<Headline>Impact of Heat Waves on Mortality — Rome, June 1–August
15, 2003
Mortality
and Morbidity Weekly Report
P Michelozzi,
MS, F de’ Donato, MS, G Accetta, PhD, F Forastiere, PhD,
M D’Ovidio,
and C Perucci, MD
Dept. of Epidemiology, Rome Local Health Authority.
L Kalkstein,
PhD.
Center for Climatic Research, Univ. of Delaware.
<Body Text> During
the summer of 2003, record high temperatures were observed across Europe,
and Italy was one of the European countries most
affected by these
this e high
temperatures. The higher frequency and intensity of heat waves might
lead to an increase in heat-related deaths in the future with the greatest
impact on urban populations, particularly the elderly and the ill. A Heat
Health Watch Warning System (HHWWS) has been activated in Rome insince 2000
to alert
health and social services of potentially oppressive weather conditions that
negatively affect health (1).
To assess the health impact of the
heat wave in Rome, the Department of Epidemiology of the Rome Health Authority
evaluated the relationship between temperature and daily mortality
for the period June 1–August 15, 2003, and compared the observed excess
mortality observed with the mortality estimated predicted by
the HHWWS. The results of the analysis indicate record excess mortality during
heat wave periods, and a strong association between daily mortality and temperature.
The HHWWS was able to identify forecast observed peaks in mortality up
to 3 days in advance. These findings emphasize the need to expand the development
of warning systems and prevention programs aimed at susceptible populations
in order to reduce the impact of heat on mortality.
In Rome significant
increases
in daily mortality is are frequently
observed during the summer (2), and in 1999 Rome was included in a World Meteorological
Organization (WMO) project on cities at high risk of heat/health impact.
The city
has an estimated population
of 2.700.000, of which 486.000 (18%) are over 65 years of age (2001, National
Census) (3). The mean annual number of total deaths is 26.000
(4). Daily deaths in Rome from June 1–August 15, 2003 were obtained
from the Mortality Registry of Rome (4). Deaths due to accidents (International Classification of
Deaths, ICD-9: 800-999) and those occurring outside Rome were excluded from
the analysis. Daily excess mortality was defined as the difference between
the number of deaths observed on a given day and the smoothed average daily
value for the reference period (1995–2002). Confidence limits were determined
on the assumingption of a Poisson distribution. To evaluate the association between excess mortality and
socio-economic status, a deprivation index based on education, occupation,
unemployment, number of family members, overcrowding and household ownership
data for the census tract of residence was applied (5).
HHWW systems
are based on a synoptic climatological procedure which defines
air masses and identifieidentifiess “oppressive” conditions historically
associated with an increase in mortality. (6,7). Meteorological
forecast data is applied to the system to predict oppressive air masses and
the related excess in mortality. The performance of the HHWWS of Rome during
summer was evaluated comparing alarm days with excess in mortality predicted
by the system to heat wave days defined as those with an observed maximum
apparent temperature (Tappmax[1])
(8) above the 90th annual percentile and with a difference
of 2°C between the first heat wave day and the previous day.
During the studymmer period considered, the
mean daily temperature was 3° C above the mean and Tappmax was
35.2°C compared to 31.1°C for the reference period. Tappmax
was higher than 33.1°C (90th annual percentile) on 55 days (72%
vs 35% in the reference period ). Three major heat wave periods were observed
in Rome; the first episode (June 9–July 2) registered a mean Tappmax
of 36.1°C, with peaks of 37.9°C and 40.3°C. The second (July 10-30) had a
mean Tappmax of 36.4°C and registered two peaks above 40°C; while
the third was shorter (August 3-13) but registered a higher mean (38.2°C)
with three days above 40°C.
Overall daily mortality trends
followed temperature trends, with peaks in mortality coinciding with peaks
in maximum apparent temperature or with a lag of one day (figure1). The total mortality was
17.5% higher than expected with 708 excess deaths. The first heat wave was associated with an increase in mortality
of 343 total excess deaths; 285 during the second heat wave (July 10–30)
and only 84 excess deaths in the third. Excess in mortality was observed only
among people over 65 years old and increased dramatically with age: 24 more
deaths occurred among people aged 65–74 years (+2.5%), 254 in those aged 75–84
years (+20.4%), and 461 (+40.2%) aged >85 years. The increase in mortality was greater among females with 499 deaths compared
to 209 male deaths. The higher mortality of women can be attributed to the
higher proportion of females in the elderly population. Excess mortality
was higher for out-of-hospital (+22.6%) than in-hospital deaths (+6.4%).
Furthermore, a strong association was observed with socio-economic status:
excess mortality was 5.9% in the highest level, 2.7% in the medium-high,
12.7% in the medium-low and 17.8% in the lowest level (figure 2).
In the summer of 2003 the HHWWS called an alarm on 21 days
(28%) and an emergency on 18 days (24%). The system was able to identify predict the major peaks in observed mortality observed, but
underestimated the total number of excess deaths (467 excess deaths vs. 738
observed for the 65+ age group). The underestimation by the system was probably
due to the
unusually extreme weather conditions that persisted throughout the summer
2003 which resulted in oppressive air masses associated with higher temperatures
that had greater impact on heat-related mortality. These weather
conditions are less represented in the retrospective data utilized to builtd the predictive
model.
Reported
by: P Michelozzi, MS, F
de’ Donato, MS, G Accetta, PhD, F Forastiere, PhD, M D’Ovidio, C Perucci,
MD, Dept of Epidemiology, Rome Local Health Authority. L Kalkstein, Prof.,
Univ of Delaware, Center for Climatic Research.
Editorial Note: Although interest
in the impact of heat on mortality is growing, events this summer have made
it clear that the majority of European countries were unprepared to cope
with this emergency. An
aAccurate
estimates
of the heat-related mortality in the
most affected European countries isare still not still available, however tens of
thousands of excess deaths have probably occurred.
The high mortality observed during summer 2003 was a consequence of extremely high temperatures for many consecutive days, the early onset of hot weather, and the prolonged heat stress conditions. In Rome a total of 708 excess deaths were observed, 17.5% higher than the reference period. The highest increase was observed in people over 75 living in the most deprived areas of the city. Socio-economic condition may reflect housing quality, access to air conditioning, individual behaviour, access to social and health services and other social factors that can potentially impact health. However, the effect might also be caused by the higher proportion of elderly people of low socio-economic status who remain in the city during summer.
A harvesting effect,
or displacement of mortality, has been reported after heat wave episodes,
with peaks followed by a reduction in mortality over the subsequent weeks
(9). During summer 2003, persistent high temperatures maintained
mortality at above average levels, but the observed increase in daily mortality
in August’s during the
heat wave in August
was smaller than in the previous 2003 heat wave periods. However, the
available data do not allow for valid estimates of time variation of the
denominator during the study period.
In 2002, the
Department of Epidemiology and the Municipality of Rome collaborated on the
HHWWS and established an intervention plan aimed at the elderly and other
susceptible groups. Public and private institutions and organizations were
identifienrolled
and acted as mediators between the warning system and the population. Specific guidelines were developed for the general population,
and for patients suffering from specific diseases by the Department
of Epidemiology in collaboration with the Association of General
Practitioners. Intervention activities
include a 24-hour a day tele-assistance program, where registered users can
call an operation centre for assistance.
Further studies will be needed to assess weather the HHWWS and the associated preventive programs had been able to reduce the heat related mortality in Rome during summer 2003.
In Italy, the increasing proportion of elderly people and global warming could make the susceptible population more vulnerable and heat-related mortality may increase.
Heat stress
conditions are predictable, and valid and efficient HHWWS combined with appropriateffective
prevention measureprograms
may reduce heat-related mortality.
In Europe there is a need to promote
interest on climate and health. A European Commission-funded project (Assessment
and Prevention of Acute Health Effects of Weather conditions in Europe-PHEWE)
on climate and health involving 16 cities from different countries started
in 2002 is in course. The main objectives will be to evaluate the acute health effects of weather, to implement HHWWS in
a subgroup of pilot cities, to develop preventive strategies to minimise
adverse health effects, and to define guidelines for public health interventions
in Europe.
1. Michelozzi P, Fano V, Kirchmayer U, Becker M, Forastiere
F, Perucci CA. Public Health Programmes for the Prevention of Negative
Health Effects of Combined Climate and Air Pollution Conditions in the City
of Rome, Presentation at XX Reunion Cientifica de la
Sociedad Espanola de Epidemiologia, September 2002.
Albertoni F, Arcŕ M, Borgia P, Perucci CA, Tasco C. Heat-related mortality Latium Region summer 1983, MMWR, 1984;33:518–21.
3. http://dawinci.istat.it/pl/
4.
Michelozzi P, Barca
A, Capon A, Fano V, Kirchmayer U, Perucci CA 2000. Mortality
in Rome and the Lazio Region 1993-98: Cd-rom Atlas.
5. Michelozzi P, Perucci CA, Forastiere F, Fusco D, Ancona C, Dell’Orco V. Inequality in health: socioeconomic differentials in mortality in Rome, 1990-95. Journal Epidemiology and Community Health 1999;11:687–93.
6. Kalkstein LS, Nichols MC, Barthel CD, Greene JS. A New Spatial Synoptic Classification: Application to Air Mass Analysis. International Journal of Climatology, 1996;16:983–1004.
7.
Sheridan SC. The
Re-development of a Weather Type Classification Scheme for North America.
International Journal of Climatology, 2002;22:51–68.
8.
Kalkstein LS,Valimont
KM. An Evaluation of summer discomfort
in the United States using a relative climatological index, Bulletin of the
American Meteorological Society 1986;67:842–8.
9.
Braga AL, Zanobetti A, Schwartz
J, The time course of weather-related deaths, Epidemiology 2001;12:662–7.