Large language models can support generation of standardized discharge summaries - A retrospective study utilizing ChatGPT-4 and electronic health records.

OBJECTIVE: To evaluate whether psychiatric discharge summaries (DS) generated with ChatGPT-4 from electronic health records (EHR) can match the quality of DS written by psychiatric residents. METHODS: At a psychiatric primary care hospital, we compared 20 inpatient DS, written by residents, to those written with ChatGPT-4 from pseudonymized residents' notes of the patients' EHRs and a standardized prompt. 8 blinded psychiatry specialists rated both versions on a custom Likert scale from 1 to 5 across 15 quality subcategories. The primary outcome was the overall rating difference between the two groups. The secondary outcomes were the rating differences at the level of individual question, case, and rater. RESULTS: Human-written DS were rated significantly higher than AI (mean ratings: human 3.78, AI 3.12, p < 0.05). They surpassed AI significantly in 12/15 questions and 16/20 cases and were favored significantly by 7/8 raters. For "low expected correction effort", human DS were rated as 67 % favorable, 19 % neutral, and 14 % unfavorable, whereas AI-DS were rated as 22 % favorable, 33 % neutral, and 45 % unfavorable. Hallucinations were present in 40 % of AI-DS, with 37.5 % deemed highly clinically relevant. Minor content mistakes were found in 30 % of AI and 10 % of human DS. Raters correctly identified AI-DS with 81 % sensitivity and 75 % specificity. DISCUSSION: Overall, AI-DS did not match the quality of resident-written DS but performed similarly in 20% of cases and were rated as favorable for "low expected correction effort" in 22% of cases. AI-DS lacked most in content specificity, ability to distill key case information, and coherence but performed adequately in conciseness, adherence to formalities, relevance of included content, and form. CONCLUSION: LLM-written DS show potential as templates for physicians to finalize, potentially saving time in the future.

Negative Symptoms and Their Associations With Other Clinical Variables and Working Memory Across the Schizophrenia Spectrum and Bipolar Disorder.

Negative symptoms (NS) of schizophrenia spectrum disorders (SSD) are also prevalent in bipolar disorder I (BD-I) and show associations with impaired working memory (WM). However, empirical work on their relationship to other clinical factors across SSD and BD-I is sparse. Here, we characterized the associations of NS with key clinical variables and WM capacity across a combined sample of SSD and BD. We included 50 outpatients with SSD and 49 with BD-I and assessed NS domains using SANS global scores for avolition-apathy, anhedonia-asociality, alogia, and blunted affect. We assessed the transdiagnostic relationship between NS and other clinical variables, including positive symptoms, disorganization, depressive symptoms, and antipsychotic medication, using multiple regressions. The strength of these associations was further determined through dominance analyses. Finally, we used multiple regression to assess the relationship between NS domains and WM. To assess the generalizability of transdiagnostic associations, analyses were repeated in each diagnostic group separately. Across SSD and BD-I, disorganization was associated with avolition-apathy and anhedonia-asociality and depressive symptoms additionally predicted anhedonia-asociality. Antipsychotic dose was associated with blunted affect while group differences only predicted alogia. Higher avolition-apathy was related to impaired WM transdiagnostically, partially mediated by the severity of disorganization, whereas only in BD-I higher anhedonia-asociality was associated with better WM capacity. This study demonstrated transdiagnostic associations of both avolition-apathy and anhedonia-asociality with disorganization and identified avolition-apathy as a potential transdiagnostic predictor of WM impairments. Overall, our findings highlight the importance of understanding the relationship between NS domains and other clinical factors with cognitive function across SSD and BD.

Examining the impact of substance use on hospital length of stay in schizophrenia spectrum disorder: a retrospective analysis.

BACKGROUND: Among patients diagnosed with schizophrenia, the presence of substance use poses an aggravating comorbidity, exerting a negative impact on the course of the disease, adherence to therapeutic regimens, treatment outcomes, duration of hospital stays, and the frequency of hospitalizations. The primary objective of the present study is to investigate the relationship between comorbid substance use disorders, antipsychotic treatment, and the length of stay in individuals hospitalized for treatment of schizophrenia. METHODS: We conducted a retrospective analysis of electronic health records spanning a 12-month period, specifically focusing on adult patients diagnosed with schizophrenia who were discharged from the University Hospital of Psychiatry Zurich between January and December 2019. We documented the number and types of diagnosed substance use disorder, the antipsychotic treatment, the length of stay, and the number of previous hospitalizations for each patient. RESULTS: Over a third (n = 328; 37.1%) of patients with schizophrenia had comorbid substance use with cannabis being the most frequent consumed substance. Patients with substance use (either single or multiple) were more frequently hospitalized; those with multiple substance use more frequently than those with a single substance use (F(2, 882) = 69.06; p < 0.001). There were no differences regarding the rate of compulsory admission. Patients with no substance use had a lower HoNOS score at discharge (F(2, 882) = 4.06). Patients with multiple substance use had a shorter length of stay (F(2, 882) = 9.22; p < 0.001), even after adjusting for duration of illness, previous hospitalizations, diagnosis, and antipsychotic treatment. CONCLUSIONS: In patients with schizophrenia, comorbid single or multiple substance use has a relevant negative impact on treatment and thus on the course of disease. Substance use in patients with schizophrenia should therefore receive special attention in order to reduce re-hospitalization rates and improve the clinical outcome.

Impact of psychopathology on day-to-day living in patients with schizophrenia: A network analysis.

Although the relationship between schizophrenia and disability is well established, the association between the symptoms of the disorder and functional domains remains unclear. The current study explored the nuances of the relationship between symptoms and domains of functioning in a sample of 1127 patients with schizophrenia. We assessed the symptoms of schizophrenia with the Positive and Negative Syndrome Scale (PANSS) and psychosocial functioning with the mini-ICF-APP (mini-International Classification of Functioning Rating for Limitations of Activities and Participation in Psychological Disorders). The mean PANSS score was 94.28 (27.20), and the mean mini-ICF-APP score was 25.25 (8.96), both of which are indicative of severe symptom load and impairment. We were able to show a strong relationship and overlap between symptoms and disability in patients with schizophrenia. We identified several symptoms related to functional impairment. Deficits in judgment and abstract thinking contribute to impairment through poor adherence (to routines and compliance with rules) and difficulties in planning and organizing. We believe that in schizophrenia, symptoms and their interactions constitute a disorder beyond any single manifestation. Furthermore, we suggest that cognitive testing and cognitive treatment should become part of the standard of care for patients with schizophrenia.

Relationship between antipsychotic medication and aggressive events in patients with a psychotic disorder hospitalized for treatment.

BACKGROUND: Disruptive and aggressive behavior is frequent in patients with a psychotic disorder; furthermore, it is a recurrent reason for compulsory admission. Even during treatment, many patients continue to show aggressive behavior. Antipsychotic medication is posed to have anti-aggressive properties; its prescription is a common strategy for the treatment (and prevention) of violent behavior. The present study aims to investigate the relation between the antipsychotic class, according to the dopamine D2-Receptor binding affinity (i.e., "loose" - "tight binding"), and aggressive events perpetrated by hospitalized patients with a psychotic disorder. METHODS: We conducted a four-year retrospective analysis of legally liable aggressive incidents perpetrated by patients during hospitalization. We extracted patients' basic demographic and clinical data from electronic health records. We used the Staff Observation Aggression Scale (SOAS-R) to grade the severity of an event. Differences between patients with a "loose" or "tight-binding" antipsychotic were analyzed. RESULTS: In the observation period, there were 17,901 direct admissions; and 61 severe aggressive events (an incidence of 0.85 for every 1,000 admissions year). Patients with a psychotic disorder perpetrated 51 events (incidence of 2.90 for every 1,000 admission year), with an OR of 15.85 (CI: 8.04-31.25) compared to non-psychotic patients. We could identify 46 events conducted by patients with a psychotic disorder under medication. The mean SOAS-R total score was 17.02 (2.74). The majority of victims in the "loose-binding" group were staff members (73.1%, n = 19), while the majority of victims in the "tight-binding" group were fellow patients (65.0%, n = 13); (X2(3,46) = 19.687; p < 0.001). There were no demographic or clinical differences between the groups and no differences regarding dose equivalents or other prescribed medication. CONCLUSIONS: In aggressive behaviors conducted by patients with a psychotic disorder under antipsychotic medication, the dopamine D2-Receptor affinity seems to have a high impact on the target of aggression. However, more studies are needed to investigate the anti-aggressive effects of individual antipsychotic agents.

Use of Long-Acting Injectable Antipsychotics in Inpatients with Schizophrenia Spectrum Disorder in an Academic Psychiatric Hospital in Switzerland.

Long-acting injectable antipsychotics (LAIs) offer many benefits to patients with schizophrenia spectrum disorder (SSD). They are used with very different frequencies due to questions of eligibility or patients and prescribers' attitudes towards LAI use. We assessed the prescribing rates of LAIs in a large academic psychiatric hospital with a public service mandate in Switzerland and compared them with other countries and health care systems. To our knowledge, this study is the first to investigate inpatient LAI use in Europe. Medical records of all patients diagnosed with SSD discharged from the Clinic of Adult Psychiatry of the University Hospital of Psychiatry Zurich over a 12 month period from January to December 2019 were evaluated regarding the prescribed antipsychotics at the time of discharge. The rates of use of LAIs among all patients and among patients receiving LAI-eligible antipsychotic substances were assessed retrospectively. We assessed records of 885 patients with SSD. Among all cases, 13.9% received an LAI. Among patients who received antipsychotic medication that was eligible for LAI use (n = 434), 28.1% received an agent as an LAI. LAI use included paliperidone palmitate (69.9%), aripiprazole monohydrate (14.6%), risperidone (4.9%) and first-generation LAIs (9.8%). Compared to international frequencies of LAI administration, the prescription rate of LAIs in SSD patients was low. Further studies will evaluate patient- and prescriber-related reasons for this low rate.

Orbitofrontal-Striatal Structural Alterations Linked to Negative Symptoms at Different Stages of the Schizophrenia Spectrum.

Negative symptoms such as anhedonia and apathy are among the most debilitating manifestations of schizophrenia (SZ). Imaging studies have linked these symptoms to morphometric abnormalities in 2 brain regions implicated in reward and motivation: the orbitofrontal cortex (OFC) and striatum. Higher negative symptoms are generally associated with reduced OFC thickness, while higher apathy specifically maps to reduced striatal volume. However, it remains unclear whether these tissue losses are a consequence of chronic illness and its treatment or an underlying phenotypic trait. Here, we use multicentre magnetic resonance imaging data to investigate orbitofrontal-striatal abnormalities across the SZ spectrum from healthy populations with high schizotypy to unmedicated and medicated first-episode psychosis (FEP), and patients with chronic SZ. Putamen, caudate, accumbens volume, and OFC thickness were estimated from T1-weighted images acquired in all 3 diagnostic groups and controls from 4 sites (n = 337). Results were first established in 1 discovery dataset and replicated in 3 independent samples. There was a negative correlation between apathy and putamen/accumbens volume only in healthy individuals with schizotypy; however, medicated patients exhibited larger putamen volume, which appears to be a consequence of antipsychotic medications. The negative association between reduced OFC thickness and total negative symptoms also appeared to vary along the SZ spectrum, being significant only in FEP patients. In schizotypy, there was increased OFC thickness relative to controls. Our findings suggest that negative symptoms are associated with a temporal continuum of orbitofrontal-striatal abnormalities that may predate the occurrence of SZ. Thicker OFC in schizotypy may represent either compensatory or pathological mechanisms prior to the disease onset.

Apathy is not associated with reduced ventral striatal volume in patients with schizophrenia.

BACKGROUND: A growing body of neuroimaging research has revealed a relationship between blunted activation of the ventral striatum (VS) and apathy in schizophrenia. In contrast, the association between reduced striatal volume and apathy is less well established, while the relationship between VS function and structure in patients with schizophrenia remains an open question. Here, we aimed to replicate previous structural findings in a larger independent sample and to investigate the relationship between VS hypoactivation and VS volume. METHODS: We included brain structural magnetic resonance imaging (MRI) data from 60 patients with schizophrenia (SZ) that had shown an association of VS hypoactivation with apathy during reward anticipation and 58 healthy controls (HC). To improve replicability, we applied analytical methods employed in two previously published studies: Voxel-based morphometry and the Multiple Automatically Generated Templates (MAGeT) algorithm. VS and dorsal striatum (DS) volume were correlated with apathy correcting for age, gender and total brain volume. Additionally, left VS activity was correlated with left VS volume. RESULTS: We failed to replicate the association between apathy and reduced VS volume and did not find a correlation with DS volume. Functional and structural left VS measures exhibited a trend-level correlation (rs = 0.248, p = 0.067, r2 = 0.06). CONCLUSIONS: Our present data suggests that functional and structural striatal neuroimaging correlates of apathy can occur independently. Replication of previous findings may have been limited by other factors (medication, illness duration, age) potentially related to striatal volume changes in SZ. Finally, associations between reward-related VS function and structure should be further explored.

Subtle white matter alterations in schizophrenia identified with a new measure of fiber density.

Altered cerebral connectivity is one of the core pathophysiological mechanism underlying the development and progression of information-processing deficits in schizophrenia. To date, most diffusion tensor imaging (DTI) studies used fractional anisotropy (FA) to investigate disrupted white matter connections. However, a quantitative interpretation of FA changes is often impeded by the inherent limitations of the underlying tensor model. A more fine-grained measure of white matter alterations could be achieved by measuring fiber density (FD) - a novel non-tensor-derived diffusion marker. This study investigates, for the first time, FD alterations in schizophrenia patients. FD and FA maps were derived from diffusion data of 25 healthy controls (HC) and 21 patients with schizophrenia (SZ). Using tract-based spatial statistics (TBSS), group differences in FD and FA were investigated across the entire white matter. Furthermore, we performed a region of interest (ROI) analysis of frontal fasciculi to detect potential correlations between FD and positive symptoms. As a result, whole brain TBSS analysis revealed reduced FD in SZ patients compared to HC in several white matter tracts including the left and right thalamic radiation (TR), superior longitudinal fasciculus (SLF), corpus callosum (CC), and corticospinal tract (CST). In contrast, there were no significant FA differences between groups. Further, FD values in the TR were negatively correlated with the severity of positive symptoms and medication dose in SZ patients. In summary, a novel diffusion-weighted data analysis approach enabled us to identify widespread FD changes in SZ patients with most prominent white matter alterations in the frontal and subcortical regions. Our findings suggest that the new FD measure may be more sensitive to subtle changes in the white matter microstructure compared to FA, particularly in the given population. Therefore, investigating FD may be a promising approach to detect subtle changes in the white matter microstructure of altered connectivity in schizophrenia.

Cerebral blood flow in striatal regions is associated with apathy in patients with schizophrenia.

BACKGROUND: Striatal dysfunction has been proposed as a pathomechanism for negative symptoms in schizophrenia. There is consensus that negative symptoms can be grouped into 2 dimensions: apathy and diminished expression. Recent studies suggest that different neural mechanisms underlie these dimensions, but the relationship between regional resting-state cerebral blood flow (rCBF) and negative symptom dimensions has not been investigated. METHODS: This study included 29 patients with schizophrenia and 20 healthy controls. We measured rCBF in the striatum using arterial spin labelling (ASL) MRI. We assessed negative symptoms using the Brief Negative Symptom Scale. RESULTS: In the ventral and dorsal striatum, rCBF was not different between patients with schizophrenia and controls. However, we did find a positive association between the severity of apathy and increased rCBF in the ventral and dorsal striatum in patients with schizophrenia. This effect was not present for diminished expression. LIMITATIONS: All patients were taking atypical antipsychotics, so an effect of antipsychotic medication on rCBF could not be excluded, although we did not find a significant association between rCBF and chlorpromazine equivalents. CONCLUSION: The main finding of this study was a specific association between increased striatal rCBF and the negative symptom dimension of apathy. Our results further support the separate assessment of apathy and diminished expression when investigating the neural basis of negative symptoms. The ASL technique can provide a direct and quantitative approach to investigating the role of rCBF changes in the pathophysiology of negative symptoms.