Using filled prescription sequences to rank antidepressants: A nationwide replication study. (February 2023)

• Record Type: Journal Article
• Authors: Ouazana-Vedrines, Charles; Lesuffleur, Thomas; Denis, Pierre; Hoertel, Nicolas; Limosin, Frédéric; Rachas, Antoine; Tuppin, Philippe; Lemogne, Cédric
• Is Part Of: Journal of psychiatric research. Volume 158(2023)
• Page Start: 180
• Page End: 184
• Publication Date: 2023-02
• Subjects: Administrative claims -- Healthcare -- Reimbursement -- Big data -- Antidepressive agents -- Cohort studies -- Depressive disorder -- Acceptability -- Ranking; RCT randomized controlled trial -- SNDS système national des données de santé -- GP general practitioner -- aOR adjusted odd ratio -- CI confidence interval -- CNAM caisse nationale d'assurance maladie; Psychiatry -- Periodicals; Mental Disorders -- Periodicals; Maladies mentales -- Périodiques; Psychiatry; Electronic journals; Periodicals; 616.89005
• DOI: 10.1016/j.jpsychires.2022.12.030
• Languages: English
• ISSNs: 0022-3956

Abstract

Abstract: Ranking antidepressants according to their acceptability (i.e., a combination of both efficacy and tolerability) in the general population may help choosing the best first-line medication. This study aimed to replicate the results of a proof-of-concept study ranking anti-depressants according to the proportion of filled prescription sequences consistent with a continuation of the first treatment versus those consistent with a change. We used a nationwide cohort from the French national health data system (SNDS) to support the use of this method as a widely available tool to rank antidepressant treatments in real life settings. About 1.2 million people were identified as new antidepressant users in the SNDS in 2011. The outcome was clinical acceptability as measured by the continuation/failure ratio over the six-month period following the introduction of the first-line treatment. Continuation was defined as at least two refills of the same treatment. Failure was defined as a psychiatric hospitalization, death or at least one filled prescription of another antidepressant, an antipsychotic medication, or a mood-stabilizer. Adjusted Odds Ratios (aOR) and 95% Confidence Interval (CI) were computed through multivariable binary logistic regressions. We ranked antidepressant medications according to clinical acceptability. Escitalopram again was the most acceptable option, and the five following antidepressants were the same as in the replication sample of the proof-of-concept study, in order Fluoxetine, Paroxetine, Sertraline, Citalopram and Venlafaxine with aOR (95% CI) for continuation ranging from 0.79 (0.77–0.81) to 0.66 (0.64–0.67). The present study provides evidence that filled prescription sequences is a widely available, robust and reproductible tool to rank antidepressant treatments in real life settings.