A Stitch in Time: Improving Public Health Early Warning Systems for Extreme Weather Events

Extreme weather events, particularly floods and heat waves, annually affect millions of people and cause billions of dollars of damage. In 2003, in Europe, Canada, and the United States, floods and storms caused 15 deaths and US$2.97 billion in total damages, and the extended heat wave in Europe caused more than 20,000 excess deaths (1); the impacts in developing countries were substantially larger. There is a growing body of scientific research suggesting that the frequency and intensity of extreme weather events are likely to increase over the coming decades as a consequence of global climate change (2). These events cannot be prevented, but their consequences can be reduced by taking advantage of advances in meteorologic forecasting in the development and implementation of early warning systems that target vulnerable regions and populations. The skill with which weather and climatic events can be forecast has increased significantly over the past 30 years as more has been learned about the climate system. During this period, weather forecasting improved from the same-day forecast to the advance forecast. Our understanding of the mechanics and teleconnections of El Niño/Southern Oscillation now provides us with the capacity for seasonal and annual forecasting—assumed as recently as the 1970s to be more science fiction than fact (3). In fact, Chen et al. (4) recently suggested that El Niño events can be predicted 2 years in advance. Public health professionals have the opportunity to integrate weather- and climate-related information into local and regional risk management plans to reduce the detrimental health effects of hazards as diverse as tropical cyclones, floods, heat waves, wildfires, and droughts (5, 6). Scientists began to keep instrumental records of temperature, precipitation, and other weather elements in the 1860s. In its Third Assessment Report, the Intergovernmental Panel on Climate Change evaluated this record and concluded that there was an increase in the frequency of extreme high monthly and seasonal average temperatures over the 20th century (5). The Intergovernmental Panel on Climate Change also concluded that precipitation increased by 0.5–1.0 percent per decade over most mid and high latitudes of the Northern Hemisphere continents, with a 2–4 percent increase in the frequency of heavy precipitation events at those latitudes during the latter half of the 20th century (7). In addition, it is very likely that El Niño events have been more frequent, persistent, and intense since the mid-1970s, in comparison with the previous 100 years. These trends are a consequence of a warming climate, particularly over the mid and high latitudes of continents in the Northern Hemisphere. Northern Hemisphere temperatures in the 1990s were higher than at any other time in the past 6–10 centuries (7). The warmth of the 1990s was outside the 95 percent confidence interval of temperature uncertainty, defined by historical variation, during even the warmest periods of the last millennium (8). The increase in temperature over the 20th century was likely to have been the largest of any century during the past millennium. The projections for the 21st century are for more, and more rapid, change in temperature and precipitation than was experienced during the last century (8). Interactions between changes in the mean and variability of weather variables complicate projections of possible future trends in extreme events (9). Assuming a normal distribution of surface temperatures, one can envision three scenarios of increasing temperature (10). In the first scenario, there is a simple shift in mean temperature without a change in the variance (e.g., the shape of the curve would remain the same). If that occurred, there would be a decrease in cold weather and an increase in both hot weather and record hot weather. The second scenario is an increase in the variance without a change in mean temperature; this would result in an increased frequency of cold and hot weather, with a decreasing frequency of weather that could be considered average under the previous climate (i.e., the shape of the curve would become flatter). Finally, if there were a shift in both the mean and the variance, there would be a small decrease in cold weather and a significant increase in both hot weather and record hot weather. The patterns for precipitation might differ, because precipitation is not well approximated by a normal distribution; there could be changes such as a shift in frequency or a shift in distribution, either of which could affect overall intensity. Because changes in the frequency of many extremes can be surprisingly large for seemingly modest changes in mean climate, there is growing concern that future weather patterns will resemble the third scenario and that what is currently considered an extreme event may become common (11). Easterling et al. (2) summarized results from the modeling of different types of climate extremes for the 21st century. For simple extremes based on climate statistics, the authors concluded that the following changes are very likely (90–99 percent probability) to occur by the end of the 21st century: higher maximum temperatures; more hot summer days; an increase in the heat index; more 1-day heavy precipitation events; and more multiday heavy precipitation events. In terms of complex event-driven climate extremes, it is very likely that there will be more heat waves, and it is possible (33–66 percent probability) that there will be more intense mid-latitude storms and more intense El Niño events. A likely consequence of increased climate variability will be surprises with regard to the timing, intensity, location, and duration of extreme weather events (12). This means that all mid-latitude regions need to be prepared for extreme weather events, regardless of whether they have occurred over the past century. Unfortunately, scientists and policy-makers have been focusing more on the uncertainty about the rate and intensity of changes in climate variability than on the certainty that without adequate preparation, more extreme events will lead to increased morbidity and mortality. Flooding and heat waves are of particular concern because of recent increases in mortality (13). Effective prediction and prevention programs that incorporate advances in climate forecasting can be designed and implemented with a better understanding of the subpopulations at risk and of the information needed for effective response to warnings (5, 6). The value of these early warning systems will increase as the projections of increased climate variability are realized. Extreme weather events cannot be prevented, but population vulnerability to these events can be reduced. Table 1 summarizes the health impacts of heat waves, extreme rainfall, floods, and droughts. Some of the health outcomes associated with extreme weather events have well-described etiologies; for example, there have been numerous studies of the association between heat waves and mortality and morbidity, and of the increased frequency of disease outbreaks following floods (14–16). Mortality and morbidity immediately and directly associated with an extreme event are often documented, but other health outcomes, such as the mental health effects associated with floods, are less well studied (14). More consistent and comprehensive data collection during and after extreme weather events will lead to a better understanding of the morbidity and mortality associated with such events. Summary of the health impacts of extreme weather events* Sources: Hajat et al. (14), Kovats et al. (15), Malilay (16), and Kilbourne (37). Summary of the health impacts of extreme weather events* Sources: Hajat et al. (14), Kovats et al. (15), Malilay (16), and Kilbourne (37). Surveillance during the time period immediately surrounding an extreme weather event is a key public health activity. Public health professionals must determine whether the event is associated with an increase in disease (such as diarrheal disease after a flood) so that appropriate measures (such as a “boil water” alert) can be instituted. In addition, surveillance of age- and cause-specific deaths over time is needed for calculation of baseline or normal mortality rates in order to recognize increases in mortality over what would have been expected (17). However, because there can be a considerable time lag between when deaths occur and when data become available for analysis by public health authorities, surveillance during an extreme event is insufficient for determining when to implement public health interventions. The limitations of existing surveillance were demonstrated during the 2003 European heat wave. This heat wave, which caused approximately 15,000 excess deaths in France during a 2-week period, took French public authorities by surprise (18). There are numerous reasons why timely responses to this event were not implemented, among them the fact that surveillance systems were not adequately designed to detect and respond to a heat wave. Surveillance systems did not provide authorities with information quickly enough to detect the increased number of deaths in time to implement interventions. A retrospective assessment found that there had been approximately 3,900 deaths at the time when only 10 deaths had been reported (18). Surveillance systems were not designed to recognize increases in morbidity and mortality among persons with chronic diseases, such as cardiovascular and respiratory conditions. In addition, even if surveillance systems had been adequate to detect an increase in morbidity, existing emergency public health interventions were not designed to address sudden increases in endemic and common diseases. Another problem was the large numbers of people at risk; it was estimated that there were 6 million people at risk during the heat wave, of which 1 million were at very high risk. There were few widely available and efficient measures in France for reducing heat-related mortality, especially with so many people at risk. Air conditioning may have saved lives but was generally not available, particularly for the populations at highest risk, such as elderly persons in nursing homes. France is now developing a heat health warning system and a response plan, which have been shown in other cities to reduce mortality from extreme heat (19). Improvements in surveillance systems can only partly address the health impacts of extreme weather events because, even if timely, recognition of an increase in adverse health outcomes does not always translate into practical and effective responses; all warnings are predictions, but not all predictions are warnings. Public health professionals need to increase their focus on prediction and prevention in order to mitigate the health consequences of extreme events. The addition of early warning systems to existing surveillance mechanisms, coupled with effective response capabilities, can both reduce current vulnerability and increase resilience to future extreme events (20). For example, early warnings of flooding risk have been shown to be effective in reducing flood-related deaths when the warning is coupled with appropriate responses by citizens and emergency responders (16). An example is the difference between the 1993–1994 flooding along the Rhine and Meuse rivers in Germany and the 1995 flooding along those same rivers (21). Persistent high precipitation caused both events. The two floods had similar characteristics, although the 1993–1994 flood had a second peak discharge. Ten people lost their lives in the 1993–1994 flood, and the total damage in Belgium, Germany, France, and the Netherlands was estimated at US$900 million (21–23). The economic cost of the 1995 flood in Germany was reduced by about half, presumably because people were aware of the risks and made appropriate behavioral and other changes. Surprisingly few heat wave early warning systems have been implemented, given the evidence that they can reduce mortality (19, 24). For example, the Philadelphia Hot Weather-Health Watch/Warning System was initiated in 1995 to alert the population to take a variety of precautionary actions during a heat wave (24). In an evaluation of the system, it was estimated that issuing a warning saved, on average, 2.6 lives for each warning day and for 3 days after the warning ended; the system saved an estimated 117 lives over a 3-year period. However, at the time of the 2003 European heat wave, only two cities in the World Health Organization European region—Rome, Italy, and Lisbon, Portugal—had operational heat health warning systems (25–27). In the United States, a recent review of 18 cities vulnerable to heat waves found that 10 had written response plans, with one third of the plans being no more than cursory (28). Australia, where (like the United States) heat waves are responsible for more deaths than all other natural disasters combined, has no operational systems. For flooding, which is the most common natural disaster in Europe, the emphasis in disaster management has been on postdisaster improvisation rather than predisaster planning (14). There is a need for more good-quality epidemiologic data before vulnerability indices can be used operationally to minimize the health effects of flooding. The studies that have been conducted have primarily focused on assessing larger events, although more frequent smaller events may result in a greater health burden. Furthermore, different types of floods may require different types of intervention strategies. Whereas surveillance systems are intended to detect disease outbreaks and measure and summarize data on such outbreaks as they occur, early warning systems for extreme weather and climatic events are designed to alert the population and relevant authorities in advance about developing adverse meteorologic conditions and then to implement effective measures to reduce adverse health outcomes during and after the event. A basic requirement for an early warning system is that the community or region have an adequate public health and social infrastructure, including the political will, to undertake the design and implementation of such a system. The principal components of an early warning system include identification and forecasting of the event, prediction of the possible health outcomes, an effective and timely response plan, and an ongoing evaluation of the system and its components. The system should be developed in collaboration with all relevant stakeholders to ensure that the issues of greatest concern are identified and addressed, thus increasing the likelihood of success. Stakeholders include the that will the development and of the system, the that will be expected to take and those likely to be by an extreme event. of persons by extreme events may provide local about responses and their an extreme event to be defined in collaboration with relevant and taking into the that determine risk in a particular can be as a of the that an adverse event will occur and the consequences of that event Public health responses should because the interventions developed for events with consequences (e.g., a heavy event not associated with will from the responses for events with consequences (e.g., the 2003 European heat A variety of determine whether an extreme event a risk, including economic the of the public health infrastructure, For example, what is considered a day by and the summer local of risk. Public health authorities must with meteorologic to the and of in order to for during an extreme event. for developing events, such as heat waves and floods, public health need to what information will be available so that the of interventions can be For example, heat wave early warning systems provide warnings in advance are implemented as are Climate change projections that past weather is no a for the early warning systems and other public health interventions need to incorporate plans for an The development of an early warning system should include the of scenarios for future weather to ensure that the system can the increasing frequency and of extreme events. For example, the weather that in the 2003 European heat wave was larger than considered in the development of the heat health warning systems implemented in and of This event may be a of future European and provides a warning to other regions In addition, it would be to scenarios events with impacts that the health For example, if a heat wave in the of that to a region and of were not available, responses would need to be developed for a of health impacts and other extreme heat a in Europe in might affect or damage such as in and that an extreme event has been forecast is insufficient information for the design of a response There to be a that can the meteorologic along with between weather conditions and health outcomes, to possible health in vulnerable populations. The of the should be of disease that can be and quickly so that effective responses can be The must be and timely, and prediction should not weather but also or that could affect the for an For example, heat waves may be associated with high of so should their and The of such must and have of uncertainty, from data to understanding of disease The and of a have for the design of interventions. a warning when was and issuing a warning when one was have not only in terms of morbidity and mortality but also in terms of public to on warnings. an understanding of these and their associated into the design of an early warning system could its predictions of extreme events and their health impacts are not for developing an early warning an effective response is for reducing the predicted of disease The design and implementation of response should be in a that all stakeholders in the and should the and political and of the region (12). The of an early warning response are by the and in which the system as well as the of the event (12). The of the including of and the interventions must be evaluated and to response for the relevant community or For example, heat wave early warning systems in the United have two of health a which that meteorologic conditions are such that a heat wave could and a which that a heat wave has The public the terms and because they have been used by the in early warning systems for other climatic events, such as This is not in other The United heat wave has of and A response for extreme weather events should incorporate components for both implementation and evaluation of The implementation at a include where the response will be when interventions will be implemented, including for what interventions will be the response will be and to the interventions will be The responsible and need to on for the response This is particularly an when a such as a flood, In addition, the should where interventions will be for example, if interventions will focus on the data on where the of elderly people need to be The of implementation of a response is by including the certainty in of extreme weather events uncertainty decreasing as the event the of lead time needed for and the at which responses are should a between actions when they are not needed and not actions when they would reduce morbidity and mortality. For example, one excess per day during a heat wave is to be to implementation of prevention at what number of deaths should actions be of an early warning system is on the of effective interventions for reducing the of there will be interventions for the such as on to body temperature to heat there should be particularly vulnerable such as the the and For example, designed to alert to the risk of in during a heat wave will from designed to to time at a These should be of a larger designed to increase public of appropriate behavioral and other responses in the event of extreme Stakeholders should be in the development of interventions to ensure their maximum There should be a written of the the interventions that will be will including the of each and information will be plans, emergency The of the response plan, along with the interventions in the plan, should be to all The should is responsible for the when the will be the key the key to be and they should be and the of the will be and evaluated are often not considered in the design of response plans and they can be particularly vulnerable because them to about actions to An of a early warning system is an evaluation of This should at a and evaluation of the overall of the and evaluation of the system and an economic assessment of the and of the system. and evaluation of the of the early warning system and interventions should be into the system design so that appropriate can be made to the of the system its Because is about the of interventions in reducing the morbidity and mortality associated with extreme events, research is needed to better where should be to the most effective research has been conducted on the of early warning systems. of the Philadelphia Hot Weather-Health Watch/Warning System found that the estimated of the system were with the estimated value of a Unfortunately, data were not available on the interventions in the system, so their could not be of the information on the of extreme weather is focused on and rather than adverse health events. a that used to the of other health can be that and outcomes associated with warning systems. should and be so that the of the system can be under different scenarios of and and should other of results should include the cost per event (e.g., will existing understanding of health economic modeling and evaluation of early warning systems to be with other health interventions. that early warning systems can reduce the morbidity and mortality associated with extreme weather events. be public health must from a focus on surveillance and response to a greater emphasis on prediction and is for public health to and with in the development of early warning systems. the increasing skill of in forecasting extreme events with effective public health interventions can disaster of more frequent and more intense extreme events in a climate increase the of and early warning systems that take into the of extremes outside the The of an early warning system will on its (i.e., whether it is designed for a heat wave or a on and on In such a system will include meteorologic predictions of health outcomes, a written response for and evaluation research is needed to the development of early warning particularly the collection of epidemiologic data on the of health risks associated with extreme events and better understanding of the of particular interventions. extreme weather events are well to be the morbidity associated with these events is surprisingly (14). In addition, data are generally only for large events, so the of smaller events is early warning systems a of interventions in response to a but few interventions have been evaluated for their are needed to ensure effective and efficient of of early warning systems in of increased climate variability can be as an of the precautionary The precautionary is an to public that can be used in of or to health or the where there is a need to to reduce hazards before there is of of a precautionary is before of is available, if impacts could be or This is the with projections of a future by increases in the frequency and of extreme weather events. The scientific that not only is climate change but its effects are also being must be into political before extreme weather events cost more The time to is

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.