Predictive value of predictive score for ICU delirium for delirium in ICU patients

Objective To evaluate the predictive value of predictive score for ICU delirium (PSID) for delirium in ICU patients.

Methods The clinical data of 133 patients hospitalized in the department of intensive medicine of Huangshan Shoukang Hospital were retrospectively analyzed. According to whether the patients had delirium during their stay in the ICU, they were divided into a delirium group (n=42) and a non-delirium group (n=91). The related clinical data of patients between two groups were compared. ROC curve analysis was used to evaluate the predictive value of PSID score for delirium in ICU patients, which was compared with delirium risk assessment score (DRAS) and acute physiology and chronic health evaluation Ⅱ (APACHEⅡ). Multiple logistic regression analysis was performed to evaluate the risk factors for delirium in ICU patients.

Results Hypoxemia, deep sedation, use of benzodiazepine drug, and mechanical ventilation of patients in the delirium group were significantly higher than those in the non-delirium group (P<0.01). The incidence of delirium was 10.0% (6/60) with PSID score of 0-4 points, 46.3% (31/67) with PSID score of 5-9 points, and 83.3% (5/6) with PSID score of more than 10 points. There was a significantly positive correlation between PSID score and the incidence of delirium (r=0.891, P<0.01). The area under the curve of PSID score predicting the risk of delirium in ICU patients was significantly higher than that of DRAS score and APACHE Ⅱ score (P<0.01). Multivariate logistic regression analysis showed that deep sedation, APACHE Ⅱ score, PSID score, and DRAS score were all risk factors for delirium in ICU patients (P<0.05 to P<0.01).

Conclusions PSID score has good predictive value for delirium in ICU patients, and higher score indicates higher risk of delirium, which was worthy of clinical application.