Andrew Haines, Paul R Epstein, Anthony J McMichael, on behalf of an international panel*
Climate change, the chosen focus of the Lancet series, could affect health in a variety of ways. Direct effects of a rise in temperature (particularly increases in the frequency and intensity of heatwaves) may include deaths from cardiovascular and cerebrovascular disease among the elderly. Indirect effects are secondary, such as changes in vector-borne diseases or crop production, and tertiary, such as the social and economic impacts of environmental refugees and conflict over fresh water supplies.
Traditional epidemiological monitoring of disease and mortality has limitations because there may be undesirable delays before changes in chronic diseases are detected. Other approaches must also be used, including biological markers to give early warning of damage, the monitoring of carriers of infection such as insects and rodents, and remote sensing for large-scale monitoring. There is growing awareness of the need to link environmental issues with health---for instance the 1993 World Bank report Investing in Health includes forest and fresh water resources. We argue for integration of health into existing and planned environmental monitoring systems. In this final article we consider five aspects of monitoring, with cross-reference to the series where appropriate: biological, environmental, and human health indicators; data needed to monitor indicators; technology for measuring them; organisations doing the work; and gaps in information.
The scientific assessment of climate change is being updated by the Inter-Governmental Panel on Climate Change (IPCC). The Second World Climate Conference in Geneva (1990) recognised the need for a Global Climate Observing System (GCOS) and a committee for GCOS has now been set up. GCOS will cover all components of atmosphere, biosphere, cryosphere, hydrosphere, and land surface climate, and that coverage is beyond the scope of current monitoring programmes such as Global Atmosphere Watch and the World Weather Watch network of satellites, telecommunication, and data processing facilities (figure 1).
Two other observing systems (ocean and terrestrial, GOOS and GTOS) will enable GCOS to provide a fuller picture. More than eighty international organisations and programmes are involved in global environmental monitoring, and the potential for overlap and lack of coordination is great. Until now health has not been adequately taken into account. A selection of these organisations is shown in figure 2.
The direct effects of temperature on health are mainly manifest as an increase in death rates amongst the elderly during periods of high temperature and can best be detected through analysis of mortality data collected daily. Such data are currently available mainly in developed countries but this information is needed for urban centres in less developed countries. Aggregation of deaths into weekly or monthly statistics is of much less value because an increase in mortality tends to be short-lasting and may be followed by a period of lower than expected mortality. Changes in morbidity and in seasonal patterns of disease can be detected in primary care data such as those collected from sentinel general practices around the UK. This database demonstrates, for instance, that consultations for asthma have risen lately (for reasons that are unclear) and that they show seasonal variations with peaks in the summer and towards the end of the year. The increasing use of computers should make it possible to collect routine data about consultations and hospital referrals in large populations.
Illnesses of plants, birds, fish and mammals can be indicators of environmental ill-health. The factors which influence the growth of parasites and pests are nutrients, competitors, predators, and climate. When more than one factor is disturbed at the same time, the system's resilience declines and its resistance to pests may decrease. Bioindicators are used to monitor environmental toxins. The abundance and distribution of key species such as insects and algae can be used as indicators of ecosystem health. When an indicator is also a disease, vector surveillance for health outcomes can be directly linked. The Global Terrestrial Observing System (GTOS) requires a network of sentinel sites. The only global network available now is that run by UNESCO's Man and Biosphere programme, comprising 311 biosphere reserves. Some of these may become part of GTOS and could also be suitable for monitoring of health-related indicators. As they are generally composed of natural systems they will need to be supplemented by other sites, including agricultural, rangeland, forestry and fresh water systems.
Vector-borne diseases (Freier, Rogers and Packer, Nicholls, Nov 20; Almendares and others, Dec 4)
The WHO task group identified several vector-borne diseases that might be influenced by climate change. Examples are malaria, lymphatic filariasis, African trypanosomiasis, dengue and yellowfever.
Changes in terrestrial ecosystems---detected for instance, by satellite imaging---can help monitor vector-borne diseases. In particular, vegetation indices produced by high-resolution radiometry have been correlated with mortality rates and population density of tsetse flies. Several types of remote sensing can be used to indicate animal and vector habitats; the LANDSAT and SPOT satellites (figure 1) have resolutions of 30 m and 10 m, respectively, and have been used to identify habitats of ticks and mosquitoes. The US National Aeronautics and Space Administration is sponsoring research on the use of satellite information for vector-borne disease monitoring and control. Improved surveillance systems should be incorporated within the next generation of earth observation platforms. Integrated systems combining meteorological, topographic, and epidemiological data must become more accessible and simpler to use.
Climate change may first have impact on vector-borne diseases at the margins of their current distributions. In global warming isotherms shift polewards and vectorborne disease may follow in the same direction 10[[ring]]C for yellowfever, 16deg.C for vivax malaria, 20[[ring]]C for falciparum malaria). Climate change might also affect the altitude at which vector-borne diseases are found, and high altitude sites in Kenya, Rwanda, Costa Rica, and Argentina may be good sites for monitoring. Field studies have been done but they must be kept going indefinitely. Low cost continuous monitoring may be possible through local primary care facilities with health staff trained to diagnose malaria and other conditions reliably and to keep accurate records.
In Latin America, Chagas' disease could be monitored in Chile and Argentina, currently at the edges of the endemic area. Schistosomiasis could also be susceptible to climate change, especially if irrigation patterns change. In the USA there is a possibility of the spread of five vector-borne diseases---malaria, yellowfever, Rift Valley fever, dengue fever, and arbovirus-induced encephalitides. The use of the Southern Oscillation Index, based on differences in atmospheric pressure, to predict outbreaks of Australian encephalitis was discussed by Nicholls.
Climate change may result in the elimination of some vectors and/or pathogens---for instance, as a result of very hot dry conditions, as in Honduras (Almendares et al). Local influences, such as deforestation, need to be distinguished from climate change.
Changes in coastal ecology from local and global influences have direct impacts on health. Environmental monitoring of nutrients, currents, algae, and fish must be supplemented by (1) monitoring algae for Vibrio cholerae; (2) surveillance of coastal communities for cholera and for fish (eg, ciguatera) and shellfish poisonings; and (3) surveillance of coral reefs (warming and ultraviolet radiation may cause bleaching).
Marine algal blooms can be detected by remote sensing and satellite radiometry is useful for monitoring sea surface temperatures to guide sampling (figure 3). Microwave bands (to measure salinity) may be helpful for following particular toxic phytoplankton species. The next generation of satellites (Sea WiFS, to be launched in early 1994) will improve monitoring. Remote sensing needs to be supplemented by local sampling to examine individual species of algae and zooplankton associated with gastrointestinal pathogens and biotoxins. Data on winds and currents, nutrients (including nitrogen and phosphorus originating from sewage), fertilisers, and industrial pollutants will help to determine when conditions are propitious for the growth of algal blooms. In 1994 the monitoring of large marine ecosystems (funded by the Global Environment Facility) is scheduled for the Gulf of Guinea, then the Yellow Sea, and ultimately the world's other 50 coastal marine ecosystems.
A temperature increase of 2.5[[ring]]C between 1990 and 2100 is projected to lead to a rise in sea level of 48 cm. The impact will depend on land subsidence, erosion, and the frequency and intensity of storms.
There are currently 204 monitoring stations for sea level rise with planned expansion to 306 in eighty-five countries. Measurements are improving under the auspices of the Global Seal Level Observing System (GLOSS), which has a tide-gauge network. The countries most vulnerable to a rise in sea level include Bangladesh, Egypt, Pakistan, Indonesia, and Thailand, all with large and relatively poor populations. Several low-lying islands such as Kiribati, Tokelau, and the Maldives would also be in danger. The health consequences will be direct (eg, due to flooding) and indirect effects (eg, due to displacement of populations and changes in vector habitats).
Fresh water is rapidly emerging as a limiting factor for human development. Rivers, lakes, and underground aquifers show widespread signs of degradation and depletion, even as human demands on water resources rise inexorably. Some twenty-six countries now have indigenous water supplies of less than 1000 m per person per year, a benchmark for chronic water scarcity. By the end of this decade, some 300 million people in Africa---one third of that continent's projected population---will be living in water-scarce countries. Although domestic water use accounts for less than one-tenth of water use, there already exists a large shortfall for safe drinking water. Globally, the expansion of irrigated areas---which currently produce one-third of the world's food---has slowed to about 1% per year whilst the world population grows annually by 1.7%.
Temperature increases resulting from the equivalent of a doubling of the concentration of heat-trapping gases will probably raise both evaporation and precipitation globally by 7-15%. Rainfall patterns will shift, with some areas getting more moisture and others less. Hurricanes and monsoons may intensify and the sea level rise will salinate some supplies of fresh water.
There is no global monitoring of water quantity, although most countries individually monitor the flows of rivers and the levels of lakes. The Global Runoff Data Centre, under the auspices of the WMO and based in Koblenz, Germany, maintains a database on daily river flows from 1664 stations in ninety-one countries. These data could serve as a baseline for examining possible shifts resulting from climate change were a global system to be established. The monitoring of water quality on a global scale is the responsibility of the WHO/UNEP Global Environment Monitoring System (GEMS). It promotes the measurement of about fifty indices of quality but practice among the 340 stations in forty-one countries varies considerably. The monitoring of pollutants and bacteria are relevant to climate change because changes in runoff may alter the concentrations; however, it is the use of fertilisers and pesticides, irrigation patterns, and industrial effluents that are key determinants of pollutant levels.
A specific fresh water indicator of warming could be algal blooms, measured as chlorophyll a. There is increasing awareness of the formation of large floating masses of blue-green algae. Certain species can produce toxins which may be poisonous, and rashes, eye irritation, vomiting, diarrhoea, and myalgia have occurred in people who swim through algal blooms. The blooms are considered to be caused by a combination of calm sunny periods and sufficient nutrients, notably phosphorus.
Several systems have been developed by international agencies to provide early warning of food shortages, notably in Africa. These systems rely on routine data of three sorts, that indicate food supply, food access, and wellbeing. Data obtained on the ground, such as food stocks and planted areas early in the season, supplement satellite data to indicate supply; food prices in local markets reflect access; and anthropometric measures or, in extreme cases, mortality rates give evidence of health impacts on populations.
Satellite data, as indicators of food supply and impending famine, improve consistency among countries and are more accurate and more timely than information had from farmers or local markets, for example. "Greenness" indices (red and near-infrared spectral reflectance) are available from daily data from satellites. This index is linked closely to cereal and forage production, and can be used to predict locust infestations. ure 4 illustrates this approach for Ethiopia. Rainfall estimates are based on duration of cloud cover (presumed to indicate rain).
One International Geosphere Biosphere Programme project is a global network modelling crop yield responses to environmental change. Another, jointly with an International Social Science Council programme on dimensions of human environmental change, will monitor long-term changes in global land for agricultural use driven by non-climatic influences such as population growth and trade agreements.
Agricultural yields can also be affected by pests and predators, which are themselves susceptible to climate change. Potential examples from the USA are anaplasmosis (a rickettsial disease of cattle) and hornfly.
To assess the impact of enhanced ultraviolet-B (UV-B) radiation resulting from stratospheric ozone depletion, two trends must be monitored---global changes in column ozone abundance and changes in UV-B flux at ground level.
Ozone trends have been monitored by instruments on a satellite, and by ground-based spectrometers. Observations on trace gases (especially chlorine and bromine containing compounds) that catalytically deplete the ozone layer are needed to predict future trends in ozone loss. A light weight unmanned aircraft shows much promise here; a fleet of them could fly for days or even weeks at a time in the lower stratosphere and provide continuous data that remote sensing techniques cannot. Serious international cooperation on monitoring UV-B has only just begun, although many governmental agencies are now acquiring the expensive instrumentation. In the UK for example the National Radiological Protection Board has been monitoring solar UV (visible, UV-A, and erythemally weighted UV-B) at three sites since 1988. Until recently, only broad-band measurements of UV-B region were available but we now have instruments that provide spectral resolution, and serial data from Toronto, Canada, published last month illustrate what can be achieved.
Epidemiological monitoring of skin cancers (basal cell, squamous cell, melanoma), cataract, and other possibly UV-B induced disorders of the eye are needed over a range of latitudes. Recently studies have been initiated in southern Chile, where there has been appreciable stratospheric ozone depletion. Whilst data on melanoma can be captured by cancer registry data, basal cell and squamous cell cancers may be less reliably reported. Reliable estimates of cataract prevalence are likely to require periodic epidemiological surveys using a standard system to grade lens opacity. However, these potential effects may take years to become manifest so markers which respond more rapidly are needed. The International Agency for Research on Cancer is exploring methods of making early estimates of changes in skin cancer risk. One possibility is to use biological markers, for instance certain dimer-forming mutations of the p53 gene in skin cells, which appear to be related to UV exposure."
Emerging infectious diseases are infections that are new in the population or are rapidly increasing in incidence or expanding in geographical range; examples are dengue, hantavirus pulmonary syndrome, and some haemorrhagic fevers. Most emerging diseases are caused by "microbial traffic"---that is, the introduction and dissemination of existing agents into human populations either from other species or from smaller populations. This process is often precipitated by ecological or environmental change and is facilitated by population movements and other social factors. Re-emerging diseases are those that had been decreasing but are now rapidly increasing again. Often previously active control programmes against well-recognised threats to public health have been allowed to lapse.
Our capabilities for health monitoring and rapid response are seriously fragmented, with insufficient coordination and communication let alone provision for future needs. Inexpensive reliable communications (eg, by e-mail) are still not available worldwide, although initiatives such as SatelLife's HealthNet, providing lowcost access to medical databases for remote areas, and Internet e-mail offer hope that this can soon be achieved. A secondary network directly linking interested field scientists could greatly aid early recognition.
In conventional epidemiological surveillance, only a small fraction of cases may be recognised and reported. With emerging diseases, even a single unusual incident can be significant and investigation of such a pointer requires linked capabilities for clinical identification of a "new" syndrome or disease outbreak, for the epidemiological investigation of the event (usually the weakest link), and for laboratory characterisation. Existing facilities with all the necessary capabilities, including some WHO collaborating centres for arboviruses and haemorrhagic fevers, can be a starting-point. ProMED (International Program for Monitoring Emerging Diseases) has lately been proposed and the idea is supported by the Federation of American Scientists and by WHO. Targeting so-called "critical geographical areas" undergoing rapid ecological or demographic change would be most effective. US Centers for Disease Control and Prevention has lately set up a programme on emerging diseases.
WHO could have a key role in coordinating a "Global Health Watch" (in quotes because there is no such system) based on environmental health initiatives in its regional centres. It will therefore need to be involved in the design and implementation of aspects of GTOS and GOOS. It can help select sentinel populations in critical regions where specific impacts seem most likely. Monitoring of health and climate change should be linked to information about the global health picture, including population growth. Existing collaborative programmes with other UN agencies (FAO, ILO, UNEP) places WHO in an excellent position to promote interdisciplinary activity on climate and ecosystem health.
The WHO database Climedat specifically focuses on work on the public health aspects of global climate change. It lists investigators, organisations, and projects dedicated to research on climate health.+ In addition the UN International Decade for National Disaster Reduction can provide practical input on preparedness and mitigation.
Greater integration of efforts to collect data on health and global environmental change is needed. Many of the potential effects of climate change will be insidious and will take a long time to manifest themselves, and sometimes the links between ecosystem damage and health are unclear. However, the creation of a monitoring network must not be used as a "wait and see" argument against action to reduce greenhouse gas emissions. The Framework Climate Change Convention signed in Rio de Janeiro last year has not yet come into force (it must first be ratified by fifty countries) but may be in 1994. It stipulates only that developed countries should reduce their carbon dioxide emissions to 1990 levels by the year 2000, whereas the IPCC states that a 60% reduction is required to stabilise atmospheric concentrations.
Much of the burden of global environmental change may fall on poorer countries, which are less well equipped to monitor, and the danger is that monitoring will focus disproportionately on the problems affecting the rich nations. This raises important ethical and practical issues. If monitoring is to be effective international collaboration on epidemiological surveys, field studies, and routine data collection to complement satellite data will have to improve.
This means a partnership between the technically advanced nations with access to remote sensing capacity, for example, and others. A global health monitoring network is essential not only to determine the impact of climate change but also to shape strategies to prevent climate change as far as possible and mitigate those effects which do occur.
* Dr Paul R Epstein (Harvard Medical School, USA), Prof Andrew Haines (UCLMS, Whittington Hospital, UK), Dr Martin Hugh-Jones (WHO Collaborating Center, School of Veterinary Medicine, Louisiana State University, Baton Rouge, USA), Dr Charles F Hutchinson (College of Agriculture, University of Arizona, Tucson, USA), Prof Laurence S Kalkstein (University of Delaware, Newark, USA), Dr Steven A Lloyd (Harvard University, Cambridge, USA), Prof Anthony J McMichael (University of Adelaide, South Australia), Dr Stephen S Morse, (Rockefeller University, New York, USA), Dr Neville Nicholls (Bureau of Meteorology Research Centre, Melbourne, Australia), Prof Martin Parry (Environmental Change Unit, University of Oxford, UK), Dr Jonathan Patz (World Health Organization, Geneva, Switzerland), Dr Sandra Postel (World Watch Institute, Washington DC, USA), Dr Kenneth Sherman (US National Oceanic and Atmospheric Administration, Narrangansett, Rhode Island), and Dr Rudi Slooff (World Health Organization, Geneva, Switzerland).
Correspondence to: Prof Andrew Haines, Department of Primary Health Care. UCLMS, Whittington Hospital, Highgate Hill, London N19 5NF, UK
1 McMichael A. Global environmental change and human population health: a conceptual and scientific challenge for epidemiology. Int J. Epidemiol 1993; 22:1-8.
2 Haines A, Fuchs C. Potential impacts on health of atmospheric change. J. Publ Health Med 1991; 13:69-80.
3 World Bank. World development report 1993: investing in health, world development indicators. Oxford: Oxford University Press, 1993.
4 Houghton JT, Collander BA, Varney SK. Climate 1992: the supplementary report to the IPCC scientific assessment. Cambridge University Press, 1992.
5 Fleming DM, Norbury CA, Crombie TL. Annual and seasonal variations in the incidence of common diseases. Roy Coll Gen Pract Occ Pap 1991; no 53.
6 WHO Task Group. Potential health effects of climate change. Geneva: WHO, 1990: 58.
7 Epstein PR, Rogers DJ, Sloff R. Satellite imaging and vector-borne disease. Lancet 1993; 341: 1404-06.
8 Maurice J. Fever in the urban jungle. New Sci Oct 16, 1993: 25.
9 Longstreth JA. Human health. In: Smith JB, Tirpak D, eds. The potential effects of global climate change on the United States. Washington, DC: Environmental Protection Agency, 1989.
10 Gleason DF, Wellington GM. Ultraviolet radiation and coral bleaching. Nature 1993; 365-836-38.
11 Wigley TMC, Raper SCV. Implications for climate and sea level of revised of IPCC emission scenarios. Nature 1992; 357:293.
12 Postel S. Last oasis: facing water scarcity. New York: WW Norton, 1992.
13 Waggoner PE, ed. Climate change and US water resources. New York: Wiley, 1990.
14 Rodenburg E. Eyeless in Gaia: the state of global environmental monitoring. Washington DC: World Resources Institute, 1991.
15 WHO Commission on Health and Environment. Our planet, our health. Geneva: WHO, 1992.
16 Elder GH, Hunder PR, Codd GA. Hazardous freshwater cyanobacteria (blue green algae). Lancet 1993; 341:1519-20.
17 Rosenzweig C, Daniel MM. Agriculture. In: Smith JB, Tirpak D, eds. The potential effects of global climate change in the United States. Washington, DC: US Environmental Protection Agency, 1989.
18 Gleason JF, Bhartia PK, Herman JR, et al. Record low global ozone in 1992. Science 1993; 260:523-4.
19 Dean SF, Rawlinson AI, McKinlay AF, et al. NRPB solar radiation measurement system. Radiol Protec Bull 1991; 124:6-11.
20 Kerr JB, McElroy CT. Evidence for large upwards trends of ultraviolet-b radiation linked to ozone depletion. Science 1993; 262:1032-34.
21 Chylack L, Wolfe JK, Singer DM, et al. The lens opacities classification system III. Arch Ophthalmol 1993; 111:831-837.
22 Nakazawa H, English D, Randell PL, et al. UV and skin cancer specific P53 gene mutation in normal skin as biologically relevant exposure measurement. Proc Natl Acad Sci (in press).
23 Centers for Disease Control and Prevention. Update: hantavirus pulmonary syndrome--United States 1993. JAMA 1993; 270:2287-88.
24 Henderson DA. Surveillance systems and intergovernmental cooperation. In; Morse SS, ed. Emerging viruses. New York: Oxford University Press, 1993: 283-89.
25 Lederberg J, Shope RE, Oaks SC Jr, eds. Emerging infections: microbial threats to health in the United States. Washington, DC: National Academy Press, 1992.
26 World Health Organization. WHO consultation for the development and use of environmental health indicators in the management of environmental risks to human health. Geneva: WHO, 1993.
27 Brown K, Maddison D. The UK and the global environment: the conventions on climate change and biological diversity. In Pearce D, ed. Blueprint 3; measuring sustainable development. London: Earthscan (in press).