Predictive model and discriminant analysis of the development of dementia in patients with delirium in the emergency department

The main aim of our study is to know the sociodemographic, clinical, analytical, and functional variables that predict the probability of developing dementia in patients with delirium who attend the emergency room. All patients with delirium (n=45) from the emergency room who were admitted to the Geriatrics service of the General University Hospital of Ciudad Real (HGUCR) in 2016-2018 and met the inclusion and exclusion criteria were included. Subsequently, we ran a bivariate and multivariate analysis of the variables that predicted a diagnosis of dementia at six months and a discriminant analysis. 15.6% of patients presented dementia at six months of follow-up, 22.2% had developed cognitive impairment. We conducted a multivariate model (R2 Nagelkerke 0.459) for the probability of developing dementia, with elevated heart rate being the most crucial variable (OR=11.5). The model could excluded dementia with 100% accuracy. Finally, we achieved a discriminant function capable of correctly classifying 95.6% of the cases. It included the following variables of influence: pH, Lawton Brody index, calcium, urea, and heart rate. A few clinical and analytical variables that are easily detectable in the emergency room, especially tachycardia, could help us better identify those patients with delirium at higher risk of developing dementia, as well as formulate hypotheses about the variables involved in the development of dementia in patients with delirium.