Value-based Healthcare: How Can Large Language Model (LLM) Technology be Integrated With Patient-reported Outcomes?

People at ultra-high risk (UHR) of developing psychosis experience early subthreshold symptoms, which, although they do not meet the criteria for full-blown psychotic disorders, are associated with a high level of distress and a number of negative mental health consequences. This putative prodromal period, which can last from weeks to years (), is characterized by functional and social impairment, the occurrence of subclinical, attenuated psychotic symptoms, and the presence of additional comorbid psychiatric disorders. This stage often precedes the development of more severe psychotic states, with approximately 25 % of individuals progressing to develop a threshold psychotic disorder within 3 years, with enhanced risk persisting for up to 10 years (). Importantly, UHR individuals who do not transition to psychosis remain at significant risk for continued attenuated psychotic symptoms, as well as other persistent or recurrent conditions such as mood and anxiety disorders (; ). The awareness that early signs may precede the onset of the disease has drawn attention to the need to identify people who may be in such a risk group. Early recognition of risk states is crucial as it allows for timely intervention to alleviate ongoing difficulties and reduce the likelihood of progression to not only full-blown psychotic disorder but also other serious mental health conditions. Indeed, over the last two decades, a number of randomized controlled trials (RCTs), systematic reviews, and meta-analyses on the efficacy of various therapeutic interventions for UHR patients have been published (). The results do not fully determine the superiority of one type of treatment over another. Nonetheless, they indicate significant benefits and efficacy in reducing the severity of clinical symptoms and delaying (or preventing) the transition to psychosis (; , ; ), especially following psychosocial interventions such as cognitive behavioral therapy (CBT) (; ; ). The current research presents secondary analyses from the Staged Treatment in Early Psychosis (STEP) study (; ), a sequential multiple assignment randomized trial (SMART) conducted within the clinical program at Orygen, Melbourne, Australia. The trial recruited young individuals aged 12 to 25 years meeting the UHR criteria. The intervention was delivered in three steps: Step 1 involved Support and Problem Solving (SPS); Step 2 compared SPS with Cognitive Behavioural Case Management (CBCM); and Step 3 compared CBCM combined with antidepressant medication versus CBCM with placebo. Participants who did not respond to treatment at a given step progressed to the next stage (please refer to the section for further details on the study design). The results of the STEP study showed no significant differences in symptom and functional outcomes, as well as transition to psychosis rates between the treatment modalities used. However, the results demonstrated overall functional and symptomatic improvement following the intervention. For instance, it has been shown that decrease in cognitive biases six months into treatment has been associated with greater improvements in general psychopathology, depressive symptoms, and quality of life among UHR patients (). Although this effect did not significantly differ between treatment groups (which included CBT-based interventions and support and problem-solving), it is essential to highlight that cognitive biases can be a crucial mediator or even a direct target of treatment for subthreshold psychotic symptoms. This is consistent with a recent meta-analysis suggesting that changes in cognitive biases in response to treatment (which may be one of the key mechanisms at work in CBT-based interventions) should be considered as the primary outcome of treatment effectiveness studies to better understand how modification of cognitive biases leads to improvements in psychotic symptoms ( ). The cognitive model of psychosis suggests that, in light of pre-existing vulnerabilities (i.e., biopsychosocial predispositions), disturbed cognitive processes lead to the development of anomalous experiences, which, under the influence of negative emotional states, contribute to the formation of psychotic symptoms and thus to the onset or maintenance of psychotic disorders (; ). Distorted cognitive processes include cognitive biases, such as jumping to conclusions (JTC) and belief inflexibility (reasoning biases), attention to threat and aberrant salience (attentional biases), as well as external attributions and personalizing (attributional biases) (; ). Recent systematic reviews and meta-analyses revealed significant associations between subclinical psychotic symptoms and cognitive biases in individuals at high risk of developing psychosis, and, in line with the cognitive model of psychosis, emphasized the contribution of cognitive biases to the emergence and maintenance of both positive and negative symptoms (; ). It is noteworthy that the STEP study reported a su