Borderline Personality Disorder Symptoms and Substance Dependence Before and After Jail Incarceration



Journal Title

Journal ISSN

Volume Title



Borderline personality disorder (BPD) and substance use disorders (SUDs) commonly co-occur across community, clinical, and correctional settings. Research has begun to examine the relationships between BPD and SUDs. However, it is unclear whether BPD is differentially related to different SUDs – in other words, is BPD more strongly related to some SUDs compared to other SUDs? Second, most of the research examining BPD and SUDs has been cross-sectional; thus, little is known about the directionality of predictive relationships between BPD and SUDs. Across two studies, this dissertation further examines the relationships between BPD and SUDs with a large sample of people currently (Study 1) and formerly (Study 2) incarcerated in an adult detention center (i.e., jail) located in a suburb of Washington, DC. Using structural equational modeling, Study 1 of this dissertation examines whether BPD symptoms assessed shortly after jail incarceration are differentially related to alcohol, cannabis, cocaine, and opioid dependence symptoms experienced in the 12 months prior to incarceration. BPD symptoms were significantly related to dependence on each of the four substances, but the link between BPD symptoms and cocaine dependence was stronger than alcohol, cannabis, or opioid dependence. Study 2 of this dissertation utilizes cross-lagged structural equational modeling to determine the directionality of the relationship between symptoms of BPD and dependence on alcohol, cannabis, cocaine, and opioids across three post-release timepoints (one year, four years, and seven years after release from jail). In examining the relationships of symptoms of BPD to dependence on cannabis and cocaine, the unidirectional model of BPD symptoms predicting subsequent substance dependence fit the data best. Results were less clear regarding the relationship of BPD symptoms to alcohol dependence. Both unidirectional models – BPD symptoms predicting subsequent alcohol dependence, and alcohol dependence predicting subsequent BPD symptoms – separately fit the data; the bidirectional model, where both directions simultaneously occur, did not significantly improve model fit compared to either unidirectional model. Regarding the relationship of BPD symptoms to opioid dependence, the baseline model fit the data best, as neither unidirectional model significantly improved model fit. Findings from both studies generalized across men and women and across White and Black people currently and formerly incarcerated in jail. Implications of findings and directions for future research are discussed.