Relationship between air temperature and mortality in two Portuguese cities

Abstract

Abstract Background Current global climate changes are significant and the concern with temperature-related mortality have been investigated, especially related to extreme heat. In the current Portuguese scenario of health decentralization to municipalities, knowing the local climate profile associated to other risks allows public policies to be implemented to protect population. Methods The present study is observational and cross-sectional in two Portuguese cities, Odivelas and Torres Vedras. The daily data on average temperature and the number of deaths were analyzed between 2015 to 2018. The data were provided by statistics Portugal and Portuguese Institute of Sea and Atmosphere. For data organization and preliminary analysis, Microsoft Power BI and Phyton software were used. Results A scatter plot of the average day-temperature versus the number of deaths showed that the variables did not present a clear functional form, but there is symmetry on the graph. Preliminary analyzes were performed between these variables which resulted in a Pearsońs coefficient of -0.17 (p-value = 0.000, confidential interval = 95%) for Torres Vedras and -0.21 (p-value = 0.000, confidential interval = 95%) for Odivelas. The negative coefficient means the lower the temperature, greater the number of deaths. However, these results show that the coefficient is weak and there is no correlation with each other. Conclusions The result shows that the average day-temperature and the number of deaths is not correlated and it will be necessary to include other variables in the study. Other data will be assessed, including period and related factors (season and seasonality, for example). The statistical evaluation for the maximum and minimum temperature data for the day will be applied. Gender and age will be available too. In addition, mortality, morbidity, medication consumption and other environmental variables will also be analyzed. Key messages We believe that the results of this study will provide data for the development of public health policies. The decentralization of health to cities is a challenge that needs data support.