Participants
A quasi-experimental design was used to recruit participants. We invited 14 junior high schools to participate in the study through the local Department of Education in seven counties and cities. We recruited two schools in each county/city and randomly assigned them to experimental and comparison groups. We invited the health education teachers of the experimental schools to attend an orientation meeting and introduced the purpose and methods of the study. After obtaining the health education teacher’s permission to participate, we invited students from two classes of these schools to enroll in the study. Students and their parents/guardians provided written consent forms. An identical procedure is carried out for the control school group. Students or parents who did not provide written consent were not included in the study. All students and parents/guardians were informed of their right to participate and were assured that students’ health education grades were not contingent on participation. The final sample comprised 323 and 325 students in experimental and comparison groups, respectively.
Procedure and Program Delivery
A flowchart outlining participant enrollment and assessments is presented in Figure 1. After selecting the seven experimental schools, the research team approached the principal and health education teacher of each school to explain the research purpose, method, and protocol. After obtaining permission to conduct the study, we delivered recruitment messages to invite students to participate in this study, and scheduled an orientation meeting to ensure that the health education teachers could fully understand the purpose of the study and the cooperation works. Subsequently, we provided a half-day workshop to introduce the drug-use prevention program (Table 1). The program was developed by a professional team including professionals in health promotion and health education, drug-use prevention professionals, nursing, and social workers. Teachers were strongly advised to use numerous interactive teaching methods including questions and answers, brain-storming, story-telling, case discussion, situational role-playing, game playing, watching an animated film, and to follow the discussion, value clarification, modeling, and skills practice exercises during the program implementation. To promote health literacy acquisition, we also provided program worksheets and a parent-child workbook to increase the potential practices of health literacy in daily life. Animated films, E-games, case stories, worksheets, and role-playing were used to increase health literacy learning experiences.
Before program implementation, the relevant materials and worksheets were implemented in a junior high school as a pilot test to ensure their appropriateness. The program consisted of ten 45-minute sessions. The sessions were delivered in a health education class, morning study time, and during the class meeting time, which was arranged by the health education teacher and school administration. The 10 sessions were completed within six months. A structured self-reported questionnaire was administered to students at baseline and the end of the program by research staff blind to the students’ group status. The principal investigator supervised the teachers during program implementation to ensure fidelity to program design. A regular monthly meeting was scheduled to support health education teachers’ program delivery (Figure 2, 3).
Instruments
The four TPB variables including attitude, subjective norm, PBC, and intention to not use illegal drugs were modified from the previous study with permission [15]. Demographic variables consisted of gender, parents’/guardians’ education level, parenting style (authoritative vs. democratic), household status (living with both parents, living with a single parent, and others), and lifetime substance use (smoking, drinking, and betel nuts chewing).
Drug-use-related health literacy measures students' ability to access and understand information and resources of substance use prevention and apply them to make the right decisions to maintain and promote their health [20].This scale consists of 14 Likert-type items, ranging from 1 (strongly disagree) to 5 (strongly agree). Items appear as five groups based on the logic flow of five presented drug-use paragraphs, respectively.
A sample paragraph is “When the school had activities, Leo heard the discipline director say to all students: In recent years, Taiwan's illegal drug-use has escalated according to the news, and the average age of users has gradually decreased. Students should pay attention not to go to at-risk environments such as Internet cafes, billiards rooms, and home parties, to prevent exposure to illegal drugs in the community. If you have family and friends with a drug use problem who need help, please dial 0800-775-885.” The three follow-up items after that paragraph are “If I were Leo, I would reduce my access to at-risk places;” “I know what kinds of places are ‘at-risk environments’ that may expose me to illegal drugs;” and “I know that the ‘special line for successful detoxification’ is 0800-770-885, and I will support my family and friends in need of those resources.” The higher the score, the higher level of drug use-related literacy. The Cronbach’s α coefficient was 0.86 in this study. Exploratory factor analysis extracted only one factor, which could explain 53.86% of the variance.
Attitude was measured using four pairs of evaluative bipolar adjectives (pairs of opposite terms) to assess students’ positive or negative evaluations and feelings regarding illegal drug use. Each item was scored on a Likert-type scale with a reversed score scale of 1–5, with higher scores indicating a higher level of agreement on not using drugs. A sample item is “To me, drug use makes me feel joyless/joyful.” Exploratory factor analysis extracted only one factor and the factor could explain 75.47% of the variance. The Cronbach’s α coefficient was 0.80 in this study.
Subjective norm was measured by five items using a five-point Likert-type scale. Each item was scored from 1 to 5 with higher scores indicating a higher level of significant others’ agreement on not using drugs. A sample item was “My teachers don't think I should use drugs.” Exploratory factor analysis extracted only one factor and the factor could explain 66.83% of the variance. The Cronbach’s α coefficient was 0.93 in this study.
Perceived behavior control was measured using two items rated on a five-point Likert-type scale. Each item was scored from 1 to 5 with higher, scores indicating a higher level of students’ confidence in not using drugs. A sample item is “I am confident I won't use drugs.” Exploratory factor analysis extracted only one factor and the factor could explain 85.88% of the variance. Cronbach’s α coefficient was 0.83 in this study.
We used the intention not to use drugs as a proxy measure for drug-free behavior because most students had not used any drugs before. The intention not to use drugs was measured using three items rated on a five-point Likert-type scale. Each item was scored from 1 to 5 with higher scores indicating a higher level of students’ agreement on not using drugs. A sample item is “I would not like to use drugs.” Exploratory factor analysis extracted only one factor, which could explain 86.48% of the variance. Cronbach’s α coefficient was 0.93 in this study.
Statistical analysis
SPSS version 22.0 was used for the Descriptive analyses of the demographic and outcome variables. Chi-square tests were used to compare percentages on the demographic status between the experimental and comparison groups. The group comparisons of outcome measures at baseline were determined by performing Hotelling’s T2 to avoid inflating type I error. A generalized estimating equation (GEE) was used to investigate the effects of time, groups, and their interactions on the outcome variables; GEEs enabled understanding the patterns of the time change and the effects at both the individual and group levels.
After the first-round analysis, we conducted further analysis to explore the intervention effects for the drug-use-related function, and communicative/interactive and critical health literacy. This further analysis is meaningful because we wanted to explore whether the life-skills training and 4 Fs (Facts, Feeling, Finding, and Future) teaching methods designed for health literacy were effective or not.