Abstract
Background/Aims:
The objective of our research was to study the weather effect on death events in patients with chronic heart failure (CHF) controlling for the natural cycles.
Methods:
The study dataset contained everyday counts of death events in CHF patients from November 2001 to June 2005, accompanied by daily weather measurements (temperature and its range, barometric pressure, humidity, and wind speed) in 4 Israeli districts (Southern, Haifa, Central, and Tel-Aviv). The nationally representative cohort was comprised of 14,020 patients hospitalized at least once with a primary discharge diagnosis consistent with decompensated heart failure, 8131 of whom died during the study follow-up. Annual, seasonal, and other cycles were used for decomposing weather variables into sums of trigonometric functions. Either Poisson or negative binomial regression model based on special time series technique was used for analysis. To account for possible delay, the model terms also included 1- to-3-day lagged weather variables.
Results:
At large, effect of weather on CHF mortality was found in each district. With that, effect of humidity (the actual and lagged values including interaction terms with the trigonometric functions) on CHF mortality was widespread; but the effect of other variables was different. That is, temperature influence (including the lagged values and the interaction terms) was significant everywhere, but absent in Haifa (the largest seaport). Influence of temperature range (including again the lags and the interactions) was in universally but out in Tel-Aviv (a coastal metropolitan). Barometric pressure and wind speed values (with lags and interactions) had an effect in each region but Central district (an inland part of the country).
Conclusion:
Obtained weather effects are reasonable and match the climate characteristics of the considered regions of Israel. Understanding the impact of weather may better help guiding therapy of patients with CHF.
The objective of our research was to study the weather effect on death events in patients with chronic heart failure (CHF) controlling for the natural cycles.
Methods:
The study dataset contained everyday counts of death events in CHF patients from November 2001 to June 2005, accompanied by daily weather measurements (temperature and its range, barometric pressure, humidity, and wind speed) in 4 Israeli districts (Southern, Haifa, Central, and Tel-Aviv). The nationally representative cohort was comprised of 14,020 patients hospitalized at least once with a primary discharge diagnosis consistent with decompensated heart failure, 8131 of whom died during the study follow-up. Annual, seasonal, and other cycles were used for decomposing weather variables into sums of trigonometric functions. Either Poisson or negative binomial regression model based on special time series technique was used for analysis. To account for possible delay, the model terms also included 1- to-3-day lagged weather variables.
Results:
At large, effect of weather on CHF mortality was found in each district. With that, effect of humidity (the actual and lagged values including interaction terms with the trigonometric functions) on CHF mortality was widespread; but the effect of other variables was different. That is, temperature influence (including the lagged values and the interaction terms) was significant everywhere, but absent in Haifa (the largest seaport). Influence of temperature range (including again the lags and the interactions) was in universally but out in Tel-Aviv (a coastal metropolitan). Barometric pressure and wind speed values (with lags and interactions) had an effect in each region but Central district (an inland part of the country).
Conclusion:
Obtained weather effects are reasonable and match the climate characteristics of the considered regions of Israel. Understanding the impact of weather may better help guiding therapy of patients with CHF.
| Original language | English |
|---|---|
| Pages (from-to) | S164-S165 |
| Journal | Epidemiology |
| Volume | 22 |
| Issue number | 1 |
| DOIs | |
| State | Published - Jan 2011 |