Kristie L. Ebi, Exponent Health Group Vulnerability and Adaptation Assessments Hands-On Training Workshop Human Health Sector Kristie L. Ebi, Exponent Health Group Full references can be found in Chapter 11, Bibliography, of the Handbook.

Outline Overview of the potential health impacts of climate variability and change Health data to determine the current burden of climate-sensitive diseases Methods and tools for V&A assessment in the health sector Methods for determining a health adaptation baseline

Overview of the Potential Health Impacts of Climate Variability and Change

Topics Pathways for weather to affect health Potential health impacts of climate change Extreme weather events El Nino and disease Temperature Floods Vector-borne diseases Diseases related to air pollution Diarrheal diseases

Pathways from Driving Forces to Potential Health Impacts This shows the pathways by which climate change and other drivers can affect human health. Climate change will act through regional weather changes to affect health directly (temperature-related illness and death; and extreme weather-related health effects) and indirectly (air pollution-related health effects; water- and food-borne diseases; vector- and rodent-borne diseases; and mental, nutritional, infectious, and other health effects). The extent to which health impacts will be realized depends on the effectiveness of adaptation measures and on modulating influences (other drivers of health outcomes, such as population density in regions vulnerable to flooding). Corvalan et al., 2003

Pathways for Weather to Affect Health: Example = Diarrheal Disease Distal Causes Proximal Causes Infection Hazards Health Outcome Temperature Humidity Precipitation Survival/ replication of pathogens in the environment Consumption of contaminated water Incidence of mortality and morbidity attributable to diarrhea Living conditions (water supply and sanitation) Contamination of water sources Consumption of contaminated food This diagram shows the various pathways by which weather and other factors can influence the morbidity and mortality due to diarrheal diseases. Contamination of food sources Contact with infected persons Food sources and hygiene practices Vulnerability (e.g. age and nutrition) Rate of person to person contact WHO

IPCC TAR–Potential Health Impacts of Climate Change Any increase in climate extremes (storms, floods, cyclones) could increase the risk of infectious disease epidemics, particularly in low-income countries Increase in heatwaves, often exacerbated by increased humidity & urban air pollution Increase in the geographic range of potential transmission of malaria & other vector-borne diseases Increase in water- and food-borne diseases The severity of impacts will depend on the capacity to adapt & its effective deployment

Drivers of Health Issues Population density Urbanization Public health infrastructure Economic and technologic development Environmental conditions Populations at risk Poor Children Increasing population of elderly residents Immunocompromised Climate change is not the sole determinant of climate-sensitive diseases. This lists some other important drivers.

ENSO and Disease Kovats et al., 2003

Exploring Linkages Between ENSO and Human Health

Dengue Epidemics in South Pacific 1970-1999

El Nino starts El Nino stops This shows the relationship between the beginning of an El Niño and the number of cholera cases in Uganda. El Niño events have been shown to be associated with outbreaks of other diseases such as malaria. Dr. Githeko, personal communication

Climate Change May Entail Changes in Variance, as Well as Changes in Mean Interactions between changes in the mean and variability of weather variables complicate projecting possible future trends in extreme events. Assuming a normal distribution of surface temperature, one can envision three scenarios of increasing temperatures. 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 this occurs, then there would be a decrease in cold weather and an increase in both hot and record hot weather. A second scenario is an increase in the variance without a change in mean temperature; this would result in increasing cold and hot weather, with a decreasing frequency of weather that could be considered average under the previous climate (e.g., the shape of the curve would become flatter). Finally, if there is a shift in both the mean and the variance, then there would be small decrease in cold weather and a significant increase in both hot and record hot weather. Folland et al., 2001

Temperature Extremes in the Caribbean, 1955-2000 These data were presented by Michael Taylor (University of the West Indies) at a workshop in Barbados on small island states organized by WHO, WMO, and UNEP in May 2002 (workshop report available from WHO; Aron et al., 2003). The graphs shows the trend for minimum and maximum temperatures in the Caribbean, along with the percentage of days that temperature was greater than the 90th percentile (for minimum and maximum temperature), relative to the period 1977-1997. Temperature extremes have been increasing since the late 1950s. These observed changes are considered consistent with patterns related to climate change.

Climate Variability and Change Impacts in the Caribbean DATE COUNTRY EVENT DEATH ESTIMATED COSTS (US$ million, 1998) 1974 Honduras Hurricane Fifi 7,000 1,331 1982/3 Bolivia, Ecuador, Peru El Niño 5,661 1997/98 Bolivia, Colombia, Ecuador, Peru 600 7,694 1998 Central America Hurricane Mitch 9,214 6,008 Dominican Republic Hurricane Georges 235 2,193   Cuba 6 N/A 1999 Venezuela Landslide 25,000 Fuente: ECLAC, América Latina y El Caribe: El Impacto de los Desastres Naturales en el Desarrollo, 1972-1999, LC/MEX/L.402; OFDA, Venezuela- Floods, Fact Sheet #10, 1/12/ 2000.

Mechanisms by Which Above Average Rainfall Can Affect Health Event Description Potential Health Impact Heavy precipitation “Extreme event” Increased or decreased mosquito abundance Flood River/stream over tops its banks Property or crops damaged Above plus contamination of surface water Catastrophic flood/disaster Above plus increased risk of respiratory and diarrhoeal disease, injuries, etc. Kovats et al., 2003

Health Impacts of Floods Immediate deaths and injuries Nonspecific increases in mortality Infectious diseases – leptospirosis, hepatitis, diarrheal, respiratory, and vector-borne diseases Exposure to toxic substances Mental health effects Increased demands on health systems Flooding results in more than immediate deaths and injuries, as listed on this slide. Philip Wijmans, LWF/ACT Mozambique, March 2000

Mechanisms by Which Drought Can Affect Health Description Potential Health Impact Soil moisture decreases Changes in vector abundance Decreased crop production Depends on socioeconomic factors Reduction in food or water supply and quality Food shortage, illness, malnutrition, increased risk of disease Food shortage leading to deaths Death, starvation, risks associated with population displacement Kovats et al., 2003

Examples of Environmental Changes and Possible Effects on Infectious Diseases Example Disease Pathway of Effect Dams, canals, irrigation Malaria Increase breeding sites for mosquitoes Urbanization Cholera Decreased sanitation & hygiene, increased water contamination Reforestation Lyme disease Increase tick hosts, outdoor exposure Ocean warming Red tide Increase toxic algal blooms Patz et al., 2003 Wilson 2001

Factors that Influence the Range and Prevalence of Infectious Diseases Sociodemographic influences Human travel, trade, and migration Disease control efforts Drug resistance Nutrition Environmental influences Land-use, including deforestation, agricultural development, and urbanization Ecological influences

Temperature and Precipitation Effects on Vector- and Rodent-Borne Diseases Survival and reproduction rate of the vector Time of year and level of vector activity, specifically the biting rate Rate of development and reproduction of the pathogen within the vector

Main Types of Transmission Cycles for Infectious Disease Patz et al., 2003

Potential Transmission of Schistosomiasis, Jiangsu Province Yang et al., 2005

Climate Change and Malaria under Different Scenarios (2080) Increase: East Africa, Central Asia, Russian Federation Decrease: Central America, Amazon [within current vector limits] Van Lieshout et al. 2004 A1 A2 Source: Van Leishout et al., 2004. Based on the MIASMA model (V2.2) developed by Martens and colleagues. The model links GCM climate scenarios with an impact module that applies the formula for the basic reproduction rate to calculate the transmission potential of the malaria mosquito population, and to estimate the population at risk. The population at risk was defined as the total population living in an area where conditions were suitable for malaria transmission as defined by the transmission potential, and an average monthly precipitation of 80 mm. The reference scenario included population growth and kept the climate conditions the same as in the baseline climatology of 1961-1990. The model estimates climate suitability for an average year and the suitability for stable or annual transmission. The model assumes the current level of adaptation to malaria (countries were classified into one of six groups based on expert judgment). Estimates of the additional population at risk for more than 1 consecutive month of transmission by the 2080s ranged from more than 220 million (A1) to over 400 million (A2) when climate factors and population growth are considered in the model. The figure shows, under each scenario, the change in risk classified by the changes in the number of consecutive months of transmission (> +2, +2, -2, < -2). B1 B2 Van Lieshout et al. 2004

China Haze 10 January 2003 This aerial photo from NASA gives an indication of the importance of dust storms. Few studies have been conducted of the extent of illness and death occurring during dust storms. NASA

Effect of Temperature Variation on Diarrheal Incidence in Lima, Peru Daily Diarrhea Admissions Daily Temperature Diarrheal diseases are also climate sensitive. These graphs show the variation in daily diarrheal admissions with the variation in daily temperature. Adjusting statistically for the long-term trend and for other seasonal effects, diarrhea increases by 8% for each 1°C increase in temperature. Where the dose-response relationship has been quantified, similar patterns have been observed in other developing countries, i.e., 3% increase for each 1°C increase in Fiji. A dose-response relationship can be applied to temperature change maps for different climate change scenarios. Diarrhea increases by 8% for each 1ºC increase in temperature Checkley et al., 2000

Resources McMichael, A.J., D.H. Campbell-Lendrum, C.F. Corvalan, K.L. Ebi, A. Githeko, J.D. Scheraga, and A. Woodward (eds.). 2003. Climate Change and Human Health: Risks and Responses. WHO, Geneva. Summary pdf available at http://www.who.int/globalchange/publications/cchhsummary/ Kovats, R.D., K.L Ebi, and B. Menne. 2003. Methods of Assessing Human Health Vulnerability and Public Health Adaptation to Climate Change. WHO/Health Canada/UNEP. Pdf available at http://www.who.dk/document/E81923.pdf

Health Data to Determine the Current Burden of Climate-Sensitive Diseases

Questions to be Addressed What climate-sensitive diseases are important in the country or region? What is the current burden of these diseases? What factors other than climate should be considered? Water, sanitation, etc. Where are data available? Are health services able to satisfy current demands?

Health Data Sources World Health Report provides regional-level data for all major diseases http://www.who.int/whr/en Annual data in Statistical Annex WHO databases Malnutrition http://www.who.int/nutgrowth/db Water and sanitation http://www.who.int/entity/water_sanitation_health/database/en Ministry of Health Disease surveillance/reporting branch

Health Data Sources – Other UNICEF at http://www.unicef.org CRED-EMDAT provides data on disasters http://www.em-dat.net Mission hospitals Government district hospitals

Indonesia Total population = 219,883,000 Annual population growth rate = 1.4% Life expectancy at birth = 67 years Under age 5 mortality rate = 41/1,000 70% of 1-year-olds immunized with 3 doses of DTP 3.2% of gross domestic product spent on health This is an example of data from the World Health Report. WHO, 2005

Methods and Tools for V&A Assessment in the Health Sector

Methods and Tools Qualitative assessments Methods of assessing human health vulnerability to climate change WHO Global Burden of Disease Comparative Risk Assessment Environmental Burden of Disease MIASMA Other models

Qualitative Assessments Available data allow for qualitative assessment of vulnerability For example, given current burden of diarrheal diseases and projected changes in precipitation, will vulnerability remain the same, increase, or decrease?

Methods of Assessing Human Health Vulnerability and Public Health Adaptation to Climate Change Publication available from WHO European Centre for Environment and Health. Kovats et al., 2003

Methods for: Estimating the current distribution and burden of climate-sensitive diseases Estimating future health impacts attributable to climate change Identifying current and future adaptation options to reduce the burden of disease Kovats et al., 2003

Estimate Potential Future Health Impacts Requires using climate scenarios Can use top-down or bottom-up approaches Models can be complex spatial models or be based on a simple exposure-response relationship Should include projections of how other relevant factors may change Uncertainty must be addressed explicitly Kovats et al., 2003

Case Study: Risk of Vector-Borne Diseases in Portugal Four qualitative scenarios developed of changes in climate and in vector populations Vector not present Focal distribution of vector Widespread distribution of vector Change from focal to potentially regional distribution Expert judgment determined likely risk under each scenario for 5 vector-borne diseases Casimiro et al., 2006

Portuguese National Assessment Vector Parasite   None Present Imported human cases only Low prevalence in vectors/hosts High prevalence vectors/hosts None Present Negligible Risk Focal Distribu-tion Very low Low Regional Medium Wide-spread High Casimiro & Calheiros 2002

Sources of Uncertainty Data Missing data or errors in data Models Uncertainty regarding predictability of the system Uncertainty introduced by simplifying relationships Other Inappropriate spatial or temporal data Inappropriate assumptions Uncertainty about predictive ability of scenarios Kovats et al., 2003

Estimating the Global Health Impacts of Climate Change What will be the total potential health impact caused by climate change (2000 to 2030)? How much of this could be avoided by reducing the risk factor (i.e. stabilizing greenhouse gas (GHG) emissions)? McMichael et al., 2004

Comparative Risk Assessment Time Greenhouse gas emissions scenarios 2020s 2050s 2080s Global climate modelling: Generates series of maps of predicted future climate The comparative risk assessment approach used scenarios of greenhouse gas emissions, which were input into global climate models, the output of which was used with health impact models to estimate the total burden of disease under various scenarios. The burden of disease estimates generated were numbers of deaths and disability-adjusted life years lost (DALYs) that take into account both morbidity and mortality associated with the health outcome. 2080s 2050s 2020s Health impact model: Estimates the change in relative risk of specific diseases McMichael et al., 2004

Criteria for Selection of Health Outcomes Sensitive to climate variation Important global health burden Quantitative model available at the global scale McMichael et al., 2004

Health Outcomes Considered Outcome Class Incidence / prevalence Outcome Direct effects of heat and cold Incidence Cardiovascular disease deaths Foodborne & waterborne diseases Diarrhea episodes Vector-borne diseases Malaria cases Natural disasters Deaths due to unintentional injuries Other unintentional injuries Risk of malnutrition Prevalence Non-availability of recommended daily calorie intake McMichael et al., 2004

Exposure: Alternative Future Projections of GHG Emissions Unmitigated current GHG emissions trends Stabilization at 750 ppm CO2-equivalent by the year 2210 Stabilization at 550 ppm CO2-equivalent by the year 2170 Average climate conditions for 1961-1990 (WMO climate normal baseline) Source: UK Hadley Centre models McMichael et al., 2004

Estimated Mortality (000s) Attributable to Climate Change, 2000 Mal-nutrition Diarrhea CVD All Causes Deaths / Million SEAR-B 1 2 7.9 SEAR-D 52 22 7 80 65.8 McMichael et al., 2004

Climate scenarios, as function of GHG emissions The results for diarrhea show relatively modest increases in relative risks the developing regions of the world. However, because the absolute number of people at risk is large, these suggest that many more people could be at risk under these climate change scenarios. Further, this is a relatively conservative estimate, because it is looking at direct temperature effects only, ignoring possible effects on diarrhea acting through lack of clean water, etc.

Conclusions Climate change may already be causing a significant burden in developing countries Unmitigated climate change is likely to cause significant public health impacts out to 2030 Largest impacts from diarrhea, malnutrition, and malaria Uncertainties include: Uncertainties in projections Effectiveness of interventions Changes in nonclimatic factors McMichael et al., 2004

Environmental Burden of Disease A. Prüss-Üstün, C. Mathers, C. Corvalan, and A. Woodward. 2003. Introduction and Methods: Assessing the Environmental Burden of Disease at National and Local Levels [pdf available at http://www.who.int/peh/burden/burdenindex.html] Climate change document will be published soon

Climate and Stable Malaria Transmission Climate suitability is a primary determinant of whether the conditions in a particular location are suitable for stable malaria transmission A change in temperature may lengthen or shorten the season in which mosquitoes or parasites can survive Changes in precipitation or temperature may result in conditions during the season of transmission that are conducive to increased or decreased parasite and vector populations In Africa, stable malaria refers to falciparum malaria, the most serious form of malaria. More than 90% of malaria in Africa is falciparum malaria.

Climate and Stable Malaria Transmission (continued) Changes in precipitation or temperature may cause previously inhospitable altitudes or ecosystems to become conducive to transmission. Higher altitudes that were formerly too cold or desert fringes that were previously too dry for mosquito populations to develop may be rendered hospitable by small changes in temperature or precipitation.

Relationship between Temperature and Daily Survivorship of Anopheles This graph shows the proportion of Anopheles mosquitoes surviving for one day at different mean temperatures. The proportion of surviving mosquitoes declines rapidly over mean temperatures of about 37°C. The proportion of mosquitoes surviving two days is a function of how many survived the first day, plus any new mosquitoes added to the population the first day.

Relationship between Temperature and Time Required for Parasite Development This graph shows the number of days required for malaria parasite development as a function of mean temperature. At mean temperatures less than 21°C, mosquitoes have to survive at least 20 days after they acquire the malaria parasite to be infective. As mean temperature rises to about 21°C, the time required becomes shorter. At 30°C, it takes just a few days for mosquitoes to become infective.

Proportion of Vectors Surviving Time Required for Parasite Development This graph shows the proportion of vectors surviving long enough for the malaria parasite to develop at different mean temperatures. At about 27- 32ºC, somewhat more than one-third of vectors survive long enough to pass along the parasite.

and regional and country-level maps MARA/ARMA was a large project designed to map malaria risk in Africa. The website provides a wealth of information on malaria prevalence and population data at regional and country levels. Inquiries to the website are answered quickly. The website [http://www.mara.org.za] contains prevalence and population data, and regional and country-level maps

MARA/ARMA analyzed the listed environmental data in relation to malaria prevalence.

MARA/ARMA Model Biological model that defines a set of decision rules based on minimum and mean temperature constraints on the development of the Plasmodium falciparum parasite and the Anopheles vector, and on precipitation constraints on the survival and breeding capacity of the mosquito CD-ROM $5 for developing countries or can download components from website: www.mara.org.za

This map, from the MARA/ARMA website, shows the distribution of endemic malaria in Africa (endemic means that malaria is always present, although not necessarily for all months of the year). As stated, the model is based on the biological constraints on the vector and parasite.

MIASMA Modeling Framework for the Health Impact Assessment of Man-Induced Atmospheric Changes MIASA was written by Dr. Pim Martens (p.martens@icis.unimaas.nl). A fee of US$ 5 is required for a self-extracting CD Includes modules for thermal stress, malaria, dengue, and schistosomiasis Select IPCC scenario and GCM

Other Models CiMSiM and DENSim for dengue Weather and habitat-driven entomological simulation model that links with a simulation model of human population dynamics to project disease outbreaks http://daac.gsfc.nasa.gov/IDP/models/index.html

India’s Initial National Communication: Goals To identify, analyze, and evaluate the impacts of climate variability and change on natural ecosystems, socioeconomic systems, and human health To assess the vulnerabilities, which also depend on the institutional and financial capacities of the affected communities To assess the potential adaptation responses To develop technical, institutional, and financial strategies to reduce vulnerability

India’s Initial National Communication Temperature-related mortality Vector-borne diseases Changing patterns of diseases – malaria, filaria, kala-azar, Japanese encephalitis, dengue Health effects of extreme weather Diarrhea, cholera, and poisoning caused by biological and chemical contaminants in water Damaged public health infrastructure due to cyclones/floods Social and mental health stress due to disasters and displacements Health effects due to insecurity in food production

Malaria in India 1976-2001

Projected Changes in Number of Months Malaria Can Be Transmitted

Factors Affecting Malaria Distribution and Prevalence in India Climate Urban settlements Poverty Irrigation Agricultural practices Land-use change

Methods for Determining a Health Adaptation Baseline

Questions for Designing Adaptation Policies and Measures Adaptation to what? Is additional intervention needed? What are the future projections for the outcome? Who is vulnerable? On scale relevant for adaptation Who adapts? How does adaptation occur? When should interventions be implemented? How good or likely is the adaptation?

Current and Future Adaptation Options What is being done now to reduce the burden of disease? How effective are these policies and measures? What measures should begin to be implemented to increase the range of possible future interventions? When and where should new policies be implemented? Identify strengths and weaknesses, as well as threats and opportunities to implementation Kovats et al., 2003

Public Health Adaptation Existing risks Modifying existing prevention strategies Reinstitute effective prevention programs that have been neglected or abandoned Apply win/win or no-regrets strategies New risks

Options for Adaptations to Reduce the Health Impacts of Climate Change Health Outcome Legislative Technical Educational-advisory Cultural & Behavioral Thermal stress Building guidelines Housing, public buildings, urban planning, air conditioning Early warning systems Clothing, siesta Extreme weather events Planning laws, economic incentives for building Urban planning, storm shelters Use of storm shelters Vector-borne diseases Vector control, vaccination, impregnated bednets, sustainable surveillance, prevention & control programmes Health education Water storage practices Water-borne diseases Watershed protection laws, water quality regulation Screening for pathogens, improved water treatment & sanitation Boil water alerts Washing hands and other behavior, use of pit latrines Source: Chapter 9 (Human Health) in the IPCC Third Assessment Report. McMichael et al. 2001

Screening the Theoretical Range of Response Options – Malaria Ebi and Burton, submitted

Analysis of the Practical Range of Response Options – Malaria Ebi and Burton, submitted

Thank you