Predictive Associations Among the Repeated Measurements of Anxiety, Depression, and Craving in a Dual Diagnosis Program

Abstract

Objective: Anxiety and depression commonly co-occur with substance use disorders. Conceptual models would presume that changes in anxiety and depression should lead to changes in drug and alcohol craving during treatment for co-occurring disorders, but no longitudinal investigation has explored this assumption. This study examined the associations among craving, anxiety, and depression in a dual diagnosis program. Methods: In all, 93 adult patients at a community-based dual diagnosis program for substance use disorders and comorbid anxiety and mood disorders provided daily ratings of anxiety and depression. We examined whether these ratings predicted changes in craving and vice versa. Results: Using hierarchical linear modeling to examine day-to-day change over time, we observed that anxiety and depression ratings uniquely and independently predicted subsequent craving ratings, and craving ratings also predicted subsequent anxiety and depression ratings. Conclusions: These preliminary findings support mutual maintenance models of substance use and psychiatric comorbidity, thus providing preliminary support for integrated programs that simultaneously address both problems. Implications and future directions are discussed.

Keywords:

• Dual diagnosis
• comorbidity
• community treatment
• treatment mechanisms

Note

Acknowledgments

We thank the clinicians and administrative staff at the UCLA Dual Diagnosis program for assisting with data collection and management.

Disclosures

The authors have no disclosures or conflicts of interest to report. The study was not supported by any grant funding.

Notes

1 We included these three participants because anxiety and depressive symptoms are often present in these other disorders (Schatz & Rostain, 2006; Petry, Stinson & Grant, 2005). Therefore, we took a transdiagnostic approach, focused on the associations among dimensional symptoms targeted in the dual diagnosis program that were likely relevant across all participants, rather than diagnostic categories. Still, we conducted a parallel set of analyses excluding the three participants and the results were identical. Thus, to retain statistical power we include them in the reported analyses.