Rural Adolescent Drinking Behavior: Three Year Follow-Up in the New Hampshire Substance Abuse Prevention Study

Article excerpt

The New Hampshire Substance Abuse Prevention Study is a cohort study of 4,406 public school children, who were in elementary school, junior high school or in the tenth grade in 1987. The subjects consisted of 2,227 females (50.5%) and 2,179 males (49.5%); 98.6% are white. This is a rural population living in communities of 668 to 11,795 according to the 1990 census. The children are from a broad range of economic groups with 1990 median incomes for a family of four ranging from $12,798 to $32,429. This cohort is representative of rural children in New Hampshire (U.S. Bureau of the Census, 1991).

Two approaches to substance prevention in a quasi-experimental design were tested with a follow-up at 36 months. The arms were: (1) a comprehensive school curriculum (Here's Looking At You, 2000) implemented in grades 1 through 12; (2) this curriculum (Parent Communication Course) and a community task force (Johnson Institute Model); and (3) delayed intervention control. The control community agreed not to implement any comprehensive school curriculum, parenting course, or community task force during the follow-up period. In the intervention communities, residents implemented and coordinated the interventions with training and support provided by research staff. In the combination intervention there was no explicit central community coordination of the three components, though there were overlapping persons in key positions in all components.

Baseline questionnaires were collected from 68% of the total public school population in March and April of 1987. The interventions began in autumn 1987 and continued through the 36 months of follow-up data collection. Annual follow-up questionnaires were collected in March and April of 1988, 1989, and 1990. Questionnaires were collected from all absentees and students assigned to home study. School dropouts, graduates, and transfers were not followed. Four years of data from 79% of the cohort were collected. In 1992 and 1993 the senior author conducted group and individual interviews.

The following definitions of drinking behavior were used. Initiation - subject has ever tried beer, wine or liquor, excluding religious observances, but including secular family occasions. Drinking - subject has had beer, wine or liquor excluding religious observances, but including secular family occasions twice or more in the 30 calendar days preceding questionnaire completion. Drunkenness - subject has had five or more drinks in a row or passed out or vomited more than once after drinking.

The two main analyses were by school and by individual, controlling for community, age, gender, and psychosocial characteristics. Some positive prevention results were achieved for cigarettes, marijuana, other illicit drugs, and spitting tobacco, but no effects were found for any of the three levels of alcohol use at 36 months' follow-up (Table 1), although one-year results had supported a positive effect. Subgroup analyses using the school as the unit of analysis in order to account for different baseline rates of drug use and different rates of attrition (defined as the percentage who failed to complete four questionnaires) supported the main analyses' results in every case.

Some research groups have found positive one-year results for drinking (Botvin, Baker, Filazzola, & Botvin, 1990; Pentz et al., 1989; Dielman, Shipe, Leech, & Butchart 1989; Hansen & Graham, 1991), but positive results for this drug seem more difficult to achieve than for tobacco and marijuana (Pentz et al., 1989; Perry, Murray, & Griffin (1990); Perry et al., 1990; Botvin et al., 1990; Mackinnon, Pentz, & Stacy, 1993) and we do not yet know how durable these results will be. Achieving positive results for drinking prevention is especially important, [TABULAR DATA FOR TABLE 1 OMITTED] since most children drink before they use any other drugs (Committee on Adolescence, 1987; Dunne & Schipperheijn, 1989; Sharp & Lowe, 1989; Alexander et al. …