Articles zyxw The Safer Choices Project: Methodological Issues in School-Based Health Promotion Intervention Research zy Karen Basen-Engquist, Guy S. Parcel, Ronald Harrist, Douglas Kirby, Karin Coyle, Stephen Banspach, Deborah Rugg ABSTRACT: zyxwvutsrqponm Randomized trials zyxwvutsr of school-based health promotion programs present unique design and analytical issues zyxw noi widely discussed in the research literature. This article describes the Safer Choices study - a school-based program for prevention of HlV, other sexually transmitted diseases, and pregnancy - to illustrate critical methodological issues involved in large-scale, school- based intervention trials, particularly those evaluating interventions with a school-wide focus. The issues presented are: I) compara- bility ofthe intervention and control groups even when few units are randomized: 2) factors that affect the decision to use a cohort or cross-sectional design: and 3 ) appropriate analysis strategy when the unit of randomization zyxwvu and intervention is at the school level, but observations are at the student level. (3 Sch Health. zyxwvuts 1997;67(9):365-371) zyxwvuts chools provide a point of access to most adolescents: S Nearly 95% of all children and adolescents attend elementary or secondary schools.' Because its primary mission is to impart knowledge and teach skills, school is the ideal place for promoting health among adolescents. Some school-based programs - such as those for children with asthma or for young people who use drugs - target specific student groups based on special health needs or risks. Other programs - such as those on smoking preven- tion, AIDS prevention, or nutrition education curricula - are designed to change risky behavior by targeting all students in the classroom. By establishing school-wide norms for drug use or unprotected sex, for example, these programs can also address students in the larger school environment. For those involved in school-based health promotion research, differences between individual-level and school-wide programs raise several important method- ological issues, such as identifying the proper unit of randomization and analysis, randomizing a small number of units, and identifying the most appropriate research design and data collection strategy. Karen Basen-Engquist. PhD, MPH, Assistant Professor, Dept. of Behavioral Science, University of Texas M.D. Anderson Cancer Center, Box 243, 1515 Holcombe. Houston, TX 77030: Guy S. Parcel, PhD, Professor and Director, Center for Health Promotion Research and Development, Universiiy of Texas at Houston Health Science Center: Ronald Harrist, PhD, Associate Professor of Biometry, University of Texas at Houston Health Science Center. P.O. Box 20186, Houston, TX 77225; Douglas Kirby, PhD, Senior Research Scientist: Karin Coyle. PhD. Director of Research, ETR Associates, P.O. Box 1830. Santa Cruz, CA 95061 -1830; Stephen Banspach. PhD, Chief, Surveillance and Evaluation Research Section, Division of Adolescent and School Health, Centers for Disease Control and Prevention, MS K33, 4770 Buford Highway, NE, Atlanta, GA 30341-3724; and Deborah Rugs, PhD, PMI Research Evaluation Coordinator, Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention, 1600 Clifon Road, NE. MS E59, .Atlama, GA 30333. This ariicle was submitted March 3, 1997, and accepted for publication May 28, 1997. This article addresses several key methodological issues concerning research on school health promotion programs with a school-wide focus. Using the Safer Choices project as an example, it discusses how to randomize a small number of units, like schools, rather than individuals; reasons for using a cohort or cross-sectional design; and how to analyze data when randomization and intervention delivery are at the school level, but observations are at the student level. THE SAFER CHOICES PROJECT Safer Choices, a five-year project funded by the Centers for Disease Control and Prevention, tested the effectiveness of a state-of-the-art program to prevent human immunode- ficiency virus (HIV) infection, the spread of other sexually transmitted diseases (STDs), and the incidence of unin- tended pregnancy among high school populations. HIV infection and other STDs are serious problems for adoles- cents. Approximately 20% of the people diagnosed with Acquired Immunodeficiency Syndrome (AIDS) are 20 to 29 years old.* Because of the long latency period between infection with HIV and diagnosis of AIDS, many adults in this age group are likely to have been infected with HIV during adolescence. Sexually transmitted diseases also are a significant problem for this group. Approximately 25% of sexually experienced adolescents contract an STD each year.' The Safer Choices project aimed to develop, implement, and evaluate a theory-driven, school-based, multiple- component HIV, STD, and pregnancy prevention interven- tion for high school students. To evaluate the program, the project team used a trial that randomly assigned schools to either a multiple-component intervention or a control condition that provided an information-based classroom curriculum on HIV and other STDs. Social cognitive theory," social influences,s,6 and models Journal of School Health November 1997, Vol. 67, No. 9 365