LONDON had been hot and dry for days, and a thick layer of smog mingled with pollen hung over the city. As the thunder clouds rolled in on 24 June 2005, the UK鈥檚 午夜福利1000集合 Forecasting Unit put hospitals on standby and issued a health warning: all hay-fever sufferers, even those with no previous history of asthma should stay indoors or risk a severe asthma attack.
The authorities had good reason to be concerned. On the same day 11 years earlier, identical weather conditions had conspired to make pollen and spores more allergenic than usual, and 鈥渢hunderstorm asthma鈥� struck thousands of people, overwhelming the city鈥檚 emergency rooms. This time the hospitals were more prepared. Despite having up to eight times as many patients admitted with asthma attacks as usual, they had had some warning of the influx, and the public had clear advice on what preventive measures they should take.
The success of this warning and others like it has led to a surge of interest in forecasting how the weather may affect health. A growing body of evidence suggests that changes in temperature, humidity, air speed and pollution can lead to highly predictable increases in the number of people who have heart attacks, strokes and respiratory problems. The idea originated in the UK, but now epidemiologists and meteorologists worldwide are coming together to predict exactly how, where and when people will fall ill, and to issue specific health warnings to hospitals and patients.
Advertisement
Weather forecasts already giving us some degree of health-related advice: pollen counts, air-quality updates and sun strength indices have become a regular feature of summer bulletins, while advice to 鈥渨rap up warm鈥� and 鈥渙nly go out if absolutely necessary鈥� figure in many a winter forecast. What鈥檚 different about health forecasting is that it combines weather data specific to a local area with medical knowledge about how these conditions could increase the severity and prevalence of disease. The result is targeted advice that could save lives.
Shuaib Nasser at Addenbrooke鈥檚 Hospital in Cambridge, UK, has been investigating the links between thunderstorms and asthma since 2002, when a cluster of attacks emerged at the end of the grass pollen season in July. 鈥淭hunderstorm asthma is a very real phenomenon,鈥� says Nasser. 鈥淭here needs to be a way of forecasting, accurately, when a certain kind of thunderstorm is about to break, which patients are going to be at risk and to have a strategy for what to do next.鈥�
Red alert
This, he says, should be fairly easy to do. Thunderstorm asthma only appears to affect a subset of hay-fever sufferers and allergic asthmatics who react to pollen and, Nasser believes, fungal spores. It is also associated with very specific weather conditions 鈥� a combination of high pollen, pollution and hot, still air before a storm. As the clouds gather, strong updraughts suck pollutants and pollen grains up from the ground. The grains freeze and shatter when they reach the tops of the clouds, spilling their contents and becoming more allergenic. When rain carries these pollen fragments back down to the ground they set off violent allergic reactions. Understanding and predicting this sequence of events makes it much easier to forecast when and where such attacks are going to occur and who is likely to be worst affected.
But while these conditions may be relatively rare, normal seasonal fluctuations in temperature can also have serious consequences for health. As temperatures fall in winter, mortality rates begin to rise. A 1997 study published in The Lancet (vol 349, p 1341) by a group of European researchers showed that the impact of the cold on health depends on where you live. In temperate London there is a 1.4 per cent increase in mortality for every degree below 18 掳C, while the better prepared Finns suffer an increase of only 0.2 per cent. Roughly half of these extra deaths are caused by respiratory problems. Below 19 掳C the membranes in the nose produce less mucus and so become less effective at trapping germs and preventing disease, while cold air also causes constriction of the airways, resulting in reduced lung capacity.
Poor air quality adds to the effect. Pollution also reduces mucus in the nose and may cause inflammation of the alveoli in the lungs, blood coagulation and, possibly, strokes.
The other half of the cold-weather death toll is accounted for by effects on the cardiovascular system. Low temperatures put added stress on the body, as blood flow is diverted from the extremities to the central organs. Sudden drops in temperature can also trigger the release of adrenalin, which raises the blood pressure and heart rate.
The impact of low temperatures on health seems to be fairly predictable. In 1997 researchers at Queen Mary and Westfield College, University of London, looked at the effects of a cold snap on mortality over a period of 16 years. They found an increase in the number of heart attacks two days after a sudden drop in temperature, and a peak in the number of strokes five days after the cold spell began. Deaths caused by respiratory problems peak 12 days after a drop in temperature (see Graphic).
Persistent high temperatures, such as during the heatwave that claimed an estimated 35,000 lives in western Europe in August 2003, can increase death rates too, particularly among elderly people. 鈥淚t seems that the human body can cope with high daytime temperatures as long as it gets the chance to cool down overnight, but consistently high daytime and night-time temperatures take a cumulative toll on the body,鈥� says William Bird, clinical director of the 午夜福利1000集合 Forecasting Unit (HFU), part of the UK鈥檚 meteorological office, the Met Office. That hot weather can lead to heat exhaustion, heat strokes, heart attacks and strokes.
The good news is that deaths from extremes of cold and heat are preventable. The lower mortality rates triggered by cold weather in Finland compared with London have been attributed to measures as straightforward as warmer houses and more appropriate clothing. Many of the heat-related deaths in Europe in the summer of 2003 were put down to people drinking too little water to stay cool and hydrated.
Good advice
In an attempt to work out exactly what advice should be given in different weather conditions the Met Office launched a project in 2001 called Forecasting the Nation鈥檚 午夜福利1000集合. The aim was to assess the viability of using real-time National 午夜福利1000集合 Service (NHS) data, infectious disease data and weather forecasts to predict workloads in different parts of the NHS. The HFU is the result of that project鈥檚 success, and is now responsible for carrying out further research on specific disease areas.
Last winter the HFU ran a pilot project across the UK to try and forecast exacerbations of chronic obstructive pulmonary disease (COPD), a chronic respiratory disease that is linked to smoking and kills around 2.7 million people worldwide each year. COPD exacerbations tend to peak around 12 days after a drop in the temperature, and are associated with high humidity, low wind speeds and poor air quality, all of which increase the transmission of viruses such as respiratory syncytial virus and influenza, both of which can lead to chest infections.
Meteorological and health data were used to create twice-weekly forecasts that were sent out to participating health authorities, which were then able to target specific people to ensure that they were as well prepared and protected as possible. Advice included limiting journeys in order to cut exposure to pollution or low temperatures, and ensuring that if vulnerable people did go outside, they dressed warmly and covered their mouths in order to prevent them inhaling cold air, which could trigger constriction of the airways and make breathing more difficult. Nurses also encouraged people to make sure that their homes were warm enough, and that they had the correct medication.
鈥淒on鈥檛 go on unnecessary journeys, dress warmly and cover your mouth鈥�
Bird says that the trial went well, and anecdotal accounts from one small doctors鈥� practice suggested that hospital admissions dropped by 85 per cent over the duration of the trial. 鈥溛缫垢＠�1000集合 forecasting is about targeting patients with specific interventions at the time when there is most impact,鈥� he says. 鈥淲hat we found was that many patients in very high-risk groups actually moved down to a lower-risk group because of the early intervention and their increased knowledge about handling the disease.鈥�
But it is not only the patients who benefit, says Bird. 鈥淚t also saves hospitals significant amounts of money because they can plan when hospital beds are going to be most busy, resulting in fewer cancelled operations.鈥�
So successful was the project that this winter the Met Office has expanded COPD forecasting and is recording the results as part of a formal clinical trial. In the next few years the Met Office plans to introduce health forecasting for stroke, asthma and cardiovascular disease.
Not every area of health is as easily linked to changes in weather as asthma and COPD. Matthew Walters, a consultant physician at Glasgow鈥檚 Western Infirmary in the UK, is working on stroke forecasting. 鈥淓pidemiological studies suggest that the burden of stroke tends to increase during the winter months, with sudden temperature changes acting as a predictor for stroke risk,鈥� he says. But inconsistencies between studies mean there is a need for further research. He is planning to compare data from around 10,000 patients attending Glasgow鈥檚 acute stroke unit between June 1990 and September 2003 with atmospheric pressure and temperature data from the Met Office from the same period.
While the link between environmental conditions and strokes is still unclear, Walters thinks that there is potential for health forecasting in certain people. 鈥淭here are a large number of factors driving hospital admissions, but in a small subset of conditions health forecasting may have a significant impact in terms of predicting the number of patients we see at certain times of the year,鈥� he says.
Meanwhile, scientists at Australia鈥檚 national research organisation, CSIRO, have recently received funding to develop their own health forecasting system, which will use meteorological and air-quality readings to predict health outcomes in COPD, respiratory diseases and admissions to hospital with cardiovascular problems. And while modelling systems are limited to predicting admissions and issuing health warnings up to seven or 14 days ahead, there is the potential for predicting further into the future. 鈥淚 see that there is potential there,鈥� says Bill Physick, joint leader of CSIRO鈥檚 Air Quality Modelling and Dispersion Team. 鈥淢odels that are used for seasonal forecasting may have a role in predicting particularly wet, dry, hot or cold seasons, which could then be linked with various diseases, including vector-borne diseases.鈥�
So far the UK is leading the way in health forecasting. But health chiefs in the US are watching the progress with interest. 鈥淓ventually I think it is a transition we are going to see worldwide,鈥� says Maureen Bisognano, executive vice-president of the Institute for 午夜福利1000集合 Improvement, a not-for-profit organisation based in Cambridge, Massachusetts, which promotes better healthcare across the US. She would like to see some form of health forecasts broadcast alongside TV weather reports within a year. 鈥淲e already do it with pollen,鈥� she says. 鈥淎nd patients over here are already used to taking on responsibility for their own healthcare.鈥�
But it may not all be plain sailing. Paul Wilkinson of the London School of Hygiene and Tropical Medicine has been evaluating the COPD pilot for the Department of 午夜福利1000集合. His report, expected to be published in the British Medical Journal later this year, says that there is still not enough evidence to establish firm links between disease and weather conditions. Another unknown is whether a cash-strapped health system could ever deliver a working health-forecasting unit that would provide value for money.
鈥淐learly the idea is interesting and it has a basis in the fact that certain diseases are very variable across the year and in some cases are very predictable,鈥� he says. 鈥淏ut it still needs to be evaluated in a way that looks at the cost benefits.鈥�
Others think that there is a danger in issuing blanket health warnings that could eventually lead to scepticism and complacency.
鈥淎s far as thunderstorm asthma is concerned, one of the most important things to realise is that while there is a susceptible group within the asthmatic population, it is not all asthmatics and hay fever sufferers who are at risk,鈥� says Nasser. 鈥淚f you issue a thunderstorm warning to 100 asthmatics and only five actually get any symptoms, they are going to get disillusioned pretty quickly. There needs to be a way of focusing in on that specific group and targeting them directly.鈥�
Despite these concerns, excitement about the potential for health forecasting is growing, with Bird predicting that it could become common practice in the UK within 10 years. He says SMS text messages could be used to encourage the most vulnerable patients to take the right precautions. And if Bird鈥檚 wider vision is realised, news and weather bulletins will soon be followed by some helpful advice on surviving the coming weeks, beamed direct to our living rooms by a friendly face with a cheery smile.
