Etude de l impactde la caniculed août

Etude de l impactde la caniculed août 2003 surla population européenne Jean-Marie Robine 1, Siu Lan Cheung 1, Sophie Le Roy 1, Herman Van Oyen 2, Clare Griffiths 3, Jean-Pierre Michel 4 & François R Herrmann 4 1 Inserm, Health and Demography, CRLC, University of Montpellier 1 2 Unit of Epidemiology, Scientific Institute of Public Health, Brussels 3 Office of National Statistics, London 4 Department of Rehabilitation and Geriatrics, Geneva Medical School and University Hospitals Journée Météo et Santé, Paris, 19 février 2008

Depending on data source, methodology and reference period 2003: 35,000 (Larsen) 2004: over 30,000 (UNEP) 2004: between 22,000 and 35,000 (Schär and Jendritzky, 2004) 2005: Over 35,000 (WHO, 2005) 2005: About 45,000 (Kosatsky, 2005); the previously published estimate of 22,080 early August excess deaths should be revised upward by 100% or more if heat events that occurred during June and July 2003 are also taken into account. 2006: Over 30,000 (Haines et al, 2006) 2006: Between 27,000 and 40,000 (Kowats et all) 2006: Over 38,000 in 7 countries (Sardon ) 2007: Between 45,000 to 50,000 (Barbieri, Nizard and Toulemon, in press)

Methods: Daily numbers of deaths at a regional level from January 1 st 1998 onwards were collected in sixteen European countries. Summer mortality was analysed for the reference period 1998-2002 to set thresholds for extreme values. Variations in daily mortality were examined by calculating the delta between the number of daily deaths observed in 2003 and the reference period. Death frequencies were used to compare countries and regions.

Map of involved countries

Acknowlegment Barbara Leitner, Statistik Österreich, Austria; Herman van Oyen and Bianca Cox, Scientific Institute of Public Health, Belgium; Robert Jurak and Maga Francic, Croatian Central Bureau of Statistics, Croatia; Terezie Kretschmerova, Czech Statistical Office, Czech Republic; Jesper Marcussen, National Board of Health, Denmark; Eric Jougla, INSERM CépiDc, France; Stefan Rübenach, Federal Statistical Office, Germany; Silvia Bruzzone, ISTAT, Italy ; Guy Weber, Direction de la sante, Luxembourg ; Carel Harmsen, Central Bureau of Statistics, Netherlands; Artur Satora, Central Statistical Office of Poland, Poland; Eduarda Góis, Instituto Nacional de Estatistica, Portugal ; Raquel Álvarez Esteban, Instituto Nacional de Estadistica, Spain ; Tatjana Rokavec, Institute of Public Health of the Republic of Slovenia, Slovenia; Dominik Ullmann, Office Fédéral de la Statistique, Switzerland; Clare Griffiths, Office for National Statistics, United Kingdom for data collection.

(1) The seasonality of deaths in Europe (reference period 1998-2002)

Death frequency.002.003.004 0.001 Summer period June- September Non summer period October- May 0 100 200 300 400 Days of the year starting in summer

Table: R squared (expressed as a percentage) of weighted, robust polynomial regressions, used to determine the proportion of variance of the daily death frequency explained by day of the year for each set of mortality conditions (summer and non-summer), sixteen European countries from 1998 to 2002 June to September October to May Crude Adjusted Crude Adjusted Year 5.8 -- 0.5 -- Country 4.8 -- 0.4 -- y = a 0 + a 1 x Day 2.4 13.0 0.1 1.0 y = a 0 + a 1 x + a 2 x 2 Day 2.5 13.1 37.3 38.2 y = a 0 + a 1 x + a 2 x 2 + a 3 x 3 Day 3.1 13.7 37.4 38.3 y = a 0 + a 1 x + a 2 x 2 + a 3 x 3 + a 4 x 4 Day 3.4 14.0 42.7 43.6

Death frequency.002.003.004 0.001 Summer period June- September Non summer period October- May 0 100 200 300 400 Days of the year starting in summer

1 30 61 92 122 Days of the year starting in summer Death frequency 0.001Death frequency.002 divided.003 by median.004 0.001.002.003.004 0 100 200 300 400 Days of the year starting in summer

Distribution of the daily death frequency during summer, from June 1 st to September 30 th, for each of the sixteen European countries, 1998 to 2002 Death frequency div vided by median 0.001.002.003.004 1 30 61 92 122 Days of the year starting in summer

Summer mortality accumulation, from June 1 st to September 30 th *, for the sixteen European countries as a percentage of annual mortality averaged over five years, 1998 to 2002 35% 30% 25% 20% 15% 10% 5% 0% 152 159 166 173 180 187 194 201 208 215 222 229 236 243 250 257 264 271 The two extremes are the Czech Republic, where 31.6% of annual mortality is cumulated during the summer, and Portugal, with only 29.1%.

(2) The characteristics of summer daily mortality in Europe (reference period 1998-2002)

1000 900 Number of summer days 800 700 600 500 400 300 200 100 0 0.05.1.15.2.25.3.35.4 Death frequency by day [%]

(3) The variations in daily mortality in all sixteen European countries during the summer 2003

Delta between the number of daily deaths recorded in the summer of 2003 and the average number of deaths recorded on the same day during the 1998-2002 reference period for the sixteen European countries studied Excess of mortality in numbe er (03 - mean 98-02) -2000 0 2000 4000 6000 Out of the 122 days, 27 days presented a delta lower than zero inducing a total deficit of 5,045 deaths compared to 95 days with a delta higher than zero creating a total excess mortality of 74,483 deaths. June July August September 152 181 212 243 273 Time in day starting on June 1st 2003

01 NL 02 DK 03 PL 04 CZ Death frequency divided by median 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 05 BE 06 LU 07 DE 08 AT 09 UK 10 FR 11 CH 12 SI 13 PT 15 ES 15 IT 16 HR 152 183 214 244 273 152 183 214 244 273 152 183 214 244 273 152 183 214 244 273 Graphs country Summer days in 2003

Accumulation of excess mortality during the summer of 2003 -sixteen European countries studied (reference period 1998-2002) 80000 70000 60000 50000 40000 30000 20000 10000 0 June July August September 152 166 180 194 208 222 236 250 264

140 140 140 120 120 120 100 100 100 80 80 80 60 60 60 40 France 20 REF 98-02 0 152 166 180 194 208 222 236 250 264 140 40 Germany 20 REF 98-02 0 152 166 180 194 208 222 236 250 264 140 40 Portugal 20 REF 98-02 0 152 166 180 194 208 222 236 250 264 140 120 120 120 100 100 100 80 80 80 60 60 60 40 Italy 20 REF 98-02 0 152 166 180 194 208 222 236 250 264 40 Spain 20 REF 98-02 0 152 166 180 194 208 222 236 250 264 40 England & Wales 20 REF 98-02 0 152 166 180 194 208 222 236 250 264

(4) The daily excess mortality in 2003 for each country (reference period 1998-2002)

Table 1: Delta between the number of deaths recorded in 2003 and the average number of deaths recorded during the 1998-2002 reference period Before Summer After Total of the summer June July August September Total summer year Nb Ratio Nb Ratio Nb Ratio Nb Ratio Nb Ratio Nb Ratio NB Ratio Nb Ratio Countries involved in the August 2003 excess of mortality Belgium -4-0.01 139 1.72 162 1.97 438 5.31 436 5.57 1175 3.62 1356 5.11 2528 2.41 Switzerland 92 0.34 253 5.30 187 3.89 469 9.81 130 2.75 1039 5.45-148 -0.93 984 1.58 Germany 9290 2.55 642 0.98 1159 1.73 7295 10.97 259 0.40 9355 3.56-5760 -2.69 12885 1.53 Spain -1464-0.90 4268 15.49 2751 9.64 6461 22.86 1611 6.21 15090 13.68 7249 7.95 20875 5.74 France -3977-1.70 1482 3.60 1706 4.06 15251 36.93 1051 2.62 19490 11.84 3415 2.53 18928 3.55 Croatia 882 3.95 193 4.85 157 3.98 269 6.83 169 4.49 788 5.04 5 0.04 1675 3.29 Italy 5575 2.24 5274 12.12 4318 9.72 9713 21.81 783 1.94 20089 11.63-2487 -1.76 23177 4.12 Luxemburg 69 3.47 33 10.81 27 9.29 75 25.00 34 12.22 170 14.34 79 7.85 318 7.95 Netherlands 304 0.50 78 0.71 11 0.10 578 5.24 297 2.79 965 2.20 503 1.42 1771 1.26 Portugal -2068-4.26 220 2.83 100 1.28 2196 27.75 179 2.44 2696 8.73 2072 7.76 2699 2.54 Slovenia 351 4.30 13 0.87 62 4.21 144 9.93 70 4.86 289 4.96 74 1.55 714 3.81 England & Wales -5695-2.41-1080 -2.64-504 -1.21 1987 4.90-103 -0.26 301 0.18 2025 1.44-3369 -0.62 Total 3355 0.23 11516 4.50 10137 3.88 44878 17.34 4917 1.99 71449 6.99 8382 0.99 83186 2.50 Countries used as controls Austria 708 2.12-42 -0.71 172 2.86 159 2.63 57 0.99 345 1.45-645 -3.30 408 0.53 Czech Republic 2408 5.17 207 2.43 190 2.18 58 0.67-37 -0.43 418 1.22-335 -1.20 2491 2.29 Poland 1916 1.21-487 -1.71-543 -1.85-918 -3.21-652 -2.29-2600 -2.26-3436 -3.60-4119 -1.12 Denmark -113-0.44-43 -0.95-92 -1.95-49 -1.04 14 0.31-170 -0.92 92 0.61-191 -0.32 Total 4920 1.86-365 -0.77-273 -0.56-750 -1.56-618 -1.31-2006 -1.05-4325 -2.74-1411 -0.23

(5) The excess mortality peak at the beginning of the month of August 2003

Median centered annual death frequency Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05 Major Lower outliers Minor Lower outliers Usual Minor Upper outliers Major Upper outliers

June the 1 st 2003 = DOY 152 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

June the13 th 2003 = DOY 164 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

July the16 th 2003 = DOY 197 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

July the31 st 2003 = DOY 212 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

August the 12 th 2003 = DOY 224 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

August the26 th 2003 = DOY 238 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

September the30 th 2003 = DOY 273 Legend 0.00-0.12 0.12-0.18 0.18-0.27 0.27-0.41 0.41-0.62 0.62-0.78 0.78-0.85 0.85-0.91 0.91-0.97 0.97-1.04 1.04-1.10 1.10-1.16 1.16-1.23 1.23-1.27 1.27-1.31 1.31-1.35 1.35-1.39 1.39-1.59 1.59-1.88 1.88-2.32 2.32-2.99 2.99-3.98 3.98-5.47 5.47-7.70 7.70-11.05

(6) The distortions in the proportion of deaths of older people and in the share of women

Distortion of the death structure by age and gender in France and Italy during the summer of 2003 1.6 1.4 France Ratio 65+ Ratio 75+ Ratio 85+ Ratio 95+ 1.6 1.4 Italy Ratio 65+ Ratio 75+ Ratio 85+ Ratio 95+ Age ratios 1.2 1 0.8 0.6 Age ratios 1.2 1 0.8 0.6 0.4 152 166 180 194 208 222 236 250 264 0.4 152 166 180 194 208 222 236 250 264 Time in day starting on June 1st 2003 Time in day starting on June 1st 2003 1.4 R atio female 1.4 R atio female 1.2 Ratio male 1.2 Ratio male Gender ratios 1 0.8 0.6 Gender ratios 1 0.8 0.6 0.4 152 166 180 194 208 222 236 250 264 0.4 152 166 180 194 208 222 236 250 264 Time in day starting on June 1st 2003 Time in day starting on June 1st 2003

Accumulation of excess mortality during the summer of 2003 by gender -sixteen European countries studied (reference period 1998-2002) 50000 45000 40000 35000 30000 25000 20000 15000 10000 5000 0 Women Men 152 166 180 194 208 222 236 250 264

Interpretation: During the summer of 2003, a series of minor mortality crises, most of which remain unnoticed, resulted in a significant number of victims in Europe, especially women, in addition to the huge number of deaths associated with the August heat wave. Global warming constitutes a new health threat in an aged Europe which may be difficult to detect at country level. Centralising and monitoring the count of daily deaths on an operational geographical scale constitutes a priority for Public Health in Europe.

MERCI!