Design issues addressed in published evaluations of adolescent HIV-risk reduction interventions: a review

This study was conducted to determine: (1) issues in intervention design that have been addressed in behavioral interventions targeting human immunodeficiency virus (HIV)-risk behaviors among adolescents; (2) specific choices made in intervention design; (3) historic changes in the likelihood that issues in intervention design will be addressed; and (4) if an association exists between quality of evaluation design and the number of intervention design issues addressed.

Literature search employing five electronic databases and 11 journals for articles published from January 1983 through December 1993 reporting evaluations of adolescent HIV-risk reduction interventions.

The frequency with which 12 issues in intervention design were addressed: basing the intervention on a theory of behavioral change; specifying a target population; involving the targeted community in the formulation of the intervention; addressing developmental issues; providing facts; strengthening interpersonal skills; describing the media (format) for delivering the intervention; specifying potentially relevant characteristics of the interventionists; describing the duration of the intervention; providing boosters; pilot testing the intervention; and including other potentially augmentative elements.

Twenty-eight published intervention articles were included in these analyses. The median number of intervention design issues addressed in any study was six (range three to nine), although this number increased significantly over time (p < .01). There was substantial variability in the frequency with which each individual design issue was addressed, with some design issues (e.g., inclusion of specific facts and the description of the channel employed) being addressed in all studies. Other design issues were addressed in less than one-quarter of studies [e.g., basing the intervention on a theory of behavioral change (18%) and addressing developmental issues (21%)]. The targeted community was involved in one-third of studies. More recent studies and studies employing a randomized evaluation design with both preintervention and postintervention assessments addressed more intervention design issues than did earlier studies and studies employing other evaluation designs (p = .01 and p = .03, respectively).

The majority of published adolescent HIV-risk reduction studies have not addressed important issues in intervention design. However, more recent studies and studies employing a strong evaluation design have addressed a greater number of these issues. Frameworks to guide intervention efforts (e.g., to serve as "practice guidelines") are needed to allow for both accurate replication and meaningful comparison of differing intervention approaches.