Tobacco, Alcohol, and Drug Use of People with Mood and Anxiety Disorders: Correlates of Family and Friend Relationships

Background: People with mood and anxiety disorders are a high risk population for co-occurring substance use disorders. The literature suggests that positive social networks may reduce the risk of substance use in general. However, this relationship has not yet been empirically tested for individuals with mood and anxiety disorders. Social networks for this population tend to be limited compared to people without these disorders, as one of the symptoms includes social withdrawal, making it hard to maintain good relationships. Also, this population may use substances to alleviate symptoms and to maintain social contact. Therefore, identifying social network functions in relation to substance use behaviors, particularly in this population, is of great importance in social work practice. The current study examined the impact of family and friend relationships on substance use for those with mood and anxiety disorders.

Methods: The National Survey of American Life (NSAL) was used for the secondary data analysis. Respondents who had a diagnosis of mood and anxiety disorders were included in this study (n=1,076). The sample was predominantly female (74.0%), non-White (84.4%), 74.3% had more than a high school diploma, and had a mean age of 39.54 (SD=14.02). Substance use was identified as tobacco (0=not smoking, 1=currently smoking), alcohol (frequency of drinking), and drug use (0=no, 1=used marijuana, cocaine, or etc.) in the past 12 months. Family and friend relationships (e.g., closeness, frequency of contact, receipt and provision of social support) were assessed using either a 4 or 7 Likert scale item (Cronbach’s α > .9). Structural equation modeling (SEM) with the means and variance-adjusted weighted least squares was conducted to identify the impact of family and friend relationships on each substance use behavior, controlling for age, gender, education, work status, poverty level, and the number of family members who had alcohol or drug use problems. 

Results: The sample consisted of 26% smokers, 46% alcohol users, and 16% drug users. The SEM model fit the data well (χ²(152)=490.154, p<.001, RMSEA=.045 _{[90% CL .041 - .050]}, CFI=.974, TLI=.966). The model revealed that tobacco and alcohol use were not associated with the family or friend relationships after controlling for other variables; however both family and friend relationships were strong predictors of drug use. Specifically, a better family relationship was associated with lower drug use (β=-.101, p=.031), while a better friend relationship was related to higher drug use (β=.143, p=.004). 

Conclusions and implications: Consistent with the literature, positive family relationships seem to be important in reducing the risk of drug use for individuals with mood and anxiety disorders. However, our study revealed the negative effect of friend relationships that may increase the risk of drug use. The cross-sectional data does not articulate the causality, yet literature has emphasized the importance of friends in substance use treatment. In particular, having friends without any substance abuse is identified as a key. The current data does not differentiate the friends’ characteristics which require future investigation. Implications on how social workers can facilitate social support to reduce substance use behavior will be discussed.