French heatwave disaster of August 2003

When peasants die, there are no comets seen
William Shakespeare, Julius Caesar (act II scene II)

The connections that make society strong also make it vulnerable
Barry Hindes, Threads .

This report considers the failure of the French government to prevent the unnecessary deaths of 14,000 predominantly senior citizens during a heatwave in August 2003. Governments have frequently used risk management techniques to identify threats, and control risks at different levels: prevention, damage control etc. Unfortunately in France the government failed spectacularly from the start, by not confronting politically tough, but conveniently well hidden, issues like labour law, building design and social capital that contributed to the disaster. As the death toll rose their second mistake was to downplay and deny the disaster for the sake of political expediency in a manner that ultimately contributed to even more deaths.

This essay uses the ideas of Sen (the marginalised are more risk-prone) and Shiels (the marginalised are ignored) to show how the problems facing the elderly were exasperated, before applying Power's theory of how the state selectively filters information to put itself in a politically sanctified position of believing it is incapable of responding to a problem faced by a marginalised group it has no interest in actively saving. It also hopes to illustrate a new type of disaster where a vital system people take for granted is overwhelmed by demographic forces.

Introduction
A government's lot is not a happy one. Since feudal times its mandate has been founded on securing the lives of its citizens, since it operates the public goods that provide protection (e.g.: armies, hospitals, meteorological services), and citizens have a Rawlsian belief that the weak (including those stricken by disaster) will be protected by the state. When a government fails to provide this basic need, it looses the trust of the people. Unfortunately for the state, it needs to exercise leadership and judgement by limiting the freedoms of its citizens in order to protect them, which may be unpopular (such as ordering cyclists to wear helmets). The state may be tempted to satisfy stronger political groups at the expense of not protecting weaker or diffused groups. Yet if a crisis does emerge, the state can try to salvage its trust with the public by downplaying its existence, its implications, and whether it really is the government's business to intervene.

The French heatwave disaster of August 2003, where 14,000 mostly elderly victims died, is an excellent case example. Unlike other disasters with physical damage and articulate victims, heatwaves are silent killers that wipe out dispersed population groups. Often only the government through its statistical and medical services are aware of its magnitude. And since the most obvious causes of death can be pinned on the weather or the failure of victims to hydrate themselves, it is easy for governments to deny responsibility. Most damning in France's case was the government's reluctance to acknowledge the crisis, and adopt counter-measures earlier that would have saved lives.

This essay will show that (a) disasters are often rooted in a complicated arrangement of external and invisible factors, (b) interest groups who are responsible for these factors will disavowal responsibility, especially when links appears weak, (c) governments are therefore reluctant to pressure these groups, especially those who are politically stronger than those exposed to risk, and thus (d) risk control measures become more reactive than preventative. Yet when disaster strikes, governments may deliberately not engage whatever reactive response instruments they have left at their disposal, in order to hide the embarrasing fact that something bad occurred on their watch.

Setting the Scene
This report can rely on certain incontrovertible facts about what happened in August 2003 in France. Specifically: (1) in the first two weeks of August 2003 Western Europe, experienced a prolonged weather pattern typified by high maximum and minimum temperatures that could be considered a `heatwave', (2) anecdotal evidence cited in the media, and later confirmed by Mattei, suggested that in this period medical and aged care facilities were understaffed, causing hospital wards to close, services to be withdrawn and/or remaining staff to undertake unmanageable workloads, (3) statistical records suggest a significant number of excess deaths occurred in this period, particularly amongst the elderly from hyperthermic causes, and (4) both the public and various political actors progressively formed the impression that France's medical system had broken down, because of a number of reasons, and consequently (5) certain public servants and politicians were compelled to accept responsibility and resign from office, while those who survived the public backlash were made to reconsider the country's approaches to aged care, labour hours and other policies.

Background
The effects of the heatwave were aggravated by a number of background factors, which prevented the French health authorities adequately responding in sufficient time.

An ageing population
France, like other developed countries, has a growing proportion of elderly citizens. However demographically France more balanced than its neighbours thanks to a large, youthful and fecund immigrant population and generous maternity leave provisions. In 2000 16.1% of its population was aged 65 or older, compared with 12.1% in 1970 (Pison 2000). France also has the highest life expectancy for aged females in Europe; on average a 65 year old woman will live a further 19.8 years, while men can expect an additional 15.4 years (Giarchi 1996:193) and third highest for males (15.4%) in Europe . As the population ages not only does the demand for nursing home, respite and other facilities rises, but the supply of labour and taxpayers to support these facilities drops. Increased education, labour participation and vocational aspirations amongst women, traditionally the main providers of aged care, has also tightened the supply of aged workers. Although their workforce participation has perhaps lead to more tax revenue that as a Pareto dividend has helped pay for pensions and health costs, aged care remains a labour intensive activity.

The 35-hour week
In France much of the population in August 2003 were spending the summer holidays away from their workplace and families. Thanks to the Loi Aubry inaugurated by the previous Jospin administration on 1st January 2000, the working week for French employees employed in firms and organisations with twenty or more employees was cut down to 35 hours a week, calculated from an annual total of 1,600 hours. The move to the Loi Aubry was prompted by self-interested labour unions who prioritised providing greater leisure time for its members over increased remuneration. The government hoped that the law would spurn job creation, since the supply of labour hours that existing employees could provide were reduced, then new employees would be required to meet the labour demands of the existing workload. In 2002 the mandatory application of the 35 hour work week was imposed on all employers in France.

Along with the reduction in working hours came a more flexible framework for employers and employees to negotiate patterns of workplace attendance that would suit cyclical troughs and spikes in demand, and thus raise overall labour productivity. In the 2003 summer holidays, many vacationers spent longer periods away as a result of having accumulated large banks of leave credits.

In the years after the Loi Aubry was introduced, unemployment indeed did fall, from 12.4% in 1997 to around 8.8% in 2001. Commentators cited a boom in leisure industries such as fitness centres where demand was influenced by the amount of free time customers had available to spend. However, strong economic growth was responsible for most employment generation - the Jospin government itself admitted that perhaps the new regulations only contributed to a fifth of new jobs being created, while less partisan economists gave even less credit. By the end of 2003 France's employment rate moved upwards to 9.8% (Australian Department of Foreign Affairs and Trade 2004).

The best laid plans of the Loi Aubry failed as a result of policy practitioners assuming that the number of jobs a labour market provides (and thus can be `shared') is constant. This mistake, known as the `lump of labour' fallacy, forgets that the labour market is both dynamic and heterogeneous. The supply of labour is restricted by many factors unlikely to be adequately addressed by the market in the short-term, such as requisite training and geographical location. And the demand for labour constantly shifts, being sensitive not only to economic factors like consumer demand and interest rates, but also to government policies. As the Loi Aubry was effectively a wage rise since salaries were not adjusted downwards in response to a corresponding drop in hours worked, companies responded by choosing alternatives other than hiring new staff to meet production demand. Where the nature of the work permitted, jobs were either replaced by automation, or relocated offshore (increasingly possible for white collar employment, thanks to telecommunication technology, cheaper transportation and transfer costs, and harmonised industry standards within the European Union). Yet for other occupations where labour supply is inelastic, the shortfall in labour hours was made up by employers by compelling the existing staff to increase their own productivity by working harder (Bartlett 2003). For many of France's medical, nursing and aged care workers, they simply were expected to carry out their existing workload with fewer staff available.

France's health care system
France has a mixed public health system. There are 1,032 public hospitals with a total of 315,687 beds, staffed by "hospital practitioners" according to civil service guidelines. In a somewhat inflexible employment regime, salaries are based on a flat rate regardless of specialisation or location, and seniority determines promotion. Practitioners may engage in private practice subject to earning and time limits. 91% of funding for public hospitals comes from health insurance funds, allocated to each facility according to various economic, productivity and medical criteria. In addition there are 2,139 private establishments (with 170,382 beds), including user-pay clinics, and mostly religious non-profit hospitals which are heavily subsidised by state grants. There is a heavy amount of state supervision and regulation, both regional and national, which many claim is cumbersome. Budgetary cut-backs to public hospitals which started in the conservative Jospin government in 1995 and continued under Jospin in 1998 led to hospital bed numbers being cut, and foreign doctors being recruited at around half the pay of French doctors to deal with staff shortages. It is little wonder that given the inflexible and parsimonious work environment in public hospitals that 80% of French medical graduates in 1998 chose to work in private practice, and also in that year 20% of hospital practitioner positions remained unoccupied (Dorozynski 2000:333).

The situation in nursing homes is similar. According to Hubert Falco, 80% of French nursing homes are understaffed outside holiday periods, based on a benchmark of no more than two residents per employee. Older people who generally have lost the thirst reflex need to be monitored and cajolled into drinking to ensure they are ingesting enough fluids, a labour intensive activity (Knox 2003).

In August 2003, medical and nursing staff in public hospitals were working under the 35 hour provisions, with those on night-shift working a standard 32 hour week (de Kervasdoué 2002). Anecdotally, many public doctors working in essential services like accidents and emergencies stated they work well beyond 35 hours due to administrative requirements they are not allowed to shut down wards, and thus they are forced to carry the workload of absentee staff (Fouché 2002). Their obligation to serve made them politically weak compared with other labour sectors. Those in the private sector however have more flexibility to determine their own periods of engagement, and many elected to spend their holidays in the first two weeks of August. Patients who would otherwise have used these private clinics were therefore reliant on the already overstretched public hospitals, the carers of the last resort.

Air conditioning
Like other European countries, the elderly tend to live in older dwellings, built before the introduction of air conditioning. 53% live in dwellings built before 1962 (21% before 1871) (Dooge 1992:173). In Europe, France has the second largest proportion of housing built prior to 1919 (21.0%) (Sak and Raponi 2002). At the start of 2003 only 6% of French houses had air-conditioning (Agence France-Presse 2004). In 2000 France had approximately 2.5 m2 of air-conditioned indoor space per inhabitant, slightly under the European Union average; by comParison in Italy, Greece and Spain there is around 6.0 m2 per person. And only about 2% of all centrally air-conditioned space in France existed in hospitals, compared with a European average of 7% (Centre d'Energétique 2003).

Social capital
social capital is the sum of all informal, altruistic social interactions occuring between people in a community. The cornerstone of social capital is the trust neighbours have with each other.

The best known proponent of social capital, Robert Putnam, recently concluded that ethnically homogeneous neighbourhoods had higher levels of social capital than more diverse communities (The Economist 2004:53), since people are more likely to be altruistic towards their `own kind'. Klinenberg noticed in Chicago that while economically disadvantaged areas in Chicago had more deaths in the 1996 heatwave, some poor ethnically homogeneous neighbourhoods which showed evidence of social capital (in the form of extended family and friendship networks amongst its residents, which led to busier streets and lower crime) suffered fewer fatalities, even amongst its own minorities (2002:87). South Lawndale (85% Latino, 22% poverty level) had a heat death rate of four fatalities per 100,000, compared with a Chicago average (56% white, 18% poverty level) of seven fatalities per 100,000. Neighbouring North Lawndale (96% black, 44% poverty level) had a horrendous toll of forty deaths per 100,000; its homogeneity afforded no succour for social capital against a backdrop of the more potent forces of crime and illicit drugs.

In France, the heatwave death rate was particularly high in Ile-de-France, with 230% excess mortality (Hemon and Jougla 2003:32). This conurbation includes not just Paris but a number of less prosperous surrounding dormitory suburbs (banlieux) collectively known as the petite couronne, with a number of indicators consistent with a fractured community. For reasons that vary from unemployment in a post-industrial world, non-participation from mainstream economic and cultural life, welfare dependency and ethnocentrism, some of its departments have the highest crime rates in France (Keijzer 2004). Many communities in the petite couronne were built as large public housing estates in the last thirty years. The Ile-de-France also has the greatest concentration of foreign residents (12.9% compared with 6.3% for France overall (Observoire Régional de la Santé 1998) and these figures ignore visibly "foreign" French nationals). Also 35% of all France's Muslims (EuroIslam 2002) reside here, a youthful community group that some think has rejected the kind of Gaullic secularist communitarian values needed for social capital to flourish in a multiethnic, pluralist country. If the electoral successes of the Front National in the banlieu electorates are any indication, there seems to be a significant degree of social disconnection on ethnic and generational lines. Paris's North Lawnsdale would be either Seine-Saint Dennis (26.2% of the population being foreign citizens according to the L'Institut National de la statistique et des Etudes Economiques) or Val-de-Marne where the highest excess mortality rates from the heatwave were recorded: 260% and 270% respectively (L'Institut National de la statistique et des Etudes Economiques, 2004).

The social problems in the banlieux can be traced to high youth unemployment, caused by labour market inflexibility - a politically sensitive issue the French government doesn't want to touch.

The heatwave Strikes
The deaths generally involved elderly, low-profile victims who often lived alone. This meant that unlike other disasters, there was a considerable time lag between when the death rate began to accelerate and when the public and the government became aware of its true scale, and could appropriately react. Quite often the victims were found days after they had died. The deaths tended to occur in isolation from each other, and could have been interpreted as being due to natural causes, aggravated by climatic conditions and the victims failing to hydrate themselves. Until it became a topical news item the true magnitude and effects of the heatwave was only felt by that proportion of the population involved in France's medical, aged and mortuary services.

Timeline
Neustadt and May, who believed that public management can be best understood by using historiographical methods, suggest that narratives are the best way to understand how problems occurred, as it best illustrates cause-and-effect phenomena. (1986:120). Temperature figures and mortality rates below refer to observations in the Ile-de-France region.

Late July. West Africa is hit by unseasonably heavy rain in July 2003, which pushes northwards an anti-cyclone system. The anticyclone makes its way over North Africa, warming itself from the hot air of the Sahara desert, before loosing momentum above the western European landmass, where it firmly anchors itself. As it holds back rain-bearing depressions from the Atlantic Ocean that cool Europe, the anticyclone funnels in more hot air from the Sahara. Throughout Europe temperatures daily maximum and minimum temperatures rose beyond their average (Stumme 2003).
In France from the 4th August, maximum temperatures of 35.0°C or greater are experienced at around two thirds of France's 180 weather stations, with 15% having temperatures exceeding 40.0°C. Importantly there are consecutive days of high minimum temperatures which do not give adequate respite, especially to those living without air conditioning and instead depend on trapping cool air currents overnight for cooling. Economic losses included forest fires, disrupted riverine transport and crop losses - in Germany, estimates of 80% of crops being lost to heat were circulated by farming associations, with losses to the grain harvest alone costing around one billion Euros (Deutsche Welle 2003). The heat not just influenced consumer behaviour (beer and air-conditioner sales enjoyed a boom), but also the actual and potential short term labour participation, as workers elected to spend their summer vacation away from their stifling cities, or extend their absence from work.

1 - 5 August. From the 1st to 5th August, mean daily maximum temperatures progressively rise from 25°C (normal for summer) to 37.0°C. Then right through to the 14th, minimum temperatures will not go below 20°C. The mortality rate does not move upwards much beyond the August average. A text analysis of articles published in the French newspaper Le Monde during the period show that excessive heat was first only reported as a meteorological phenomenon, a threat to agricultural production, or as a factor aggravating other hazards, such as pollution, the forest fires and the inability for nuclear reactors to function at full capacity due to cooling issues.

6 - 7 August. Maximum temperatures hit 39°C, then slightly dipped to 38°C. The mortality rate jumps to double then nearly trebles (i.e.: 300%) the August mean. The first media report in Le Monde of a human fatality due to the heatwave came on the 7th August, involving a single 32 year old man (Le Monde 2003a) Even as mortality peaked in the following days the few Le Monde articles containing the word canicule (heatwave) mention nothing about heat deaths. Neither did the Ministry of Health make any comment.

8 - 10 August. The maximum temperature starts rising upwards to 39°C again, while the mortality rate plateaus at around 300%. The Ministry of Health becomes aware of the growing number of hospital admissions around Friday 8th August, when it identified a marked rise in admissions in the previous 48 hours (2003a). The Ministry tasks the National Institute of Medicine to specifically monitor for heat-induced deaths in conjunction with regional medical authorities. A public health warning is also issued on the 8th, with the advice seemingly targetted more towards active people or people responsible for infants. No reference was made about any fatalities that had occurred (2003b).

The 9th and 10th August fell on a weekend, and the Health Ministry did not release any communiqués. The first signs of public disquiet came from the president of France's association of emergency physicians, Patrick Pelloux, who with first-hand knowledge of rising hospital admissions criticised the government's inaction (Bosch 2003:624). Pelloux denies the deaths are `natural', as the government stated in a communiqué, and anticipating worse to come, he mentions France faces a `veritable massacre' (BreakingNews.ie 2003).

Monday 11 August. The maximum temperature stays at 39°C, but the mortality rate skyrockets to 450%. The National Institute of Medicine detects an escalation of admissions in public hospitals in France, mostly in the Ile-de-France. Some medical personnel and resources are mobilised to reopen wards closed due to staff absences, and elective surgery is postponed. In a press release the Health Ministry acknowledges there has been a rise in heat-related deaths since 6th August, in particular amongst `vulnerable' people. The press release furthermore mentioned that measures are in place to monitor the situation and assist hospitals, and another public advisory is published (2003b).

Health Minister Dr Jean-François Mattei gave a news briefing from his holiday villa that evening, where he casually downplayed the threat (Erickson 2003).

Tuesday 12th August. Paris suffers its highest maximum temperature, with 40°C, and the mortality rate peaks at 620%. A hotline telephone service is established offering preventative advice, and receives 12,000 inquiries on the first day. Mattei orders the mobilisation of the Red Cross and some military hospitals, while the Ministry of Health in another press release acknowledged a shortage of space in mortuaries in Paris region. Two articles in Le Monde make reference to claims by a group of doctors that fifty deaths due to excessive heat had occured, and the crisis is worsening in the public hospitals. Raffarin from his Alpine retreat dismissed complaints from his political opposition about the rising death toll as `inappropriate', and said that for weeks the health service had been preparing for this busy period (Gentleman 2003).

Wednesday 13th August. The maximum temperature drops suddenly to become only 30°C, but the death rate is a dramatic 550%. A number of articles aree written about deaths in hospitals with a more alarmist tenor, and included reports of the first government reaction - that it was monitoring the situation and would convene a meeting within two days. Mattei, still on holiday, was verbally abused as he inspected a nearby hospital where 20 patients had died, and the Greens Party calls for his resignation (Les Vertes 2003) . He comments for the first time that hospitals around the country were overfilling with casualties. Finally in the evening Raffarin activates Plan Blanc(`White Plan'), an emergency set of protocols meant for major disasters. Plan Blancempowers hospital authorities to recall medical staff from leave back to duty and coordinate with government authorities the use of additional facilities and personnel, ranging from opening military hospitals to the public to using refrigerated trucks as temporary morgues (United Press International 2003). Many of these measures had already been in place, albeit piecemeal and poorly coordinated between different levels of government, suggesting that declaring Plan Blancwas more symbolic than practical.

Thursday 14th August. Both Mattei and Raffarin return to Paris. Plan Blancgoes into effect six days after the Health Ministry knew the hospitals weren't coping, but on this day the maximum temperature would only reach 27°C. The death rate too is dropping, down to 200%. At a press conference Mattei suggests a preliminary figure of 3,000 excess deaths (Nau 2003), describing the magnitude of the heatwave to be of `epidemic proportions'.

Friday 15th August, and beyond. By now the weather and death rate are normalising. The only thing useful left for the government to do is to deal with a backlog of bodies needing storage and disposal. Ironically the 15th falls on a bank holiday, so burials need special dispensations from mayors.

The outcry from the public and the political opposition builds up as the estimated death toll rises with each new revision. Initial estimates produced by the Health Ministry of 1,500 to 3,000 are eclipsed by a professional organisation of undertakers estimate of around 10,000. President Chirac returns to Paris on the 21st after a three-week holiday in Canada, where he had not released any media statements concerning the heatwave. He convenes a Cabinet meeting to assess what had been the government's response during the disaster, and he publically pledges that deficiencies in the health system will be fixed. In what was cynically regarded as an attempt to deflect blame away from the government in the form of a eulogy, President Chirac frames the disaster in terms of a social issue, and not a failure of leadership: "Our society must become more responsive and attentive to others, to their problems, their suffering and their vulnerability," he warned. "These dramas again shed light on the solitude of many aged or handicapped citizens" (Gentleman 2003).

Later he and other senior members would hold a state funeral for the bodies of 57 unclaimed Parisians, uncomfortably closing this unfortunate episode.

The Aftermath
Eventually on 25th September a finalised official figure was given by assessing excess mortality - the number of fatalities that occurred over and above a baseline level of mortality that would have been expected at the same time of the year. INSERM arrived at a figure of 14,802 excess deaths (60% more deaths than usual) for the period from 1st to 20th August, with these observations: