Proximal relationships between PTSD symptoms and drinking among female college students: results from a daily monitoring study

Abstract

Self-medication has been theorized to explain comorbidity between posttraumatic stress disorder (PTSD) and drinking, whereupon problem drinking develops in order to modulate negative affect and ameliorate PTSD symptoms. Daily monitoring methodologies may help refine our understanding of proximal relations between PTSD, affect, and alcohol use. One hundred thirty-six female college drinkers with a past history of sexual victimization and 38 female college drinkers with no past trauma history completed electronic monitoring of PTSD symptoms, affect, alcohol use, and alcohol cravings, daily for 4 weeks. A two-part mixed hurdle model was used to examine likelihood of drinking and amount of alcohol consumed on drinking days. We found significant relationships between daily PTSD symptoms, affect, and drinking. On days women experienced more intrusive and behavioral avoidance symptoms of PTSD, they experienced stronger urges to drink and were more likely to drink on that day. On days in which women experienced more negative affect than their average, they experienced stronger urges to drink, whereas on days in which women experienced more of the dysphoric symptoms associated with PTSD than their average, they drank less. On days with higher positive affect, women reported stronger urges to drink and were more likely to drink. Results suggest the need to examine both aspects of affect and specific PTSD symptoms as they may differentially predict drinking behavior. Differences in the ways in which PTSD symptoms and affect influence drinking suggest that interventions more specifically address the function of drinking behaviors in reducing alcohol use among college women.

Figure 1

Between (M) and within (D) person effects of daily PTSD symptoms and affect on drinking urges. Regression coefficients as well as 95% confidence intervals (CI) are presented in the dotplot. Predictors with CIs that include zero are nonsignificant.

Figure 2

The effects of daily PTSD symptoms and affect on drinking behavior. The count and logit models each contain both between (M) and within (D) person effects. Rate ratios (RR) or odds ratios (OR) as well as 95% confidence intervals (CI) are presented in the dotplot. Predictors with CIs that include zero are nonsignificant.