Tobacco smoking is one of the major causes of mortality and morbidity among former illicit drug users [1]. The prevalence of smoking among patients on methadone maintenance treatment is very high; nearly 90% of patients who visit methadone clinics smoke tobacco [2]. At the same time only 1 in 3 US methadone maintenance facilities provides smoking cessation counseling for their patients [3]. For the last decade, however, the situation in this field has improved, as more research and organizational efforts have been applied to enhance smoking cessation among methadone users [4-6].

Computers can be a powerful and cost-effective means of providing health education [7]. The necessity for some computer experience, however, was seen as one of the major obstacles to introducing more sophisticated approaches to smoking cessation, such as interactive computer programs and the Internet [8-11]. Methadone-treated patients are frequently from the low-educated, low-income strata of society. Their computer and general reading literacy is often poor, and they usually do not have access to the Internet [12]. To date, the effectiveness of computer-mediated approaches for promoting smoking cessation has not been systematically studied in this population. The potential of interactive, web-based interventions for health education and counseling in this population is unknown. In addition, little systematic information regarding the smoking decision balance of methadone tobacco users is available in the current literature. No data is available about the possible effect of methadone-treated smokers’ socio-demographics, computer literacy, Internet use, and smoking-related behavioral constructs on their ability to successfully use computer-mediated hazards of smoking education. Such information is necessary for developing a targeted anti-smoking, multi-component, computer-mediated program. In this paper we describe development and assessment of an interactive educational module of a multi-component, smoking cessation, counseling computer program for former illicit drug users treated in a methadone outpatient clinic.

The main aims of this study included: (1) development of theoretical framework for computer-mediated hazards of smoking education guided by adult learning theories; (2) implementation of the interactive education program using tablet PC; (3) assessment of the feasibility and patient acceptance of the educational module in methadone-treated patients; (4) collection of systematic information on smoking profiles and attitudes of methadone tobacco users for developing a targeted multi-component, computerized, smoking cessation counseling system; and (5) establishment of possible factors facilitating or impeding successful computer-mediated education in these patients.

Learning theories have been shown to improve significantly the efficacy of educational software [13,14]. Conventional education means, such as lectures, seminars, workshops, books, and videos already incorporate, more or less successfully, practical approaches to learning developed over centuries. Computer-mediated education is an interactive tool, and many approaches that are used intuitively in other spheres may not be applied without a clear understanding and formulation. A computer program has no ability to summarize, repeat, provide feedback or give an additional example if this capacity is not specified and implemented in advance. When designing our learning program, we reviewed over 50 of the most frequently cited theories of adult learning and found that only some of them can be used for constructing a computer program, because many of the theories have no clear experimental support and are applicable only for certain subjects or could be used only under specific conditions or only for a certain part of the learning process. The reviewed theories can be grouped into 4 domains presented in the Table 1.

For the purpose of this study, the most important attributes of the existing learning theories were not the general psychological assumptions underlying each theory, but their applicability to the development of interactive educational software. Our analysis of different learning theories resulted in 10 main principles in accordance with which we designed our educational program. These principles are described below.

The study sample consisted of 23 women (65.7%) and 12 men, 45.4 ± 6.7 years old, with an average 11.2 ± 1.7 years of education. Most of the patients were unemployed and had income of $20,000 a year or less (Table 2). Most of the patients (21/35, 60.0%) had never used computers in their lives.

The patients had smoked, on average, for 23.2 ± 10.2 years. Those patients who had discussed their smoking at least once with their doctors numbered 20 (20/35, 57.1%), although it was not connected with their knowledge about hazards of smoking or their stages of change. More than half of the patients considered smoking to be a severe problem for them, while only 1 patient thought it was not a problem at all. The total smoking self-efficacy score was 27.3 ± 12.4, which corresponds to a low perceived ability to abstain from smoking. According to Prochaska’s Stages of Change Scale, 8 patients were in the preparation stage (8/35, 22.9%), the majority were in the contemplation stage (16/35, 45.7%), and the remaining patients were in precontemplation (10/35, 28.6%).

The odor of smoke was the most frequently reported unpleasant effect of smoking (12/35, 34.3%), along with breathing and lung problems (Table 3). The main trigger to smoke was feeling nervous or depressed (25/35, 71.4%). About 20% of participants (7/35) expected to become nervous or depressed after quitting. Feeling relaxed or calm was the most frequently reported benefit of smoking (23/35, 65.7%). No desire to quit was the most frequently cited barrier to quitting, reported by 6 patients (6/35, 17.1%) (Table 3).

Although the questions asked were very basic and simple, the average knowledge score at pre-test was 60.5 ± 16.3 points (corresponding to the percentage of correct answers). Participants scored lower than the total average for questions under category of ‘Nicotine addiction,’ and ‘Quitting smoking.’ Examples of quiz statements the validity of which many patients could not assess correctly for ‘Nicotine addiction’ include ‘Nicotine is an addictive chemical’ and ‘Nicotine causes the arteries to loosen and allow more blood flow’; for ‘Quitting smoking’, examples are ‘Quitting smoking starts to benefit your health about 1-2 years after quitting’ and ‘If you smoked for more than 10 years quitting won’t help’. As a result of computer-mediated education, participant knowledge about the hazards of smoking significantly increased to 70.4 ± 11.7 points, P < .001. The increase in knowledge was not associated with the patients’ level of education or computer experience. In 3 major curriculum topics out of 4, the increase in knowledge scores was statistically significant (Table 4).

In order to establish factors facilitating or impeding successful computer-mediated education, analysis of variance (ANOVA) was employed. Considering the study design as a 1-factor experiment, ANOVA was used to test for differences between pre- and post-test knowledge scores controlling for sociodemographics, computer experience, smoking history, stages of change, and number of perceived inhibitors and facilitators of smoking cessation. The education was stratified into ‘less than high school’ and ‘high school or more’ groups. Age was categorized as < 45 and ≥ 45 groups. The income was divided into 3 groups: no income, < 20K, and ≥ 20K. Job status was stratified into ‘employed’ and ‘no job’ and the amount of smoking was categorized as ‘2 packs a day or more’, ‘1 pack a day’, ‘1/2 pack a day’, and ‘1/2 pack a week.’ Frequency of using the Internet or a computer was categorized into 2 groups: ‘never’ and ‘use at least once a month’. The stages of quitting smoking were categorized into 3 groups: pre-contemplation, contemplation, and preparation. Patients were asked to list any factors that inhibit and facilitate their intentions to quit smoking. The number of these factors was introduced into the model, which were stratified into 2 groups using their median value as a dividing point (the number of inhibitors: 1-7 and > 7, the facilitators: 1-6 and > 6). After controlling for all these variables, the difference between pre- and post-test knowledge scores remained significant (P = .004). The variables which significantly affected knowledge gain were gender (males were more likely to have a higher knowledge gain than females, P = .015), age (people over 45 were less likely to have a knowledge gain, P = .02), stage of change (subjects in contemplation were more likely to have higher knowledge gain, P = .03), and the number of facilitating factors to quit smoking (patients with more facilitators were more likely to study more successfully, P = .002). Patient education level, computer experience, and Internet use did not affect the results of computer-assisted education.

Acceptance of the program interface according to the Attitudinal Survey is presented in Table 5. The program was very well accepted by the overall participants. One patient did not complete the attitudinal survey and another patient missed filling in questions #9 and #18. Therefore, percentages for questions #9 and #18 are calculated based on 33 participants and the rest are calculated based on 34 participants. For 78.8% (26) of patients, using the computer was not complicated at all. Most of the patients (32 persons, 94.0%) rated the program as good or excellent. The majority of patients had little or no problem understanding the presented information, but 27.3% (9 patients) encountered very significant amounts of unknown words. About 18% (6) of patients claimed that they encountered information which was difficult to understand, and 11.8% (4) frequently found it confusing (Table 5).

Patients’ feedback about their learning experience was ascertained using semi-structured qualitative interviews. Except for a short training session (~ 10 minutes) provided by a research assistant, 28 (82.4%) participants did not need any help during the educational session. One participant did not complete the qualitative interview; therefore percentages are calculated based on 34 participants. Only 7 (20.6%) patients felt tired at the end of the educational session, and 5 (14.7%) believed the educational session was too long. Regarding ways to improve the program, 7 (20.6%) patients thought they would prefer to listen to the educational messages; 18 (52.9%) thought they would like to both read and listen to the information; and 27 (79.4%) thought that including more video clips would make the program better. For 31 (91.2%) patients it was amenable to answer a multiple choice question after each educational message; 26 (76.5%) were compliant with the conditional self-quiz design of the educational session (ie, patients had to demonstrate sufficient knowledge of a module before being allowed to move to the next one); and 24 (70.6%) patients felt “it was ok to repeat the whole section again.” Finally, more than half of the patients (20, or 58.8%) preferred using a computer program to learn about smoking over more traditional types of education, such as brochures, videotapes, healthcare providers, and browsing the Internet.

Computer-assisted patient education utilizing the main concepts of adult learning theories was successfully implemented in methadone-treated smokers. The results showed statistically significant increases in the subject knowledge scores. This result remained statistically significant after adjusting for major sociodemographic factors, smoking profile, and behavioral factors. The majority of the patients were able to navigate successfully the user interface even though most of them had never used computers before. These results corroborate our previous findings in which we showed that even low-income individuals with limited education and no previous computer skills were able to navigate successfully an educational computer program after 5-15 minutes of supervised training, when specific user interface principles were implemented [38,39,56].

Former drug users on maintenance methadone treatment are quite different from other groups of patients who need smoking cessation interventions. They are often depressed and have other psychiatric problems including chemical and non-chemical dependencies. Our research team is in the process of the development of a comprehensive computerized smoking cessation program. Testing of the educational module of this program was done to get feedback from patients about its acceptability and design, and also to elicit additional information about their smoking profiles. After only a brief training session, the patients, most of whom were using computers for the first time in their lives, were able to go through the educational program. Our education intervention was successful, and the program installed on tablet PCs was very well accepted by the patients.

Learning theories cannot serve as a universal recipe or magic pill to improve patient education [57]; rather they should be used and applied thoughtfully and selectively [13]. In designing our educational program we used a mix of ideas derived from different theories. Though this approach prevented us from assessing which particular ideas mostly contributed to outcomes, we thought that for the success of the intervention it was more important to provide a theoretical framework incorporating a broad spectrum of ideas and, therefore, allowing inclusion of a more diverse population. Other studies tested applicability and efficiency of some specific theories to patient education, such as the problem-based approach [58]. Using a combination of approaches can potentially be more effective when target users are represented by diverse clientele in various settings where education is implemented [59-61].

When constructing the program to install on tablet PCs we also took into account the low educational level of the patients in methadone outpatient clinics. Such design details as using self-explanatory error-proof navigation, large fonts, and audio support substantially improved the usability of the program and were highly valued by the patients.

Patients in our sample lacked basic knowledge about hazards of smoking and smoking cessation, and underestimated the negative health effects of smoking. Only 6 participants recollected such long-term consequences of smoking tobacco as lung cancer, and only 1 patient was worried by the possibility of heart problems caused by smoking. This supports the previous data [62] that former drug abusers frequently underestimate the risks of smoking. The odor of smoke was the most frequently reported negative side effect of being a smoker. Only 4 (11.4%) patients thought that getting help from others or taking some kind of medication could aid in coping with smoking urges.

Our data agree with Nahvi et al [6] that methadone users are interested in smoking cessation. In that study, nearly half of the smokers were in the contemplation stage, and about 20% were in the preparation stage, corroborating our results. These results are very interesting, when one considers that these people are already struggling with at least one kind of addiction, have numerous psychiatric comorbidities, are subjected to a lot of stress in their lives, and are frequently unemployed. Our data helps with understanding why it can be so difficult for these patients to decide to quit. Depression and nervousness are seen as major consequences of quitting, while at the same time these feelings are triggers for the desire to smoke.

In this study the main outcome was knowledge gain in hazards of smoking. We did not expect after only a single brief intervention to motivate patients to decide to quit smoking. We see the computer-assisted hazards of smoking education as a component of a comprehensive smoking cessation program, including computerized and/or in-person counseling. Tailoring computer-mediated smoking cessation counseling to a certain stage of change, gender, cultural background, and smoking profile can facilitate smoking cessation.

The content of our educational curriculum was simplified by adjusting the readability of the content to the fifth-grade level. Despite this, according to the attitudinal survey, reading comprehension was one of the major problem areas in this pilot study. Readability and accessibility can be an issue for a variety of consumer health applications, and there are well established approaches on how to measure and avoid this problem [63,64]. We think that this group of patients requires additional measures to facilitate learning (eg, audiovisual aids and significantly simplified grammar). Most of the patients were not used to studying and had no previous computer experience. Despite that, most of them were not tired at the end of the educational session. Our initial concerns that the ‘exam-like’ format of the program may not be the best fit for this group of patients were not supported by our findings. An absolute majority of the patients accepted multiple-choice questions and quizzes very well. When asked, patients preferred computer-assisted education to other conventional means of education.

Different computer-mediated smoking cessation approaches have been shown to be successful. These include computer-generated individually tailored letters [65] and internet-based smoking cessation programs [66-69]. Educational computer programs about the hazards of smoking can be used separately or as a part of smoking cessation computer intervention in outpatient drug treatment facilities. The same model can be applied to other types of health behaviors, such as healthy nutrition, alcohol drinking, condom use, and physical activity, all of which also constitute significant problems for former drug users [70,71]. Internet-enabled, touch-screen computers could be easily utilized for a widespread dissemination of computer-assisted health education in methadone clinics.

When analyzing which factors influence knowledge gain in the sample, we found that age, gender, stage of change, and number of facilitators (potentially beneficial consequences of quitting smoking for patients) influenced it. Neither computer/Internet experience nor the level of education was significant for patient ability to learn. We may conclude that the program was simple and effective enough for the patients independent of their education and computer skills, but patients who perceived smoking cessation more positively demonstrated a higher a knowledge gain. Our results supported the notion that adult learning theories could provide an effective framework for successful computer-mediated education in a group as challenging as methadone-treated smokers.