Abstract

In this study the contribution of meteorological parameters to the total variability of respiratory infections (RI) is analysed. For this purpose, data on the daily numbers of general practitioner (GP) consultations for RI during the year 2002 were used. This dataset has been compiled by the Local Health Service in the surroundings of Athens, Greece (Acharnes city). The meteorological data obtained by the Meteorological Station of the National Observatory of Athens comprise daily values of mean, maximum, and minimum air temperature, air temperature range, relative humidity, absolute humidity, sunshine, surface atmospheric pressure, wind speed, as well as day-to-day changes of these parameters. Furthermore, the following biometeorological parameters and thermal indices were also evaluated: mean radiant temperature (T (mrt)), predicted mean vote (PMV), physiologically equivalent temperature (PET) and standard effective temperature (SET*) as well as their day-to-day changes. First, the relationship between every meteorological-biometeorological parameter and consultations for RI was examined by applying the Pearson Chi-Square Test (chi (2)) to the data of the 25 compiled contingency tables. In the second stage, the application of generalised linear models (GLM) with Poisson distribution to the data revealed how much the weather variability leads to statistically important changes in consultations for RI. The results of this study contribute to the evidence that there is an association between weather conditions and the number of GP consultations for RI. More specifically, the influence of air temperature and absolute humidity on consultations on the same day is weaker than the lag effect ( approximately 2 weeks) related to cold existence and absolute humidity, while a strong wind during the preceding 3 days drives a peak in GP consultations.