Abstract

COUTIN MARIE, Gisele. Prognoses of mortality from non-communicable selected diseases. Rev Cubana Hig Epidemiol [online]. 2008, vol.46, n.3. ISSN 0253-1751.

The improvement of surveillance is one of the main functions of the strategic component of the surveillance model in Cuban health. The obtention of prognoses to reduce the uncertainty at the time of making decisions is crucial in a variable environment where the early detection of new problems is necessary. The use of techniques of analyses of temporary series with prediction ends is generalized in the public health sphere. The isolations or exponential smoothing are the most used methods to obtain prognoses of temporary series due to their simplicity, computational efficiency, and acceptable accuracy. In this paper we showed the results achieved with the application of these methods compared with the adjustments of some polynomials to attain prognoses for mortality from some non-communicable diseases selected between 2007 and 2009. The prediction models for all death causes characterized by little prediction errors and narrow prediction intervals are exposed as the main result. It was concluded that exponential isolations allow to design models in a relatively simple way and are very useful for the surveillance of health events, since they make easy the opportune decision making.

Keywords : Prognosis techniques; time series; exponential isolations; polynomial adjustments; statistical techniques for surveillance.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )