Project Development and Implementation. Three types of studies were accomplished in Year One of Project TND (1992-1993), the assessment interview (and refuse analysis and naturalistic observation), comparison evaluation study, and theme study. These studies provided empirical background for curriculum-constructing activities in Year Two. These studies indicated a clear preference for novel versus social influence activities. In particular, a "talk show" delivery modality which discussed the effects of drug use addiction and drug use and driving, and two cognitive perception-type activities (attitudinal perspective taking and stereotyping) were preferred. Of the social influence-type activities, decision making and effective communication lessons were preferred.

Three types of studies were accomplished in Year 2 (1993-1994), component study, school-as-community development, and pilot study. These studies provided empirical background for main efficacy trial implementation activities in Year 3. The component study indicated that, for future work, we could develop a self-instruction delivery modality of the program that would be rated of equal quality to a health educator-delivered program. However, when both types of sessions were placed in the same program in the pilot study, it was quite evident that all lessons should be classroom-oriented. Previous self-instruction lessons, while well-received when provided along with other self-instruction material and compared with classroom instruction as a between-groups variable (in Year 2, Study 1), were not as well-received as classroom-oriented material when both were offered in the same curriculum. The final health educator-delivered curriculum delivery system consisted of three lessons per week. The first trio of lessons helps the student to obtain an open mind to further material (listening, stereotyping, myths and denial), the second trio of lessons instructs students in chemical dependency issues (chemical dependency, talk show, stress-coping), and the third trio of lessons motivates alternative means of coping (self-control, perspectives, decision-making and commitment).

The pilot study also indicated that Associated Student Body groups (ASBs) can be formed and maintained as cohesive units, and that activities can be planned and implemented through these groups. The size of the ASBs could range probably from 5 to 20 stable student members. An increased emphasis on relevance to drug abuse prevention also was judged needed to help keep a drug abuse prevention activity focus at these meetings. The need to make at least half of the activities more than one-shot events also was emphasized, and weekly meeting forms were stream-lined to ease their completion.

All Year 3 studies were conducted as planned (1994-1995). Three types of activities were accomplished in Year 3, classroom-based curriculum component implementation, school-as-community component implementation, and baseline, process, and immediate posttest evaluation assessments. These activities, taken together, composed the main efficacy trial implementation. The nine-lesson TND curriculum was successfully implemented at all 14 program schools. A total of nine project health educators were arbitrarily assigned to schools based on availability and demand, but health educators were counterbalanced across conditions. Health educators were administered an extensive 3-week training. Delivery was conducted throughout the school year. Those students who were absent on days that a lesson was implemented were provided with single-page summaries of the material from each lesson which they could utilize as a means to "make-up" learning of missed lesson material. Implementation was successful and process evaluations were positive. Knowledge scores at posttest indicated discriminate validity between the experimental and control conditions.

The School-as-community Component was completed at 7 schools, which included the implementation of weekly ASB-TND meetings for six months, implementation of six events per school, and distribution of a Continuation Community Newsletter across these 7 schools. Events completed include job training, sports participation, self-defense training, drug-free parties, drug awareness weeks, canned food and clothes drives, drug prevention instruction for elementary youth, drunk driving/drug-related movie activity, prevention of unsafe sex and drug use, law enforcement and Alcoholics Anonymous speakers, and contribution to the newsletter. Banners were provided ("Check it out. Drug-free event") to help schools maintain inclusion of a drug use prevention-related focus at each event.

Recruitment of Schools and Program Evaluation. A total of 22 school districts from a five county region of southern California were recruited for participation in the study in a procedure approximating random selection. Each district contained one continuation high school. The smallest sized school was dropped from the set of main efficacy trial schools, and was used in piloting tasks. Thus, a total of 21 of these schools were blocked by risk factors such as estimates of drug use prevalence and achievement testing scores using archival data and were randomly assigned by block to one of three experimental conditions - standard care (control), classroom program, and classroom program plus school-as-community. An immediate pretest (baseline assessment) was followed by a three week long drug abuse prevention program at 14 continuation high schools (7 schools received a semester-long school-as-community component as well as a classroom program whereas 7 did not), and 7 other schools received only the pretest (standard care). Posttest collection at all schools was conducted one-month following the pretest.

Pretest or posttest data were collected from approximately 70% of those for whom access had been achieved. Also, we located 76% of this targeted sample at 1-year followup (1994-1996). However, 3% of the youth or their parents refused to continue participation at 1-year followup. Another 6% of the students were not available for interview. Interviewers had either reached their family or them at least once, but the student was not available then or was not at home after repeated phone calls (an average of 7 attempts, SD=7.5, to an average of 1.6 telephone numbers, SD=1.0). A follow-up survey was provided to 67% of this targeted sample. This follow-up rate of 67% (n=1250) is only slightly lower than that obtained with a traditional school sample at one-year follow-up. The analysis n=1074; only program subjects who reported being exposed to at least one program session were included.

One-year follow-up data was analyzed. Baseline comparability of conditions through use of the random blocked random assignment procedure was quite good. Prediction of level of current use (use in the last 30 days) of cigarette smoking, alcohol use, marijuana use, or hard drug use (this last variable consisting of the mean of five illicit drugs) from condition, and schools nested within condition, was completed using the PROC MIXED procedure. Means were adjusted for baseline use level, interaction between condition and baseline use level, and method of collection at follow-up (in class, 27%; by telephone, the remainder of the sample). The models for cigarettes, alcohol, and marijuana failed to indicate a significant main effect for condition. The model for hard drug use did indicate a significant main effect (F=3.85, p<.04). Those who received the intervention showed nearly half the monthly drug use frequency at follow-up compared to those in the control group. The a priori contrast of classroom-only condition versus control was significant (t=1.78, p<.05), whereas the contrast between the two program conditions was not (t=-.08, p=.93).

Two drug use measures showed a significant interaction between pretest use level and condition: alcohol (F=3.71,p=.02) and hard drug use (F=9.74, p<.01). Separate regression lines by condition were plotted for those measures over the pretest range of 0 (i.e., non-use) to 30 (i.e., daily) times per month. For alcohol, there appeared to be no effect of the program among pretest non-users and lower levels of use, with the regression intercept values all roughly equal (average value=11.64, standard error of the alcohol intercepts = 2.33). At higher pretest alcohol use levels, the classroom-only condition subjects did exhibit lower alcohol use at one-year follow-up compared to the control condition subjects at comparable pretest use levels (at pretest use level=30, t=7.42,p<.01; approximately a 25% relative reduction in use). The contrast between the two program conditions among subjects with higher levels of pretest use was not significant in the expected direction (t=-2.30,ns 1-tailed). For hard drug use, the pretest non-users showed a benefit from program exposure, with control group intercept=11.60 (se=2.84), the classroom-only condition intercept at 2.78 (se=1.15), and the school-as-community plus classroom intercept=7.91 (se=2.10). At these low levels of baseline use, the contrast between classroom-only and control was significant (t=4.42,p<.01); the contrast between the two program conditions was not significant in the expected direction (t=-2.33, ns 1-tailed). The classroom-only condition subjects maintained this advantage in lower one-year follow-up use levels across a wide the range of use, whereas the school-as-community plus classroom condition only found an effect among intermediate and higher level pretest users (i.e., approximately a 65% relative reduction in use was achieved at higher levels of use in each program condition).

The intervention program was able to impact on high risk youth, although the school-as-community component apparently did not enhance this effect (even though many of this component’s activities were completed school-wide). The booster at the one-year follow-up was 20-minutes long. It followed the post-test assessment. In general, youth they were asked how they used this material in the last year and how they might use it in the next year. Also, a Continuation Community Newsletter continued to be implemented at the 7 combined-component schools, and was sent out to all those who had attended the TND combined program the previous year. Effects held up at a two year follow-up for females, and at a three year follow-up regarding hard drug use, for females only. The trend of the data at the three-year follow-up is for classroom-only program to work relatively well for those who had reported higher levels of use at baseline, whereas the school-as-community/combined program worked relatively well for those who had reported low levels of use at baseline (1997-present). Both programs found better results than control condition at all levels of baseline use.

Generalizability Trial. The generalizability trial was implemented at three comprehensive high schools (1995-1997). The comprehensive high school delivery involved a two-condition experimental design: TND-curriculum and standard care conditions, with a pretest-posttest, one-year follow-up evaluation identical to the continuation high school trial. A classroom-level of assignment was completed (i.e., 13 classrooms per school were assigned to each of the two conditions). The baseline consent rate was 90%. Follow-up rate at one-year was 62.5% (n=679). This sample was 42% male (a lower percent male than at the continuation schools), 36% White, 38% Latino, and 26% African American. A mixed- model analysis of variance was used again, as with the other study. The results were the same as for the continuation high school youth. Effects were found on higher levels of alcohol use (up to a 60% relative reduction at the highest levels of baseline use) and at all levels of hard drug use (here, a 60-75% relative reduction). In summary, a one-year follow-up period, the effects in the RHS replicated the findings found at the CHS.

Based on some of our pilot testing work, we knew that 12 lessons apparently is a maximum number that continuation high school personnel will permit at their schools. Since this curriculum needs to be generalizable to the "real-world" setting, as well as be maximally effective, only three new lessons were added (33% increase in length) to try to elicit effects on cigarette smoking and marijuana use in the renewal grant, TND-2 (1997-2001). One and-one-half lessons focused on marijuana use. A second lesson focused on cigarette smoking cessation, with provision of key concepts from Project TNT, and also including use of a self-help booklet. The other new material expanded on the self-control lesson in Project TND. In that earlier lesson, the dimension of self-control that was discussed primarily was modification of one's provocative behavior, or learning to behave in ways to match situations. The current lesson focused on anger coping, which is highly correlated with drug use and violence. The ways in which drug use and violence interact were discussed, and self-control of aggression was instructed.

In addition to lengthening the program, two modes of delivery were contrasted in the present study: an efficacy condition (health educator-led classroom program) and a treatment effectiveness condition (health educator-assisted self-instruction program, that program modality generally used at continuation high schools). The self-instruction curriculum consists of self-reviews with answer keys, all program material re-written, very similar to the health educator-delivered program but with a programmed learning style of presentation, and use of several icons to enhance the readability and palatability of the material.

The health educator-delivered classroom program was compared to a health educator-assisted self-instruction program versus a standard care (control) condition, in a three-condition randomized design (n=18 schools), with a two-year follow-up that included memory-enhanced booster programming. Regarding the comparison of the two program conditions, we simply asked whether or not there will be differences in preventive effects between these two conditions. One may guess, however, that the health educator-delivered condition may achieve a greater preventive effect than the health educator-assisted condition, because of much greater interaction among groups of students in the classroom (e.g., student group interaction has been conjectured to be an important characteristic of effective substance abuse prevention programming). Immediate outcomes data indicated small to moderate effects on behavioral intentions to use all four categories of drug use in the future, and moderate effects on knowledge, particularly on stereotyping, the new marijuana panel lesson, perspectives, the talk show, self-control, and the new tobacco cessation lesson. The one-year follow-up results showed that the self-instruction condition did no better than the control condition. However, the health-educator led condition showed effects on all four substances, analyzed at a conservative school-level of analysis. This condition was superior to the other two, the latter of which did not differ from each other. A 12-session interactive program is able to find effects across all classes of substances. TND has now shown 3 replications of program effects in large experimental trials. Analysis of the memory booster and longer-term follow-up is ongoing. Project TND-2 will terminate 7/2001.

Project TND has engaged in a variety of etiological, empirical review, curriculum development, and methodological studies. Etiological studies have included investigation of stereotyping of high risk youth, group self-identification of high risk youth, ethnic differences in sensation seeking and drug use, violence and drug use, depression and drug use, memory association and drug use, prediction of health risk factors, drug use myths and drug use, predictors of self-initiated cigarette smoking cessation, and correlates of drug abuse and dependence, as examples. Empirical reviews include an operationalization of the concept of "high risk", the current status of marijuana research, and the generalizability of prevention research to ethnic minorities. Numerous curriculum development studies were completed and published, some involving use of novel development strategies including use of the "theme study." Finally, methodological studies (other than the curriculum development work) include the associations of naturalistic observation, refuse analysis and school personnel reports with adolescents self-reported drug use, predictors of tobacco use misreports, implementation and process evaluation of the school-as-community and classroom components, the importance of parental consent requirements, and design issues in community trials.

The significance of our findings to the scientific field and their potential impact on health, so far, include:

(1) Prevalence of drug use among continuation high school students is much higher than students from regular high schools. This is a context in which prevention efforts are needed, and which serves as a natural laboratory in which high risk youth can be studied longitudinally.

(2) Generic social influences programming is not preferred by high risk high school students. Some social influences programming can be adapted to them (e.g., effective communication and decision making), but the addition of prosocial behavior coping skills and cognitive perception correction programming are needed .

(3) It is possible to consent, collect data, and track youth from this high risk sample.

(4) High risk students attend at least two-thirds of their drug abuse prevention lessons over a 3 week period with no decreased attendance over time, show learning of TND material relative to the standard care condition, and report high interest and belief in the material. A school-as-community component can be implemented successfully, but it does not appear to add much effect to that of a classroom-only program. Effects on hard drug use and alcohol use have been achieved at one-year follow-up, using this motivation-skills-decision making model. Some effects last up to 3-years post-programming, among females.

(5) Such programming can be adapted to youth in traditional school contexts (thus far, regular high school youth).

(6) It is possible to create a 12-session version of Project TND that will effect cigarettes, marijuana, alcohol, and hard drug use at a 1-year follow-up.