# Three years of medication-use sequences in incident bipolar disorder in Sweden reveal divergent patterns in native-born and immigrant populations

## TL;DR

This study finds that immigrants in Sweden with bipolar disorder are less likely to receive adequate medication compared to native Swedes, highlighting disparities in mental health treatment.

## Contribution

The study introduces state sequence analysis to compare medication-use patterns in bipolar disorder across native and immigrant populations in Sweden.

## Key findings

- Immigrant groups, especially refugees, show higher rates of treatment failure compared to Swedish-born individuals.
- Transitions to adequate treatment are rare across all population groups.
- Adjusting for covariates reveals non-refugee and refugee immigrants have the highest probabilities of treatment failure.

## Abstract

Guideline-conform treatment of mental disorders is compromised in immigrant populations, but longitudinal pharmacoepidemiologic patterns in bipolar disorder (BD) remain unknown. We aimed to close this knowledge gap by applying state sequence analysis (SSA) to comprehensively assess individual-level medication use. Psychopharmacological medication use was assessed among Swedish-born, second-generation, non-refugee and refugee first-generation immigrants with incident BD diagnosed in Sweden 2006–2015 (n = 24,578, 16–65 years). Three years of medication-use were conceptualized with SSA as consecutive sequences of three-month periods. Anticonvulsant mood-stabilizer, lithium and antipsychotic use was considered adequate treatment. Typologies were identified by clustering and associated with population groups and covariates applying multinomial logistic regression, yielding odds ratios (OR) for comparison to the majority typology as well as estimated probabilities for each typology. Immigrant populations discontinued medication within 6 months more frequently than Swedish-born (42.1–45.7% vs 36.8%). Transitions from periods lacking medication to adequate treatment showed low likelihood across population groups (8.9–10.1%). Treatment failure (48.3% of refugees, 32.3% of Swedish-born), representing lack of adequate and antidepressant medication, predominated among seven identified typologies. Compared to Swedish-born and treatment failure, adjusted OR for other typologies were lower for refugees (0.3–0.5) and other immigrant groups (0.5–0.8). Adjusting for covariates, highest probabilities for treatment failure were computed for non-refugee (44%) and refugee first-generation immigrants (51%), followed by individuals with low education level (42%) and psychiatric comorbidities (attention-deficit/hyperactivity disorder 38%, substance-use disorder 37%). In conclusion, immigrant groups, particularly refugees, with incident BD are less likely to receive adequate treatment, requiring special emphasis on guideline-conformance.