This page was translated automatically from Turkish. Read the original

Bu sayfa, 26. Dünya Psikiyatri Kongresi'nde sunulan çalışmanın yapay zekâ desteğiyle hazırlanmış Türkçe özetidir. Çalışmanın sahibi sunan yazarlardır; atıf için orijinal kaynağa başvurun.

SO515

Clustering of clinical and social profiles in psychiatric emergencies in Paraguay’s public health system: implications for clinical decision-making and management

CLUSTERING OF CLINICAL-SOCIAL PROFILES IN PSYCHIATRIC EMERGENCIES OF PARAGUAY'S PUBLIC HEALTH SYSTEM:IMPLICATIONS FOR CLINICAL AND MANAGEMENT DECISION-MAKING

Five distinct clinical and social profiles were identified among psychiatric emergency presentations, demonstrating that admission decisions are independently influenced not only by acute symptom severity but also by social vulnerability.

The aim was to identify clinical and social profiles among psychiatric emergency presentations in low-resource systems and assess their associations with Psychiatric Intensive Care Unit (PICU) severity criteria and initial management decisions. Structured emergency records of 229 adults assessed at Paraguay’s National Psychiatric Hospital between January and June 2025 were examined using a retrospective analytical cross-sectional design; multiple imputation, descriptive and bivariate statistics, TwoStep cluster analysis, and binary logistic regression to identify predictors of admission were applied. The mean age of the sample was 31.4 years; 52.0% were women and 53.3% were unemployed; 41.0% had a history of psychiatric admission, 37.1% had discontinued medication, 50.7% had medical comorbidities, and 16.6% had active suicidal ideation with a plan. Regarding PICU severity criteria, psychiatric deterioration was present in 95.2%, failure of outpatient treatment in 72.5%, risk of self-harm in 62.0%, and risk of aggression toward others in 32.3%. Clustering revealed five profiles: vulnerable but compensated patients receiving outpatient follow-up, suicide risk with impulsivity, chronic decompensation—the largest cluster (n=73), acute agitation or a first episode, and decompensation compounded by severe social vulnerability; validity was supported by a silhouette coefficient of 0.34 and 67.8% of the variance accounted for by between-group differences. Significant differences between clusters were found in age (F(4,224)=3.92; p=0.004), sex (chi-square=8.21; p=0.042), and suicidal ideation (chi-square=21.36; p<0.001). The main predictors of admission were active suicidal ideation with a plan (odds ratio (OR)=3.14; 95% confidence interval 1.38 to 7.11), marked psychiatric deterioration (OR=2.67; 1.15 to 6.21), and failure of outpatient treatment (OR=2.45; 1.04 to 5.76). The authors suggest that recognizing this multidimensional heterogeneity, which diagnosis-based classification fails to capture, could improve triage, early risk stratification, and resource allocation, particularly in settings where inpatient beds and community-based follow-up options are limited.

Tüm WCP 2026 özetleri