The benefit-risk ratio of a long duration of antidepressant drug treatment (ADT) may not be favourable. A pragmatic way to study the consequences of ADT duration in accordance with practice guidelines is to assess its impact on health care use.

To assess the economic impact of ADT duration in a real-life setting through the changes in direct health care costs.

A historical fixed cohort study included 27,134 patients aged 18 and over with a new ADT registered in the national insurance database. The economic impact concerned health care expenditure in the first three months after treatment discontinuation. Generalized linear models compared two groups of treatment duration: less than six months and six months and over.

After adjustment for care costs before and during ADT, gender, age, chronic diseases, welfare entitlement and prescriber specialty, total health care costs (in log) [-0.06 (-0.14;0.01) p = 0.11] and psychiatric care costs (in square root) [-0.08 (-0.41;0.25) p = 0.6] were similar in both groups after ADT discontinuation. Non-psychiatric care costs were significantly lower in the “long treatment duration” group compared with the “short treatment duration” group [-11.4 (-15.8;-7.0) p < 0.0001]. The decreases in total and non-psychiatric care costs after the ADT episode were larger in the “long treatment duration” group compared with the “short treatment duration” group.

With regard to health care costs and global health, ADT of short duration appear less effective than treatment of recommended duration.