Classification models for subthreshold generalized anxiety disorder in a college population: Implications for prevention

• We define four classification models of subthreshold GAD (the at-risk population).
• We discuss the potential of online self-help interventions to prevent/treat GAD.
• We estimate the cost of implementing a stepped-care model of prevention/treatment.
• Cost depends on intervention cost and efficacy and efficient selection of at-risk.
• Longitudinal prospective studies are needed to evaluate various at-risk models.

Generalized anxiety disorder (GAD) is one of the most common psychiatric disorders on college campuses and often goes unidentified and untreated. We propose a combined prevention and treatment model composed of evidence-based self-help (SH) and guided self-help (GSH) interventions to address this issue. To inform the development of this stepped-care model of intervention delivery, we evaluated results from a population-based anxiety screening of college students. A primary model was developed to illustrate how increasing levels of symptomatology could be linked to prevention/treatment interventions. We used screening data to propose four models of classification for populations at risk for GAD. We then explored the cost considerations of implementing this prevention/treatment stepped-care model. Among 2489 college students (mean age 19.1 years; 67% female), 8.0% (198/2489) met DSM-5 clinical criteria for GAD, in line with expected clinical rates for this population. At-risk Model 1 (subthreshold, but considerable symptoms of anxiety) identified 13.7% of students as potentially at risk for developing GAD. Model 2 (subthreshold, but high GAD symptom severity) identified 13.7%. Model 3 (subthreshold, but symptoms were distressing) identified 12.3%. Model 4 (subthreshold, but considerable worry) identified 17.4%. There was little overlap among these models, with a combined at-risk population of 39.4%. The efficiency of these models in identifying those truly at risk and the cost and efficacy of preventive interventions will determine if prevention is viable. Using Model 1 data and conservative cost estimates, we found that a preventive intervention effect size of even 0.2 could make a prevention/treatment model more cost-effective than existing models of “wait-and-treat.”

Figure optionsDownload as PowerPoint slide