Abstract
Background: Posttraumatic stress disorder (PTSD) and problematic alcohol use commonly co-occur among military service members. It remains critical to understand why these patterns emerge, and under what conditions. Objectives: This study examined whether PTSD symptoms (PTSS) and alcohol involvement (quantity and frequency of use, heavy episodic drinking, and alcohol problems) are indirectly related through four distinct drinking motivations. A secondary aim was to identify factors, specifically forms of social support, which buffer these associations. Methods: Using baseline data from a randomized-controlled trial of health and well-being among civilian-employed separated service members and reservists, the present study examined these issues using a subsample of 398 current drinkers. Results: Parallel mediation models revealed PTSS–alcohol consumption associations were indirect through coping and enhancement motivations. PTSS was only related to alcohol problems through coping motivations. In addition, the indirect effect of PTSS on average level of consumption via coping motives was conditional on perceived support from friends and family, whereas the indirect effect for alcohol problems was conditional only on friend support. In contrast, the indirect effects of PTSS on alcohol consumption variables (but not problems) via enhancement motives were conditional on perceived support from friends and family. Conclusions/Importance: Future research and screening efforts should attend to individual motivations for drinking as important factors related to alcohol use and problems among service members experiencing PTSS, and emphasize the importance of communication, trust, and effective supports among military and nonmilitary friends and family.
Acknowledgment
The contents are solely the responsibility of the authors and do not necessarily represent the official views of NIOSH, CDC or HHS
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.
Notes
1 Results did not differ between models which included values of 0 for average number of drinks consumed and those with recoded 0.5 values.
2 A pathways from PTSS to drinking motives were identical for models predicting number of drinking days, average number of drinks consumed, and likelihood of engaging in heavy episodic drinking. Models predicting alcohol problems controlled for both number of drinking days and heavy episodic drinking.
3 Alcohol problem models were reexamined using the full AUDIT. The internal consistency of the full AUDIT was slightly higher than the adapted version (α = .82 and .79 respectively). Overall, results did not differ from those presented in the manuscript. For models examining the conditional influence of perceived support from friends, coping motives were associated with higher AUDIT scores (b = 1.50, p < .001), and there was a significant conditional indirect effect for PTSS on AUDIT scores through coping motives where the indirect effect was significantly different from zero at low (ab = .34) and average levels of friend support (ab = .28), but not at high levels of friend support (ab = .20). For models examining the conditional influence of perceived support from family, coping motives were again associated with higher AUDIT scores (b = 1.43, p < .001), but the indirect effect of PTSS on AUDIT scores through coping motives was not conditional upon one’s perceived level of support from family.
4 Effects of study covariates (age, race, gender, military status, neuroticism, and combat exposure) are not reported. In all conditional indirect effect models, only age was associated with number of drinking days and likelihood of heavy episodic drinking, whereas gender was associated with average number of drinks consumed. Younger participants were more likely to engage in heavy episodic drinking during the past month (b = −.04, z = −2.79, p < .01), whereas older participants reported more drinking days, on average (b = .16, t(379) = 3.57, p < .001). Men consumed significantly higher levels of alcohol, on average (b = −.22, t(379) = −2.03, p < .05), although when applying the Benjamini–Hochberg procedure, the gender difference was no longer significant.