Estimating the adaptation needs in the health sector is challenging. Most of the health outcomes that are projected to be affected by climate change are current problems; there will not be death certificates, hospital admissions, or records of visits to health care providers indicating that a particular event was due to climate change. Instead, as with some other environmental exposures (particularly indoor and outdoor air quality), models are used to estimate the proportion of a disease burden that can be attributed to climate change based on exposure-response relationships and projected changes in weather patterns. Uncertainties in models, from limited data through to inadequate specification of factors that influence the exposure-response relationship, will therefore lead to uncertainty as to the precise magnitude of the climate change impact.
The analysis makes a number of necessary, but unlikely assumptions, including that the number of annual cases of diarrhoeal disease, malaria, and malnutrition, and the cost of treatment would remain constant. Population growth is projected to increase under the medium variant from 6.1 billion in 2000 to 8.3 billion in 2030 . Conducting a sensitivity analysis that incorporated these population increases would require assumptions of future incidence rates of these health outcomes, based on assumptions of socioeconomic development, including improvements in health care delivery, the rate of deployment of current interventions, and the development of more effective technologies. Using the current number of cases in the analysis in effect assumes that incidence will decrease as population increases, without attribution of the possible reasons for such a decline. If disease rates remain constant until 2030, then the number of cases due to climate change would increase.
Because of the large uncertainties, the costs estimated should be viewed only as indicators of the relative magnitude of health adaptation costs. Countries improve their public health and health care systems as they develop, which should decrease the burden of many climate-sensitive diseases. Costs of current treatments tend to decrease over time, although development of new, more effective treatments may cost more. However, there is an underlying assumption that currently developing countries will develop along similar pathways to those followed by the developed countries. There is ample evidence to suggest that the reality may be much more challenging. A key issue is water; most developing countries do not have as much available water as developed countries did when they were developing. Therefore, it will be more difficult to resolve issues such as access to safe water and sanitation. Also, malaria is much more difficult to control in Africa than it was in Europe and the US.
Another complexity is estimating the economic cost of injuries, illnesses, and deaths across multiple countries and regions. Issues include not just how to value a human life, but how to measure economically the life-course consequences of malnutrition, for example. Mortality is a commonly used metric, but is an inadequate measure of the affect of a health outcome on the family and on society; a death at age 80 and a death at age 2 would be counted equally while having different impacts. Similarly, malnutrition decreases learning ability, thus affecting lifelong earning potential, among a myriad of other impacts. Therefore, counting cases of disease also is insufficient for estimating total impacts.
Additional research could reduce some of the uncertainties in the analysis. The literature base underlying the exposure-response relationships is fairly thin; additional estimates in more regions would increase confidence in projected relative risks and would allow estimates of future climate change-attributable cases on smaller spatial scales. Additional research also is needed to better project how population growth, socioeconomic development, and other factors would likely influence future rates of climate-sensitive health determinants and outcomes. Development of a health model would facilitate both projections and identification where additional information would reduce uncertainty . Linking such a model with integrated assessment models would take advantage of the their efforts to model population growth and economic development.
Bosello et al.  estimated the economic impacts of climate change in 2050 on temperature-related illnesses, diarrhoeal diseases, malaria, dengue fever, and schistosomiasis. Changes in morbidity and mortality were interpreted as changes in labour productivity and demand for health care. There was a mixed pattern of increases and decreases in GDP, welfare, and investment across world regions, with benefits estimated in high-income countries and losses primarily in low-income countries. The results showed that direct cost estimates, such as the present analysis, underestimate the full health costs (and benefits) of climate change.
Because of the uncertainties in the estimated costs, they should be taken as indicators of the size of the financial needs and not as accurate predictions. The estimates are likely to include both under- and over-estimates of the actual costs. Emerging technologies, along with significant investments in research and development, are likely to reduce current health burdens over the next 20+ years. On the other hand, the estimated costs were for only three of the health outcomes projected to increase with climate change; and then only a fraction of the burden of malnutrition was included. According to Caulfied et al. , the estimated prevalence of weight-for-age less than -2 SD (a measure of malnutrition) are 18% for Asia and the Pacific; 6% for Eastern Europe and Central Asia, and for Latin America and the Caribbean; 21% for the Middle East and North Africa; 46% in South Asia; 32% in Sub-Saharan Africa; and 2% in high-income countries. In addition, the model used to estimate malnutrition does not take into account new projections that a few degree increase in global mean temperature may render some areas unsuitable for rainfed agriculture; if this occurs, the short-term health consequences would likely be severe.
The costs estimated for adaptation are consistent with other estimates of financial needs for health care investment. Stenberg et al.  estimated the costs to scale-up essential child health interventions to reduce by two-thirds child mortality under the four MDGs aimed at children's health by 2015 in 75 countries; the countries chosen accounted for 94% of death among children less than five years of age. The interventions focused on malnutrition, pneumonia, diarrhoea, malaria, and key newborn causes of death. Calculations were bottom-up, based on intervention, country, and year. Costs included program-specific investments needed at national and district levels. The authors estimated that an additional US$ 52.4 billion would be required for the period 2006–2015. Projected costs in 2015 were equivalent to increasing the average total health expenditures from all financial resources in the 75 countries by 8% and raising general government health expenditure by 26% over 2002 levels. The authors noted that countries with weak health care systems may experience difficulties mobilizing enough domestic public funds.
Kiszewski et al.  estimated that US$ 38 to 45 billion would be required from 2006 to 2015 to scale up current malaria control programs to reach international goals, or about US$ 3.8 to 4.5 billion annually. If resources were to be made available and malaria goals were achieved, then the numbers of climate change-related cases of malaria in 2030 would likely to significantly lower, thus requiring fewer additional resources for treatment than the estimated US$ 4 – 12 billion under the 750 ppm CO2 scenario.
Although current governmental health expenditures can be anticipated to increase with development, there are health problems other than those associated with climate change that need to be addressed, such as HIV/AIDS, tuberculosis, diabetes, and other diseases. Assuming that Ministries of Health, NGOs, and other actors will completely cover the additional costs related to climate change is not realistic for many low-income countries; to do so would mean that other health issues of importance are left wanting. Financial and policy arrangements will need to be altered to address the projected additional cases of diarrhoeal diseases, malnutrition, and malaria.