Discontinuation of medications for alcohol use disorder in the United States: Patient, prescriber and medication predictors.

BACKGROUND AND AIMS: Little is known about medication for alcohol use disorder (MAUD) adherence. The goal of this study was to examine patient, prescriber and medication characteristics associated with time to discontinuation of MAUD.

DESIGN, SETTING AND

PARTICIPANTS

Observational cohort study from Optum Labs de-identified administrative claims data of a United States national sample of commercial and Medicare Advantage enrollees with alcohol use disorder (AUD) who initiated MAUD (naltrexone, acamprosate, disulfiram, topiramate) January 2017-June 2024 (n = 51 915). Using an accelerated failure time (AFT) log logistic survival model, we examined MAUD discontinuation, adjusting for patient demographic and clinical characteristics established within 6 months prior to or on date of initiation, specific MAUD initiated with, MAUD prescriber type and study year. We defined effect size of at least 10% change in AFT survival (i.e. AFT estimates > +/- 0.100) as being of clinical significance.

MEASUREMENTS: Time-to-discontinuation of MAUD, defined as a gap in pharmacy fills or medication administrations of MAUD days-supply >7 days.

FINDINGS: In unadjusted analyses, the median time to MAUD discontinuation was 30 days; median days-supply prior to discontinuation varied by medication (e.g. 30 days for acamprosate and naltrexone, 56 for disulfiram, 60 for topiramate). Adjusted analyses showed that among the patient characteristics, longer MAUD duration was associated with older age [e.g. AFT duration estimate (95% confidence interval) for age ≥65 vs. 18-30 = 0.164 (0.088-0.240)], while lower county median household income was associated with shorter duration [e.g. <$40 000 vs. ≥$200 000: -0.106 (-0.136 to -0.076)]. MAUD initiations via topiramate (vs. naltrexone) were associated with longer duration [0.339 (0.311-0.366)], as were initiations via disulfiram [0.103 (0.076-0.131)]. MAUD duration increased in recent study years, except 2024 [(e.g. 2023 vs. 2017 = 0.116 0.087-0.144)].

CONCLUSIONS

Among the few individuals in the United States who receive medication for alcohol use disorder (MAUD), early discontinuation is common, minimizing the opportunity for MAUD to support patients' recovery goals. While MAUD duration has increased in recent years, the improvements are small.