BACKGROUND: The treatment of schizophrenia requires a prolonged, multidimensional intervention that includes antipsychotic drugs. Treatment adherence is essential to effectively control the disorder. Shared decision-making (SDM) is a strategy, supported by numerous practical and ethical arguments, that seeks to involve patients in the therapeutic process to improve treatment adherence and satisfaction. The use of this model in mental health has been limited for many intrinsic and extrinsic reasons. The results of clinical trials conducted to date have largely been disappointing, potential due to study design-related limitations. AIM/QUESTION: To evaluate the efficacy, in terms of treatment adherence and improvement in clinical variables, such as severity of symptoms, days of hospitalization or insight, of a carefully timed SDM model initiated immediately prior to hospital discharge in patients with schizophrenia. METHODS: Single-blind, randomized clinical trial in an acute psychiatric care unit within the Andalusian Health Department to compare SDM (experimental group) to treatment as usual (TAU; control group) in a sample of patients hospitalized for an acute episode of schizophrenia or schizoaffective disorder. The study was performed between January 2014 and June 2017. The experimental group participated in SDM sessions prior to discharge with regular booster sessions over the one-year follow-up. The health care team responsible for SDM was predisposed to concordance (LatCon II scale) and received specific training in SDM. A hierarchical multiple linear regression analysis was performed to evaluate the factors independently associated with adherence, controlling for sociodemographic, clinical, and admission-related variables. Variables were assessed at admission, discharge and at 3, 6 and 12 months after discharge during the one year follow up. BARS, DAI, WAI-S, COMRADE and PANSS were used to evaluate adherence, attitude to treatment, therapeutic alliance, satisfaction and confidence with decision and clinical status, respectively. RESULTS: A total of 227 schizophrenic patients hospitalized with acute decompensation were evaluated; of these, 102 met all inclusion criteria and were included in the study. Most patients (95%) had prior experience with antipsychotics and most (82%) had experienced related side effects. Despite randomization, psychopathologic severity was greater in the experimental group, with a mean (SD) PANSS score of 104.08 (80) vs. 93.45 (20.30) (p < 0.05). The final regression model to explain adherence was significant (adjusted R2 = 0.384; F [df= 6] = 4.386; p < 0.001), with a direct, significant and independent association with SDM mediated by the number of booster sessions. DISCUSSION: Shared decision making with booster sessions appears to increase treatment adherence in patients with severe mental disorders. IMPLICATION ON PRACTICE: Ethical, practical, and clinical reasons support the use of strategies designed promote the use of long-term, shared decision-making in psychiatric patients, especially in schizophrenia spectrum disorder.