Clinical preventive services efficacy and adolescents' risky behaviors
S. M. Downs and J. D. Klein
Department of Pediatrics, University of North Carolina School of Medicine, Chapel Hill, USA.
OBJECTIVE/BACKGROUND: To analyze the value of studying or implementing
office-based clinical preventive services for adolescents. Most adolescent
mortality and morbidity is attributable to risky behaviors, yet clinical
preventive services to reduce risky behaviors are often challenged because
their efficacy has not been demonstrated. DESIGN: A cost-effectiveness
model of adolescents' risky behaviors that compares standard practice with
a program of screening visits for all adolescents and counseling visits for
youth identified as high risk. We considered two risky behaviors, alcohol
abuse and unsafe sexual activity, and five outcomes. MAIN OUTCOME MEASURES:
Baseline cost-effectiveness of the program, minimum efficacy at which the
program would be cost-effective, and sample sizes required for a trial of
the program. RESULTS: Assuming that the program is 5% effective at
preventing risky behaviors, it would cost $3035 to prevent any one adverse
outcome and $471,000 to prevent a death from an automobile crash or from
human immunodeficiency virus infection. Assuming society were willing to
pay $600,000 to prevent a death (a generally accepted figure), the program
would be cost-effective only if it were 5.6% effective at changing
behavior. At this efficacy, the program would have a cost per year of life
saved comparable to or better than many other accepted medical
interventions. However, to demonstrate changes in outcomes at this efficacy
would require a clinical trial with between 4000 and 95 million adolescents
in each treatment group, depending on the outcome measured. CONCLUSIONS:
Studying the ability of clinical preventive services to prevent outcomes of
adolescents' risky behaviors would be impractical. The decision to
implement these programs should be made based on current knowledge and
beliefs; their efficacy can probably be studied only as part of widespread
implementation.
