Transcript Diapositiva 1 - Universidad de Talca

ATLAS OF CHILEAN CARDIOVASCULAR DISEASE
MORTALITY 1997-2003
M. Gloria Icaza1a, M. Loreto Núñez2a, Nora Díaz3, David Varela4
1: Instituto de Matemática y Física, 2: Departamento de Salud Pública, Facultad de Ciencias de la Salud, 3: INTA, Universidad de Chile, 4: Geographer
a: Programa de Investigación de Factores de Riesgo de Enfermedades Cardiovasculares (PIFRECV)
Introduction: Mortality atlas provided an image the geographic distribution of disease.
Their underlying objectives range from descriptive epidemiology to hypothesis generation.
On the other hand, cardiovascular disease is the leading cause of death in the world,
being responsible for 30% of all deaths. Moreover, isquaemic heart diseases and
cerebrovascular diseases are the two leading causes of death in Chile.
The relative frecuency of the cardiovascular mortality has increased over time due to
aging and health behaviors changes. The Chilean population has experience
demographic transition which has lead to epidemiological transition where chronic
diseases prevailed over infectious diseases. The relative frecuency of isquaemic disease
is 9.7% from 1997 to 2003 in Chile. Cerebrovascular disease is the second cause of
death with a relative frecuency of 9.3% in the same period.
The aim of this study is to make an atlas of mortality for isquaemic heart diseases and
cerebrovascular diseases using small geographic areas (comunas) in Chile for the period
1997-2003.
Methods: Mortality statistics from the last seven years published by Ministry of Health
and population from 2002 census were used to calculate mortality rates for the 339
comunas in Chile (islands excluded) for people from 30 years old.
Smoothed mortality rates were estimated using Poisson regression mixed models which
take into account the small area population variation. Adjusted rates were estimated
controlling by age, sex and age-sex interaction. The methodology used, assumes
independency between comunas. We were not able to carry out other smoothing
approaches (e.g. Bayesian methods) since we have problems to define neighbourhoods
and to find spatial autocorrelations due to the peculiarities of the Chilean geography.
Population density, from 2002 census, varies from 159 inhabitants in Ollague (north
region) to 231,836 inhabitants in Puente Alto (metrop. region).
Thematic maps are presented in quintiles of mortality rates estimated by the models.
There are counties with large standard errors because they are based on sparse data,
fewer then 20 deaths, and therefore should be interpreted with caution. This areas in the
map are represented with an overlaid hatch pattern.
Models: We used generalized linear mixed models (Breslow, Clayton, 1993;
Wolfinger, O’Connell, 1993) with Proc GLIMMIX from SAS 9.1.3
Results: For isquaemic diseases, there is variation between counties.
There are eight counties witn outliers rate values. Some of them from
highly populated counties, Punta Arenas, Valparaíso, and Villa Alemana.
Punta Arenas is an isolated county in the very south of Chile, where there
is a high prevalence of elevated BMI, Diebetes and phisical inactivity.
There is also a hotspot in the urban area of Valparaíso and their
surrounding counties.
High mortality rates of cerebrovascular diseases are concentrated in southcentral of Chile.
Conclusion: This results are relevant for decision making at local and
national level. They will help to allocate resources and design prevention
programs as well as continue investigation to further understand
differences among counties.
References:
Breslow, N. Clayton, D. "Approximate Inference in Generalized Linear Mixed Models".
Journal of the American Statistical Association, 88, 9-25, 1993.
- Wolfinger, R. O’Connell, M. “Generalized Linear Mixed Models: A Pseudo-Likelihood
Approach”, Journal of Statistical Computation and Simulation, 4, 233-243, 1993.
Finantial support: This work has been founded by FONIS program
(SA04I2005) and sponsored by Chilean Ministry of Health.