Comparison of the previous and present associations of daily mortality and temperature

Based on international literature data and national assessment, a nonlinear association can be stated between temperature and mortality in different temperature ranges in the countries with moderate climate. The association is U shaped in Budapest (Paldy, 2005a), mortality increases in both the low- and high- temperature ranges. The optimal or threshold temperature differs in different populations in relation to the adaptation capacity to local weather conditions. The slope of mortality steeply increases in relation to temperature increase. We wanted to investigate whether the association changed in 2007, when a heat health-warning system was in operation; the extreme weather had a different impact as it was observed in the time series of 31 years.

In the previous studies the permanent population of Budapest was studied, the data of 2007 related to the Central Hungarian Region. In order to compare the effect, the analyzed data were restricted to the death cases that occurred in Budapest (in and out of hospital) and the daily death counts were adjusted to the number of the population of the capital which allowed the assessment of differences expressed in percentage. Plotting the curves of nonlinear association in the same graph, the major characteristics can be examined (Fig. 44.6).

31 years est.

180 160 140 120 100 80 60

180 160 140 120 100 80 60

5 10 15 20 25 30 daily mean temperature °C
22 24 26 28 30 32 daily mean temperature '

Fig. 44.6 Association of daily mean temperature and mortality in Budapest between 19702000 and in 2007.

By linearization of the part of the curve belonging to the high-temperature ranges the differences can be studied. The minimum of the curve of the time series of 31 years was around 18°C; it reached its maximum value at 29.5°C of daily mean temperature. The minimum of the curve of 2007 was above 21°C, it showed a fluctuation till 26°C then it increased steeply. In the range over 30°C, never recorded before, the association displayed an increase of mortality around 3.7 times higher (10.2 cases/1°C) than could be expected by extrapolating the results of the 31 years' analysis (2.7 cases/1°C). The 2007 observation was the first one for this temperature range, which allowed stating that excess mortality was significantly higher than it was forecast, in spite of heat alert.

The relationship of daily mortality was assessed by using the data of Central Statistical Office and permanent population of Budapest. These validated mortality data are however accessible with a considerable delay. Based on the Eu-roHeat project real time mortality data were collected in the Central Hungarian Region in 2007. The benefit of real time data collection is that the impact of heat wave can be promptly assessed by using mortality data submitted within 48 hours. On the other hand these are crude mortality data, do not provide information of the causes of death, and the data provision may not be complete. The results gained by the evaluation of real time data need reinforcement by analyzing the validated data of the Central Statistical Office.

The impact of heat waves on mortality is not only heat stroke but heat can affect many other disease outcomes. Excess mortality attributable to heat can be assessed by comparing mortality of hot days to a baseline mortality, e.g., to that of the cool days. There are different methods described in the literature, the baseline mortality can be a moving average of mortality of the year previous to heat wave, or the mean mortality of the same days of the previous year by applying a statistical smoothing function. The results depend on the methods applied (Whitman, 1997). The data of mortality due to heat waves and applied methods published so far in literature are summarized in Table 44.3. There are great differences of excess mortality of different heat waves, even relating to the same event. The events should be compared with caution, while not only the statistical methods are different, but the heat waves as well. The heat wave of the year 2003, for example, affected the countries in Europe to a different extent, temperature and duration were different. Similar temperature situations may cause different effect depending on the duration of the event and the adaptive capacity of the population. Heat waves of the early summer period may have a stronger effect than that of the late summer. The findings of the 2007 heat wave in Hungary are in line with this statement.

Table 44.3 Heat-wave events and attributed mortality in Europe.

Heat-wave event

Attributable mortality (all cause)

Baseline measure

References

1976-London, United Kingdom

9.7% increase England and Wales and 15.4% Greater London

31-day moving average of daily mortality in same year

McMichel A.J. and Kovats R.S. 1998

1981-Portugal (month of July)

1906 excess deaths in Portugal, 406 in Lisbon

Predicted values

Garcia A. et al. 1981

1983-Rome, Italy

35% increase in deaths in July 1983 in 65+ age group

Compared to deaths in same month in previous year

Todisco G. 1987

1987-Athens,

Greece

21/7-31/7

Estimated excess mortality >2000

Time trend regression adjusted

Katsouyanni K. et al. 1988

1991-Portugal 12/7-21/7

997 excess deaths

Predicted values

Nogueira P.J. and Dias C.M. 1999

1995-London, United Kingdom 30/7-3/8

11.2% (768) in England and Wales, 23% (184) Greater London

31-day moving average of daily mortality in previous 2 days

Rooney C. et al. 1998

1994-The Netherlands 19/7-31/7

24.4% increase, 1057 (95% Cl 913, 1201)

31-day moving average of daily mortality in previous 2 years

Huynen M. et al. 2001

2003-Italy 1/8-31/8

9704 (23.7%) throughout the country

Deaths in same period in 2002

Instituto Nazionale di Statistica (ISTAT) 2004

2003-Italy 1/6-15/8

3134 (15%) in all Italian capitals

Deaths in same period in 2002

Conti S. et al. 2005

2003-France 1/8-20/8

14 802 (60%)

Average of deaths for same period in years 2000-2002

Hemon D. and Jougla E. 2003

2003-Portugal 1/8-31/8

1854 (40%)

Deaths in same period in 1997-2001

Bothello J. et al. 2005

2003-Spain 1/8-31/8

3166 (8%)

Deaths in same period 1990-2002

Martinez-Navarro F. et al. 2004

2003-Switzerland 1/6-31/8

975 deaths (6.9%)

Predicted values from Poisson regression model

Grize L. et al. 2005

2003-The Netherlands 1/6 - 23/8

1400 deaths

Number of degrees above 22.3. C multiplicated with the estimated number (25-35) of excess deaths per degree

CBS. 2003

2003-Baden-Wurttemberg, Germany 1/8 - 24/8

1410 deaths

Calculations based on mortality of past 5 years

SBW. 2004

2003-Belgium 15/5-15/9

1297 deaths for age group over 65

Average of deaths for same period in years 1985-2002

Sartor F. 2004

2003-England and

Wales

4/8-13/8

2091 (17%). Mortality in London region: 616 deaths (42% excess)

Average of deaths for same period in years 1998-2002

Johnson H. et al. 2005

2006-England and Wales, first and second heat waves

680 (6%) excess mortality on the second heat wave

No data available

Health Statistics Quarterly. 2006

The frequency of heat waves has been increasing since the early 1990s. Between 1992 and 2000 six heat waves arrived to Hungary, two heat waves affected the country in 1994, 1998, and 2000. Excess mortality was detected in each event, the rate was between 12% (August 1994) and 52% (June 2000). Between 2001 and 2007, 14 heat waves affected the country; excess mortality was between 17 and 32% for the years 2001-2003.

In the year of 2007, Hungary suffered the hottest heat wave ever recorded, affecting the whole country. Daily mean temperature was above 30°C for 5 days. The impact of the heat wave could not be forecast properly, while the mean temperature was always below 30°C during the previous 17 heat waves. The excess mortality was 38% during the 10 days of the July heat wave; a significant increase could be seen from the third day of the heat wave reaching a maximum on the sixth day (95%) when the night minimum temperature was the highest. In the previous, milder and shorter heat waves excess mortality increased already in the first day itself. During the July 2007 heat wave mortality did not increase in the first 3 days, but a delayed effect of heat could be observed like in Paris in 2003 (Pirard, 2005). This delayed effect was not detected in the previous milder and shorter heat waves in Hungary (Paldy, 2005b). During the first heat wave of 2007 no excess mortality could be detected, and the August event caused only a slight increase in mortality (0.9% excess). The third heat wave did not affect the whole country, as it was forecast by the National Meteorological Service, consequently there was no heat alert announced; only a warning was issued. Daily mortality did not decrease after the July heat wave and, no short time mortality displacement was observed. Further research is necessary to study years of life lost due to the impact of heat.

Real time data collected from general practitioners and hospitals allow a rapid timely evaluation of the impact. These data can, however, be compared to that of previous years with caution. However, new knowledge was gained by the evaluation of real-time data. Mortality was much higher in Budapest than in County Pest and significantly increased in those 5 days when the mean temperature was over 30°C. Mortality of urban population was higher in hospitals than at home during the heat waves. The analysis did not allow answering the question whether the patients delivered earlier to hospital died or patients hospitalized during the heat waves lost their lives. Further, the real-time data collection did not allow analyzing the data stratified by sex, age, and cause of death. Some differences may be explained by the assumption that population living in the agglomeration of the capital prefers hospitals in the capital; furthermore, specialized national institutions (oncology, cardiology, pulmonology) are accessible in Budapest. Probably higher rate of rural population needs home care than of big cities.

Several aspects were considered in the extrapolation of excess mortality due to the July heat wave. If the 33% excess mortality of the population of the Central Hungarian Region were compared to the population of the country, excess mortality for the country would be 970 cases. If the national excess mortality were defined from national daily mortality statistics - taking into account the yearly distribution - excess mortality would be 1050 cases. This value can be regarded as a maximum estimated value for the country. The minimum values could be the sum of 278 cases of the region and the value of the extrapolation of the mortality rate of Pest County to the country which equals to 570 cases ignoring the estimated effect of bigger cities. We consider that the rates of big cities and countryside and the accessibility of hospitals can better model the national impact of heat waves. On this basis the national excess mortality could be between 600 and 800 cases. If the temperature-mortality association significantly differs in some parts of the country from that of the Central Hungarian Region, national excess mortality may differ to a greater extent. The estimation should be controlled by analyzing the validated mortality data in the future.

The summer temperature and mortality relationship was quasi-linear in the higher temperature ranges in the time series analysis of the Budapest data of 19702000. Mortality was the lowest at daily mean temperature at 18°C, above 22°C of daily mean temperature the association is steeper. The optimal daily mean temperature was 22°C in the 2007 analysis. The difference between the two results can be explained by the adaptation of the population to increasing summer temperature. In 2007 mortality was lower in the temperature range of 22-29°C than in the period of 1970-2000, the decrease can be explained by the heat health-warning system introduced in 2004 by raising the awareness of the general population and information of the health-care system. However, lower rate of excess mortality should have occurred (instead of 57% increase registered in 2007, only 15% increased) based on the extrapolation of the estimates of 31 years" time series in the range over 30°C.

Based on the results it is necessary to consider what kind of short-, mid-dle-and-long term measures can effectively and efficiently decrease excess mortality, regarding that several climate models forecast the probability of similar or even hotter heat waves taking place more frequently (Bartholy, 2007; Revesz, 2007). Real-time data collection can be recommended for monitoring effect of heat waves. Continuous evaluation of the heathealth watch warning system even allows modifying the measures during a heat wave.

Was this article helpful?

0 0
Getting Started With Solar

Getting Started With Solar

Do we really want the one thing that gives us its resources unconditionally to suffer even more than it is suffering now? Nature, is a part of our being from the earliest human days. We respect Nature and it gives us its bounty, but in the recent past greedy money hungry corporations have made us all so destructive, so wasteful.

Get My Free Ebook


Post a comment