Background <p>Despite extensive research on adherence in first-episode psychosis (FEP), reliance on self-report or clinician judgment has limited accuracy and obscured nuanced clinical associations.</p> Objective <p>This longitudinal study examined the predictive value of psychopathology, insight, and functioning for objectively measured treatment adherence across a one-year follow-up using therapeutic drug monitoring (TDM).</p> Methods <p>Seventy-eight FEP patients were assessed at baseline, 2 months, and 12 months. Clinical measures included the Brief Psychiatric Rating Scale (BPRS), Scale for the Assessment of Negative Symptoms (SANS), Social and Occupational Functioning Assessment Scale (SOFAS), and the Schedule for the Assessment of Insight–Expanded version (SAI-E). Adherence was defined by comparing plasma antipsychotic concentrations to prescribed doses according to international consensus guidelines. Generalized linear mixed models (GLMMs) were used to model time-dependent associations.</p> Results <p>While overall adherence declined, time alone was not a significant predictor. Instead, current positive symptom severity (approximately 4–5% lower odds of adherence per BPRS point) was associated with lower odds of adherence, while prior non-adherence predicted subsequent adherence. In contrast, insight, negative symptoms, and functioning showed no consistent association with adherence. A marginal association between improved functioning and better adherence was observed at 12 months.</p> Conclusions <p>Adherence in FEP appears to fluctuate in relation to current clinical state rather than follow a stable linear pattern. These findings highlight the value of objective TDM monitoring and early intervention, supporting predictive models that incorporate evolving symptom patterns rather than static baseline measures.</p> Trial registration <p>Clinical trial number: not applicable.</p>