Detecting delirium in elderly medical emergency patients: validation and subsequent modification of the German Nursing Delirium Screening Scale.

Detecting delirium in elderly emergency patients is critical to their outcome. The Nursing Delirium Screening Scale (Nu-DESC) is a short, feasible instrument that allows nurses to systematically screen for delirium. This is the first study to validate the Nu-DESC in a German emergency department (ED). The Nu-DESC was implemented in a high-volume, interdisciplinary German ED. A consecutively recruited sample of medical patients aged ≥ 70 years was screened by assigned nurses who performed the Nu-DESC as part of their daily work routine. The results were compared to a criterion standard diagnosis of delirium. According to the criterion standard diagnosis, delirium was present in 47 (14.9%) out of the 315 patients enrolled. The Nu-DESC shows a good specificity level of 91.0% (95% CI 87.0-94.2), but a moderate sensitivity level of 66.0% (95% CI 50.7-79.1). Positive and negative likelihood ratios are 7.37 (95% CI 4.77-11.36) and 0.37 (95% CI 0.25-0.56), respectively. In an exploratory analysis, we find that operationalizing the Nu-DESC item "disorientation" by specifically asking patients to state the day of the week and the name of the hospital unit would raise Nu-DESC sensitivity to 77.8%, with a specificity of 84.6% (positive and negative likelihood ratio of 5.05 and 0.26, respectively). The Nu-DESC shows good specificity but moderate sensitivity when performed by nurses during their daily work in a German ED. We have developed a modified Nu-DESC version, resulting in markedly enhanced sensitivity while maintaining a satisfactory level of specificity.

Validation of the Brief Confusion Assessment Method for Screening Delirium in Elderly Medical Patients in a German Emergency Department.

BACKGROUND: Delirium is frequent in elderly patients presenting in the emergency department (ED). Despite the severe prognosis, the majority of delirium cases remain undetected by emergency physicians (EPs). At the time of our study there was no valid delirium screening tool available for EDs in German-speaking regions. We aimed to evaluate the brief Confusion Assessment Method (bCAM) for a German ED during the daily work routine. METHODS: We implemented the bCAM into practice in a German interdisciplinary high-volume ED and evaluated the bCAM's validity in a convenience sample of medical patients aged ≥ 70 years. The bCAM, which assesses four core features of delirium, was performed by EPs during their daily work routine and compared to a criterion standard based on the criteria for delirium as described in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. RESULTS: Compared to the criterion standard, delirium was found to be present in 46 (16.0%) of the 288 nonsurgical patients enrolled. The bCAM showed 93.8% specificity (95% confidence interval [CI] = 90.0%-96.5%) and 65.2% sensitivity (95% CI = 49.8%-78.7%). Positive and negative likelihood ratios were 10.5 and 0.37, respectively, while the odds ratio was 28.4. Delirium was missed in 10 of 16 cases, since the bCAM did not indicate altered levels of consciousness and disorganized thinking. The level of agreement with the criterion standard increased for patients with low cognitive performance. CONCLUSION: This was the first study evaluating the bCAM for a German ED and when performed by EPs during routine work. The bCAM showed good specificity, but only moderate sensitivity. Nevertheless, application of the bCAM most likely improves the delirium detection rate in German EDs. However, it should only be applied by trained physicians to maximize diagnostic accuracy and hence improve the bCAM's sensitivity. Future studies should refine the bCAM.