Predictors of mortality in atypical antipsychotic-treated community-dwelling elderly patients with behavioural and psychological symptoms of dementia: a prospective population-based cohort study from Italy.

PURPOSE: To evaluate the effect of a broad range of covariates on the survival of a real-life long-term follow-up cohort of community-dwelling patients with behavioural and psychological symptoms of dementia who were new users of atypical antipsychotic medications (APMs). METHODS: This was a prospective cohort study of 1,618 subjects aged ≥65 years with dementia and BPSD ("behavioural and psychological symptoms of dementia") who were new users of atypical APMs and registered in a Dementia Evaluation Unit of Campania Region (Italy) from September 2006 to March 2010. The potential of baseline features to predict mortality was assessed with the Cox proportional hazards model. RESULTS: The average follow-up was 309 days. Of the 1,618 new users of atypical antipsychotics, 9.3 % experienced at least one adverse event, including death (5.1 %), drug therapeutic failure (3.0 %), extrapyramidal symptoms (0.5 %) and stroke (0.2 %). The crude all-cause mortality rate was 6.0 per 100 person-years [95 % confidence interval (CI) 4.8-7.4]; the rate was higher in patients aged >85 years (9.0 per 100 person-years, 95 % CI 6.4-12.7) and among male patients (7.5 per 100 person-years, 95% CI 5.3-10.6). In the multivariate analysis, only age was associated to all-cause mortality [hazard ratio (HR) 1.1; 95 % CI 1.0-1.1 and HR 1.4; 95 % CI: 0.9-2.2, respectively). In contrast, hallucination (HR 0.4; 95 % CI 0.2-0.6) and dosage change (HR 0.4; 95 % CI 0.2-0.78) were significantly associated with a lower risk of all-cause mortality. CONCLUSIONS: Among our patient cohort, the mortality rate of patients with BPSD receiving long-term treatment with atypical APMs was lower than that reported in other studies, and only age was found to be significant predictor factor of mortality.

Dimension Reduction Techniques for Distributional Symbolic Data.

In the framework of symbolic data analysis (SDA), distribution-valued data are defined as multivalued data, where each unit is described by a distribution (e.g., a histogram, a density, or a quantile function) of a quantitative variable. SDA provides different methods for analyzing multivalued data. Among them, the most relevant techniques proposed for a dimensional reduction of multivalued quantitative variables is principal component analysis (PCA). This paper gives a contribution in this context of analysis. Starting from new association measures for distributional variables based on a peculiar metric for distributions, the squared Wasserstein distance, a PCA approach is proposed for distribution-valued data, represented by quantile-variables. An application of the proposed PCA method, performed on simulated distribution-valued data, shows interesting interpretative results in terms of location, variability, and shape of the distributions on the factorial planes.