Use of antidepressant medication is associated with slower response to open-ended psychotherapy for depressed patients

The study uses a quasi-experimental design without random assignment to medication status, introducing potential confounding variables and limiting causal inference about the effects of antidepressants. It relies on observational data collected for other purposes (NMSPOP study) without a specific intervention designed for this research question, leading to potential biases related to patient selection, treatment assignment, and other unmeasured factors.

Therapists and patients were aware of medication status, introducing potential bias in treatment delivery and outcome assessment. Blinding is difficult in psychotherapy research but could have strengthened the study's internal validity by reducing the influence of expectations on outcomes.

Unknown patient adherence to medication regimens and the lack of measured serum levels weaken any conclusions about the direct impact of drug levels on treatment outcomes. This information would enhance the understanding of the medication's role in influencing treatment response.

The sample comes from a Norwegian public mental health service providing open-ended psychotherapy (average 30.6 months), limiting generalizability to other healthcare settings with shorter treatment durations or different patient populations (e.g., private practice, different cultural contexts).

Comorbid disorders are only controlled statistically, not experimentally. Other factors (e.g., treatment expectations, therapist experience, patient motivation, and social support) could influence outcomes. The study controlled for personality disorder status, which could otherwise act as a confounding variable.

While comorbid personality disorders (PD) were controlled statistically, the lack of experimental manipulation of these conditions makes it harder to isolate the specific impact of medication from the overall effect of these complex comorbidities.