The original MTP sample (N = 1016) was compared with the subset of participants who were included in the current investigation (n = 526) by using t tests and χ2 tests for age, education, gender, marital status, route of MA administration, employment, and baseline ASI composite scores. In all analyses, there were no significant differences between the patients in the current study and the original MTP sample.
At 3-year follow-up, the majority of the sample was white (69%; n = 362), female (60%; n = 316), employed (60%; n = 316) and had a high school education (33% had college education or higher); average age was 33.4 (SD = 8.0). At baseline, participants reported using MA an average of 11.9 days out of the past 30 (SD = 9.6). The preferred route of administration was smoking (63%; n = 331), followed by injection drug use (27%; n = 143) and intranasal use (9%; n = 49). There were no differences in demographic or substance use characteristics among those who completed the psychiatric assessment (N = 526) relative to those who did not (n = 61).
Pretreatment Depressive Symptoms as a Predictor of Outcomes During and After Treatment
We first examined the relationship of baseline BDI scores to treatment adherence and MA use status in the 30 days prior to discharge and follow-up. A multivariate regression model controlling for demographics, pretreatment frequency of MA and alcohol use and route of MA administration revealed that depression severity and treatment adherence were inversely related (β = −0.18, SE = 0.07; p = 0.01). Depression severity was significantly associated with self-reported MA use status in the 30 days before discharge (t = 2.80, p < 0.01); those who used MA in the month preceding discharge had higher BDI scores (M = 13.7. SD = 9.5) relative to those who abstained (M = 7.7, SD = 8.1). However, logistic regression analyses revealed that baseline BDI scores did not predict self-reported MA abstinence status in the 30 days prior to follow-up (z = 1.71, p = 0.09). All findings with regards to the relationship between baseline BDI and MA use outcomes were replicated using urine test data as outcomes.
We next examined the relationship of baseline BDI scores to AD diagnoses at follow-up. Baseline BDI scores were significantly higher among those with AD (n = 80; 15.2%) relative to those without AD (t = −2.8, df = 524, p < 0.01).
Changes in Depression During Treatment
Overall, BDI scores changed significantly during treatment (t = −13.9, df = 524, p < 0.0001) with baseline scores (M = 16.5, SD = 10.2) higher than end-of-treatment scores (M = 10.2, SD = 9.2). To investigate the clinical course of these symptoms in relation to that of MA dependence, we next examined whether the magnitude of change in depressive symptoms varied as a function of MA use status in the 30 days before discharge. A multivariate regression model controlling for demographics and frequency and route of MA administration revealed that the reduction in depressive symptoms among those who reported abstinence from MA in the month before discharge was significantly greater (β = 5.1, SE = 0.69) than that observed in those who used MA (p < 0.0001) (). Results were replicated when comparing those who tested positive for MA on urine test at discharge to those whose tests indicated abstinence.
Mean total BDI score as a function of time, separated by self-reported MA use status during the 30 days before treatment discharge (Abstinent = no MA use; Used MA = 1 or more days of MA use during the 30 days before discharge).
Depression Diagnoses and Severity at Follow-Up
Overall, 15.2% of the sample at 3-year follow-up met current MDD criteria. A significantly greater proportion of those who reported using MA during the month preceding follow-up were diagnosed with MDD (25.9%; n = 41) relative to those who were abstinent (10.6%; n = 39), χ2 = 20.20, df = 1, p < 0.0001; OR = 2.95; CI = 1.8 to 4.8. Moreover, those with MDD used MA more frequently during the follow-up period (β = 6.0, SE = 1.69; p < 0.0001) than those without MDD. Lifetime MDD and dysthymic disorder were not significantly associated with MA use in any analyses; thus, the remaining analyses focused on current MDD.
Follow-up BDI scores were significantly related to route of MA administration; injectors reported more depressive symptoms than those who used any other route of administration, t = −2.4, df = 524, p < 0.05. Moreover, the odds of being an injection user were significantly greater among those with current MDD (OR = 1.9, 95% CI = 1.2–3.2) relative to those without this diagnosis.
We examined the association between MDD at follow-up and several pretreatment substance use variables using t tests. MDD diagnosis was not significantly related to pretreatment MA use frequency, age of first MA use, number of years of lifetime MA use, or the ASI drug composite.
The relation of MDD at follow-up to demographic variables was evaluated using t tests and chi square analyses. MDD was not significantly associated with age, ethnicity, marital status, or employment, but a marginally significant relationship emerged between MDD diagnosis and gender (χ2 = 3.87, df = 1, p = 0.05) indicating that a greater proportion of women in the overall sample (n = 56; 17.7%) than men (n = 24; 11.4%) met MDD criteria.
Participants with AD were compared with those without AD for current MDD and post-treatment depression severity. Those with AD were significantly more likely to be diagnosed with current MDD (χ2 = 5.34, df = 1, p = 0.02). Likewise, among those with MDD, individuals with a concurrent AD diagnosis evidenced higher BDI scores relative to those without this diagnosis both at discharge (M = 12.7 vs. M = 9.8; t = 2.33, df = 524, p = 0.02) and follow-up (M = 13.3 vs. M = 8.4; t = −4.34, df = 524, p < 0.0001).
Association of Depression With Other Psychosocial, Psychiatric, and Substance use Variables
ASI composite scores and BSI scale scores at baseline, discharge and follow-up for those with and without MDD are plotted in and , respectively, and the results of mixed-model repeated measures analyses testing the effects of time, depression diagnosis, and their interaction are provided in . Controlling for demographics, pretreatment MA use frequency, and route of administration, analyses revealed a significant time × depression diagnosis interaction on 4 of the 7 ASI composites (alcohol, drug, employment, and psychiatric) and 7 of the 10 BSI scales (somatization, obsessive-compulsiveness, interpersonal sensitivity, depression, phobic anxiety, psychoticism, and global severity). All interactions indicated that those with MDD reported problems of significantly greater and increasing severity over time in all areas.
Mean ASI composite scores as a function of time among MA dependent adults with (N = 80; upper panel) and without (N = 446; lower panel) MDD at 3-year follow-up.
Mean BSI scale scores as a function of time among MA dependent adults with (N = 80; upper panel) and without (N = 446; lower panel) MDD at 3-year follow-up.
Changes in Psychosocial, Psychiatric, and Substance-Related Impairment Among MA Dependent Adults at 3-year Follow-Up: Linear Mixed-Effects Models