Reducing Health Risk Behaviors and Improving Depression in Adolescents

A Randomized Controlled Trial in Primary Care Clinics

Sunhye Bai, PHD, MPH; Luis R. Zeledon, MD; Elizabeth J. D'Amico, PHD; Steve Shoptaw, PHD; Claudia Avina, PHD; Anne P. LaBorde, PHD, PSYD; Martin Anderson, MD, MPH; Olivia M. Fitzpatrick, BA; Joan R. Asarnow, PHD


J Pediatr Psychol. 2018;43(9):1004-1016. 

In This Article

Abstract and Introduction


Objective: Primary care (PC) is a major service delivery setting that can provide preventive behavioral health care to youths. To explore the hypothesis that reducing health risk behaviors (HRBs) would lower depressive symptoms, and that health risk and depression can be efficiently targeted together in PC, this study (1) evaluates an intervention designed to reduce HRBs among adolescent PC patients with depressive symptoms and (2) examines prospective links between HRBs and depressive symptoms.

Method: A Randomized controlled trial was conducted comparing a behavioral health intervention with enhanced Usual PC (UC+). Participants were 187 adolescents (ages 13–18 years) with past-year depression, assessed at baseline, 6 months, and 12 months. Primary outcome was the Health Risk Behavior Index (HRBI), a composite score indexing smoking, substance use, unsafe sex, and obesity risk. Secondary/exploratory outcomes were an index of the first three most correlated behaviors (HRBI-S), each HRB, depressive symptoms, and satisfaction with mental health care.

Results: Outcomes were similar at 6 and 12 months, with no significant between-group differences. HRBI, HRBI-S, and depressive symptoms decreased, and satisfaction with mental health care increased across time in both groups. HRBI, HRBI-S, and smoking predicted later severe depression. Conversely, severe depression predicted later HRBI-S and substance use.

Conclusions: UC+ and the behavioral health intervention yielded similar benefits in reducing HRBs and depressive symptoms. Findings underscore the bidirectional links between depression and HRBs, supporting the importance of monitoring for HRBs and depression in PC to allow for effective intervention in both areas.


Less than 50% of adolescents with mental health or substance use problems (hereafter referred to as behavioral health problems) receive adequate treatment, in part because of poor detection of problems, and barriers to accessing empirically supported services (Merikangas et al., 2011). The pediatric primary care (PC) setting is a major point of health service contact for these youths, as most adolescents visit their physician annually or more (Nordin, Solberg, & Parker, 2010). Developing and evaluating care models that aim to integrate behavioral health services into PC may help address this service gap and improve adolescent health (Fallucco, Seago, Cuffe, Kraemer, & Wysocki, 2015; Kolko & Perrin, 2014; Rapp, Chavira, Sugar, & Asarnow, 2017; Stancin & Perrin, 2014; Tynan, 2016).

Integrated care offers the possibility of improving adolescent health while reducing health care costs associated with patient care (McGrady & Hommel, 2016). A recent meta-analysis of interventions that aimed to improve access to behavioral health services through PC (hereafter referred to as "integrated care") detected a small but statistically significant intervention effect, relative to Usual PC (Asarnow, Rozenman, Wiblin, & Zeltzer, 2015). Specifically, integrated care interventions improved mental health outcomes such as depression, anxiety/somatic complaints, and behavioral problems. However, effects were weaker and not statistically significant for interventions targeting substance use/abuse (Asarnow et al., 2015). Indeed, among the nine studies aimed at reducing substance use in the meta-analysis, only one study yielded statistically significant intervention benefits (Pbert et al., 2008). These results underscore the importance of further work to evaluate and develop PC strategies that optimally target behavioral health concerns.

Reducing the emergence and exacerbation of substance use and related health risk behaviors (HRBs) in adolescence is a top health priority (D'Souza-Li & Harris, 2016). HRBs, such as smoking, drug and alcohol use, unsafe sex, and unhealthy diet and exercise habits frequently occur with adolescent depression, adding to the personal and societal burden of this common and impairing disorder (Asarnow et al., 2014; Luppino et al., 2010; Wickrama & Wickrama, 2010). Given the cluster of HRBs that co-occur with and possibly exacerbate depression during adolescence (Fluharty, Taylor, Grabski, & Munafò, 2017; Luppino et al., 2010; O'Neil, Conner, & Kendall, 2011; Wickrama & Wickrama, 2010), targeting multiple HRBs with one intervention program may be an efficient strategy for reducing health risk and depression in youths. This approach fits well within PC, as the PC clinician is well positioned to screen, monitor, and intervene on HRBs and depression, and promote physical and behavioral health (Stancin & Perrin, 2014). Given the time constraints of PC visits, multiple HRB interventions that target several co-occurring problems may be a useful alternative to a complex suite of many evidence-based behavioral health treatments, each targeting a different but related problem.

Multiple HRB interventions have proven successful in diverse contexts and samples including school-based universal prevention trials (Hale, Fitzgerald-Yau, & Viner, 2014), HIV prevention trials targeting high-risk youths (Rotheram-Borus et al., 2003; Rotheram-Borus et al., 2001), and PC trials targeting adult patients at risk for cardiovascular disease or cancer (Goldstein, Whitlock, & DePue, 2004; Prochaska & Prochaska, 2011). The present study applied this approach to target HRBs and depression in the PC setting, selecting adolescents with past-year depression histories. Given the frequent co-occurrence of HRBs and depression, we focused on a depressed sample to allow the evaluation of a critical hypothesis underlying the intervention; that reductions in HRBs would lead to fewer depressive symptoms.

The study advances the field by evaluating whether a PC intervention strategy that simultaneously addresses multiple HRBs can lead to reduced HRBs and depressive symptoms. The intervention was designed to target risk and protective factors for the emergence and exacerbation of HRBs. We aimed to decrease and prevent HRBs in adolescents selected for depression, and to determine whether this approach would have benefits on HRBs and depression levels. Our ultimate goal was to identify PC strategies for decreasing HRBs and depression, improving health and behavioral health, and minimizing the personal and economic costs of behavioral health problems. We hypothesized that the intervention—in comparison with enhanced usual care—would decrease participation in HRBs and improve depression at 6 months. Intervention effects were explored over the 6- to 12-month follow-up. This intervention was guided by our hypothesis that less engagement in HRBs would be associated with fewer depressive symptoms. We tested this underlying hypothesis by examining the prospective links between HRBs and depression. We predicted that youths who engaged in more HRBs would report more depressive symptoms, and consistent with our targeted patient-selection strategy, that youths with more depressive symptoms would engage in more HRBs.