Medication adherence following pharmacogenomic testing in patients with major depressive disorder.

Weighted multi-gene pharmacogenomic test sorting psychiatric medications into three levels of gene-drug interaction, used to guide antidepressant switching.

Around half of patients treated for major depressive disorder discontinue their antidepressant within six months, increasing relapse risk. This retrospective observational study used de-identified claims data from the Optum Labs Data Warehouse in 6,224 adults who received a weighted multi-gene pharmacogenomic test between 1 January 2015 and 30 September 2021 and then switched medication, with the test report sorting psychiatric medications into three categories: no known gene-drug interaction and moderate interaction, both congruent, and significant interaction, incongruent. Patients were grouped according to the change in congruency of their worst-congruency medication 90 days before and after testing, then followed for 180 days after the switch for adherence, measured as proportion of days covered, and discontinuation, defined as a gap of at least 45 days in fills. The incongruent-to-congruent group had the highest adherence (mean proportion of days covered 0.65, SD 0.33) compared with the congruent-to-incongruent (0.58, SD 0.34) and no-change groups (0.61, SD 0.34) (p < 0.05), and the lowest discontinuation rate (46% versus 55% and 50%, p < 0.05).

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

Groups are not formed by randomisation but by prescriber behaviour after the report: those who switch to a congruent medication are those whose clinician engaged with the result, an engagement itself associated with better adherence. The absolute gap remains modest, 4 to 7 points of proportion of days covered, and 46% of patients in the best group still discontinue treatment. The finding deserves to be read alongside the randomised trial published the same week: this favourable observational signal does not survive randomisation once the endpoint becomes patient well-being.

Analysis by Dr Thibaut Benquey