Cross-Lagged Analyses Across Middle-Adulthood of the Bidirectional Associations Between Substance Use Disorders and...

We sought to clarify the causal nature of the bidirectional associations in adulthood between substance use disorder (SUD) and psychosocial dysfunction (PSD).

As assessed from National Swedish registers, SUD is measured by alcohol use disorder (AUD) and drug use disorder (DUD) and PSD by unemployment (UN), low income (LI), and high community deprivation (HCD). A cross-lagged structural equation model from ages 31 to 48 is applied to the native Swedish population born 1960-1980 and residing in Sweden at age 29 followed through 2017 (n = 2,283,330), excluding individuals with prior SUD and PSD.

All models fitted well. Focusing on the cross-lagged paths, across sexes, substances, and forms of PSD, parameter estimates for the SUD to PSD paths consistently exceeded those for the parallel PSD to SUD paths. SUD to PSD paths were nearly all statistically significant. Although the UN to SUD and LI to SUD paths were usually significant, most HCD to SUD paths were not. The differences between the UN to SUD and SUD to UN paths grew larger with increasing age, whereas the opposite pattern was seen for the HCD to SUD and SUD to HCD paths.

Across sexes, forms of SUD, and dimensions of PSD, in a fully parametrized and good-fitting cross-lagged model of middle adult life, a diagnosis of SUD consistently predicted future PSD whereas PSD often but not always predicted future SUD. The SUD to PSD paths were consistently larger than the parallel PSD to SUD paths. Our results suggest a bidirectional causal relationship between SUD and PSD across adulthood, driven largely, but not entirely, by the negative effects of SUD on future psychosocial functioning.