Prescription patterns of antidepressants: findings from a US claims database

Abstract

Background:

Introduction of serotonin reuptake inhibitors in the 1990s has increased the use of antidepressants and modified their prescription patterns.

Objective:

To identify reasons for prescriptions of antidepressants and factors associated with absence of a labelled indication on the prescription patterns of antidepressants and healthcare costs in a claims database.

Methods:

Antidepressant users with a new treatment episode with bupropion, citalopram, duloxetine, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline or venlafaxine in 2003 and 2004 were identified in the PharMetrics database. Any ICD-9 code for an approved or clinically-accepted diagnosis for antidepressant treatments (‘diagnosis of interest’) occurring within the month before or after the index claim was considered as a reason for prescription. Socio-demographic and medical characteristics were described between users with and without a diagnosis of interest and analysed using logistic regression.

Results:

A total of 392 409 antidepressant users were identified. Diagnoses of interest were recorded for 46.7% of users, the most frequent diagnosis being depressive disorders (29% of the patients), anxiety disorders (17%) and abuse and dependence (5%). There were no major differences in patterns of diagnoses of interest between the antidepressants except for fluvoxamine and bupropion.

Users without a diagnosis of interest had similar somatic comorbidities and overall baseline costs to users with a diagnosis of interest. However, they used specialised care less often (4.3 vs. 17.8%, OR = 0.50 [0.48; 0.51]), received psychotherapies less frequently (2.7 vs. 26.6%, OR = 0.12 [0.12; 0.12]), and had a shorter duration of use of antidepressants more often (36.9 vs. 28.5%, OR = 1.18 [1.17; 1.20]).

Conclusions:

The reason for prescribing antidepressants was often not reported in claims databases, and although antidepressant users with or without a diagnosis of interest can have similar somatic medical profiles and overall costs, they do not follow the same trajectory in the mental healthcare system. Depending on the research question to be answered, it is therefore important to specify which users are being targeted.

Key words::

• Antidepressant
• Anxiety
• Comorbidities
• Database
• Depression
• Diagnosis of interest
• Prescription

Transparency

Declaration of funding

This study was supported by Lundbeck SAS, Issy-les Moulineaux, France

Declaration of financial/other relationships

D.M. and P.V. have disclosed that they are Lundbeck employees. F.G. has disclosed that he is an employee of, and consultant to, Altipharm, which has received honoraria from Lundbeck in relation with this study. M.L. and M.T. have nothing to disclose.

Some peer reviewers receive honoraria from CMRO for their review work. The peer reviewers of this paper have disclosed that they have no relevant financial relationships.

Acknowledgement

The authors thank Guillaume Hébert, PhD, for his editorial assistance.

Lundbeck participated in study design, data analysis and interpretation.