Internet. Patients or other participants. Web site visitors, with a focus on visitors who completed an alcohol-screening questionnaire about their own drinking. Intervention. Brief intervention via the Web site, consisting mainly of feedback, advice, and a menu of change options and referral information. Main outcome measures. Self-reported drinking amounts and alcohol screening test scores, and utilization of Web site components. Results. Visitors completed online alcohol screening questionnaires at a rate of 50,711/year of 115,925 visitors/year. In a 14-month period, 39,842 adults completed the questionnaire about their own drinking habits; 66% were men, 90% reported drinking hazardous amounts (per occasion or typical weekly amounts), 88% reported binge (per occasion) drinking, and 55% reported typically exceeding weekly risky drinking limits. Most (65%) had alcohol screening test results (AUDIT z 8) consistent with alcohol abuse or dependence; similar proportions of women and men were hazardous drinkers. One-fifth of visitors visited portions of the Web site that provided additional information about alcohol use and referrals. Visitors with possible alcohol abuse or dependence were more likely than those without these disorders to visit a part of the Web site designed for those seeking additional help (33% vs. 8%, P < 0.0001). Conclusions. A well-publicized, easily accessible, research-based screening and intervention Web site can attract many users, most of whom are drinking excessively, and many of whom avail themselves of referral information after receiving individualized feedback. D 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved. Keywords: Screening; Alcoholism; Alcohol use; Internet; Brief intervention

Introduction $ This work was presented in preliminary form at the meeting of the American Public Health Association in Philadelphia, PA on November 13, 2002. * Corresponding author. Clinical Addiction Research and Education (CARE) Unit, Section of General Internal Medicine, Boston Medical Center, 91 East Concord Street #200, Boston, MA 02118-2393. Fax: +1617-414-4676. E-mail address: [email protected] (R. Saitz).

Alcohol use disorders are costly and a leading cause of disability and death worldwide [1,2]. Brief screening tools can identify people with alcohol problems, and, once identified, those people can receive brief interventions [3,4]. But many people with alcohol use disorders do not seek care nor are they screened even when they contact the health care system [1,5,6]. Barriers to seeking care include

0091-7435/$ - see front matter D 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2004.04.011

970

R. Saitz et al. / Preventive Medicine 39 (2004) 969–975

lack of motivation to change, beliefs about treatment, attitudes, fear of discrimination or job loss if detected, and lack of perceived effective access to care [7 – 9]. Active screening followed by intervention can improve patient outcomes for problem drinkers, and improve access to specialty treatment for those with dependence [3,10]. Despite these available options, many people remain unidentified and untreated [11,12]. Most screening and brief intervention programs have been designed for health care settings, usually paper and pencil questionnaires or in-person interviews [3]. Web-based measures of alcohol use appear to be at least as reliable as these older methods [13]. Using the Internet for screening has the potential to greatly increase the number of people screened and improve access to brief intervention. More than half of all U.S. adults use the Internet and more than half of these Internet users search for health information online (77 million people); 8% of Internet users (9 million people), 14% of young adult Internet users (18 – 29 years old), and 7% of middle-aged adult users (30 – 49 years old) have searched for information on alcohol or drug problems [14]. AlcoholScreening.org provides online self-assessment tools and health-based information to help individuals identify their own risky drinking patterns or current alcohol problems. The site delivers personalized feedback and helps users locate assistance if they are ready to seek help. In this paper, we describe the development of the Web site designed to anonymously screen adults and provide personalized online feedback regarding alcohol use, and the feasibility of its use for alcohol screening on the Web. After creating and posting the Web site, we tested whether this Web site was used widely, whether it reached a target population of hazardous drinkers, and whether these persons would seek further assistance from the site when appropriate.

Methods Web site design AlcoholScreening.org is an anonymous, free online selfscreening service to assess an individual’s alcohol consumption and its consequences. AlcoholScreening.org was based on the health belief model [15]. According to the core concepts of the health belief model, risky drinkers are more likely to reduce their alcohol consumption or otherwise control their at-risk behavior when they believe the threats are real, the benefits of change are valuable, and the barriers to behavioral changes are lessened. With these concepts in mind, AlcoholScreening.org was created as a vehicle to provide the user with personalized feedback speaking directly to the issues of risks, benefits, and action steps. Additional features were created to provide supplemental information for Web site users wishing to take action. Visitors to this Web site answer 12 questions about their drinking and an additional question for research purposes,

and are presented with results that outline the likelihood that their reported drinking patterns indicate risky or harmful alcohol consumption. Visitors may also access an online library of health information about alcohol consumption and alcohol problems, search a national database of substance abuse treatment facilities, or follow links to additional information online. All users are presented with a disclaimer emphasizing that the Web site does not provide a medical diagnosis and cannot substitute for a full evaluation by a health professional. Visitors complete an online version of the Alcohol Use Disorders Identification Test (AUDIT) [16]. In addition to the 10 AUDIT questions, users are asked two additional questions to further detail the quantity and frequency of their alcohol consumption, and whether their responses reflect their personal alcohol consumption patterns. These questions read: 

‘‘Thinking about a typical week, on how many days do you have at least one alcoholic drink? (If you don’t drink every week, answer for a typical week in which you do)’’;  ‘‘Thinking about the past year, what is the greatest number of drinks you’ve had on any one occasion?’’; and  ‘‘Optional: This question is for research purposes only: I am completing this test based upon my own alcohol-use experience OR I am just curious about the test and the related feedback, or answered the questions with someone else in mind.’’ Upon completion of the online screening test, users receive nonjudgmental feedback based on their AUDIT score [17], the current U.S. Department of Agriculture Dietary guidelines for moderate alcohol consumption [18], and U.S. alcohol consumption norms [19]. Feedback is based on the principles of successful brief interventions [20] and distinguishes between drinking amounts that place people at risk for future consequences, and drinking with consequences that have already occurred [21,22]. All users are advised whether their screening results indicate a likelihood of hazardous or harmful alcohol consumption, and all are informed about current guidelines for low-risk drinking. In addition, links are provided to alcohol and health information, and the site provides the option of searching a national database of substance abuse treatment facilities for local, in-person assistance. A traffic light theme is graphically incorporated into the site to help illustrate the screening results. Users scoring below eight on the AUDIT and whose alcohol consumption falls within the U.S. Dietary guidelines for moderate alcohol consumption are presented with a ‘‘green light’’ results page. They are told that their alcohol consumption appears to fall within healthy limits, but are warned that there are certain circumstances when any amount of alcohol may not be safe, for example, when operating a vehicle or machinery, while pregnant, or if certain medical conditions exist.

R. Saitz et al. / Preventive Medicine 39 (2004) 969–975

Those scoring below eight on the AUDIT, but who exceed the U.S. Dietary guidelines for weekly or per-occasion alcohol consumption see a ‘‘yellow light’’ results page. They are told that while their results do not suggest that a pattern of excess drinking is currently harming their health, the amount they reported consuming on at least one occasion increases their risk for injury or other immediate consequences, or that the amount they reported consuming per week places them at risk for future, mainly chronic, health consequences. They are encouraged to cut down, abstain, or set a safer personal limit for how much alcohol they consume, and, as in the ‘‘green light’’ scenario, are reminded that for some people and in certain situations, no amount of alcohol is safe. In addition, those who exceed weekly alcohol consumption guidelines receive normative information comparing their alcohol consumption to that of the general adult American population and to the adult population of their gender. A sample of this information based on the data provided by a user is: ‘‘More than 95% of the general adult American population, and 91% of men consume fewer drinks per week than you reported consuming.’’ Users scoring eight or above on the AUDIT are presented with a ‘‘red light,’’ and are told that likely their current drinking is hazardous or harmful to their health and well being. This group also receives normative feedback comparing their alcohol consumption to that of the general adult American population and to the adult population of their gender. They are told that the AUDIT cannot diagnose any condition or tell them for certain if alcohol use is harming their health, and are advised to seek further evaluation from their doctor or other qualified health professional. At any time, users can follow links to a national alcohol treatment facility database and information on alcohol and health. The treatment database is a current copy of a data set maintained by the Center for Substance Abuse Treatment (CSAT) of the federal Substance Abuse and Mental Health Services Administration (SAMHSA). Further information on alcohol and health is republished by AlcoholScreening.org and provided by U.S. Department of Health and Human Services (U.S. Department of Agriculture and the SAMHSA Center for Substance Abuse Prevention) and the National Institute on Alcohol Abuse and Alcoholism (NIAAA) of the National Institutes of Health (NIH) [18,23,24]. AlcoholScreening.org was launched in April 2001 and has been promoted to the general public in both online and offline modalities. In May 2001, it was featured on hundreds of commercial Web sites in a month-long national banner-ad public service campaign through the DoubleClick advertising network, and received donated banner ad placement on the Boston Globe Web site. The site was featured as a resource on the television program CNN Presents in April 2002 (the month during which National Alcohol Screening Day occurs) [25], and has been linked from online news stories at WebMD,

971

CNN.com, and MSNBC.com. Printed flyers (34,000) were distributed to the public through state and local health departments, alcohol treatment providers, and recovery organizations throughout the United States, and also to every employee of a major U.S. airline through May 2002. The site receives hundreds of daily referrals from Internet search engines (e.g., google.com). Data collection Data are collected anonymously by the Web site and cannot be traced to any identifiable individuals. Users are asked to provide their age and gender, but no further personal information is solicited. The responses to all screening questions are recorded to a secure database, as are the results of the real-time computer scoring that determines which feedback module is presented to the user. If, after viewing their screening feedback, a user immediately accesses either the ‘‘Learn More’’ or ‘‘Find Help’’ sections, their choice is recorded and associated with the screening responses and results. The user’s age, gender, screening responses and results, and choice to access the ‘‘Learn More’’ or ‘‘Find Help’’ sections are linked with a record number and stored on a secure password-protected server. The study of Web site use was approved by the Institutional Review Board at Boston University Medical Center. In addition to recording and reporting herein counts of Web site visits, page views, and completed AUDIT screening tests, we describe a sample of data entered by users between April 2, 2001, and June 2, 2002. These data were entered by users who reported they were adults 18 years or older and who completed the questionnaire. From this description, we excluded Web users who stated that they were ‘‘just curious’’ or that they were answering the questions for someone else. Results are reported as descriptive statistics (means and standard deviations, proportions), and the chi-square test and t test were used as appropriate for bivariable analyses. We used analysis of variance and Duncan’s multiple range test using P = 0.05 as the level of statistical significance to compare mean age across three drinking levels. For descriptive purposes, we defined hazardous drinking amounts consistent with NIAAA and U.S. Department of Agriculture recommendations: >14 standard U.S. drinks (13.7 g ethanol) per week for men; >7 drinks per week for women and those 65 years or older; or >4 drinks per occasion for men, >3 drinks for women. We defined those receiving an AUDIT score of z8 as ‘‘possible alcohol abuse or dependence,’’ those receiving AUDIT scores