Clinical predictors of mortality in dual diagnosis patients receiving integrated in-patient treatment—A cohort study with 21 years follow-up

Aims

To examine mortality, and identify predictors of mortality, in a cohort of dual diagnosis (co-existing psychiatric disorder and substance use disorder) patients.

Design

Cohort study based on national register and electronic health record data.

Setting and participants

Between 2002 and 2017, 2359 dual-diagnosis patients received treatment at a specialized Danish department and were followed until death, migration or a maximum of 21 years.

Measurements

Data on diagnosis, substance use, demographics, medication use and mortality were linked to national registers. Kaplan–Meier plots illustrated mortality differences across diagnostic groups, while the Cox model identified factors associated with mortality.

Findings

In this study, 23.5% of dual diagnosis patients died within ten years of discharge and 33.7% at end of follow-up. The highest mortality risk was among those with a primary substance use disorder (SUD) and no psychiatric diagnosis, with 45.9% dying. Alcohol [adjusted hazard ratio (aHR) = 1.42; 95% confidence interval (CI) = 1.18–1.71], opioids (aHR = 1.26; 95% CI = 1.03–1.54), as well as Charlson Comorbidity Index: 1 (aHR = 1.70; 95% CI = 1.41–2.04) and 2 + (aHR = 2.56; 95% CI = 1.94–3.36), increased mortality risk. In contrast, cannabis use disorder (aHR = 0.77; 95% CI = 0.65–0.93) and being female (aHR = 0.83; 95% CI = 0.71–0.97) reduced it. Psychiatric diagnoses F30–39 (aHR = 0.70; 95% CI = 0.56–0.87) and F40–49 (aHR = 0.76; 95% CI = 0.59–0.97) were associated with lower mortality risk compared with F20–29. Of 733 deaths with mortality cause data, nearly two-thirds were from natural causes, 8% from suicide and 20% from accidents or violence.

Conclusion

The observed high mortality rates among clinically confirmed dual diagnosis patients (co-existing psychiatric disorder and substance use disorder) in Denmark highlight the urgent need for comprehensive treatment for this population. The risk of death appears to be related to type of substance use, with alcohol and opioids associated with higher mortality rates than other substances.