Introduction Recently, analysis of health effects of climate change has become a global issue. This new and complicated problem calls for a new approach, since existing models cannot analyse changing effect of climate on health outcomes over a long period of time. We suggest methodology of analysis and estimation of changing effect of climate on health outcomes over time and apply it to estimating the effect of meteorological changes and hospitalisation and hospital mortality rates through a long time-period.
Methods Data on daily hospitalisation and hospital mortality (all-cause and specific) rates for 16 years were obtained from a database of the medical center providing tertiary care to the most population of Southern Israel. Regional daily average temperature and humidity were used as meteorological variables. Three different methodologies have been developed:
1. Constructing and comparing trend functions adjusted to possible confounders of health and meteorological variables, using harmonic analysis based on generalised linear models.
2. Estimating the effect of meteorological variables on health, changing over time by using numeric derivations.
3. Constructing special spline models for analysis of significant changes in time, based on conventional meteorological health effect models.
Results The health and meteorological variables were characterised by increasing and non-linear trends. The trends of hospitalisation and age-adjusted hospital mortality rates are strongly and significantly associated with the trends of temperature and humidity.
Conclusion Increased health outcomes are significantly associated with increased heat and humidity in our region. Similar findings were obtained by three proposed methodologies.