RAEL Home Mission About RAEL Lab Members Projects

Publications RAEL News Public Outreach Support RAEL Contact Us


Majid Ezzati
Ph.D. student

Science, Technology, and Environmental Policy (STEP) Program
Woodrow Wilson School of Public and International Affairs
Princeton University

Field Research Program on Energy Technology, Health, and the Environment: Integrated Risk and Decision Management


The rural environment of developing nations constitutes a particularly significant and appropriate context for research on technology transfer, environmental change, and public health.  In this setting, the diffusion and application of new energy technologies have directly observable local impacts on economic opportunity, human health, and the ecosystem, as well as on the global flux of greenhouse gases.  Despite such significance, the integration of energy technology management, land use, and ecology with human health in a common framework has few, if any, precedents; but it is essential if fundamental improvements in human development and conservation policy formulation are to be implemented.

The long-term objective of this research is to investigate the interrelationship between technology management and innovation, and the resulting impacts on economic development, public health, and ecological conservation and sustainability in a rural context.  The focus of  my dissertation research is to develop a new notion of "household ecology" that considers the links between human activity, human environment - within and surrounding the home, subsistence agricultural plots, cattle compounds, and fuelwood resources - and public health with emphasis on: the role of household energy technology (traditional and improved cookstoves) in shaping this ecosystem, the impacts of the latter on human health, and the local knowledge of these issues.

The research is organized around three themes: (i) Household choice of technology; (ii) Interaction between household activity and local environment (indoor and outdoor); (iii) Impacts of local environment on human health.


Historically few studies of public health in Africa have utilized the connection of health to the local environment, and especially how the whole system is influenced by the introduction of new technologies.  Examples of works that do take such an integrated approach include the classical study of John Ford (1971) on the interaction between human activity, vegetation, wildlife, cattle health, and the spread of sleeping sickness in East Africa, and Sir Malcolm Watson’s (1953) analysis of the relationship between economic activity, landscape change, insect density, and malaria in Central Africa.  Both works encourage control of disease by environmental management in a manner similar to early efforts elsewhere in the world to manage malaria through ecological manipulation and minimal use of pesticides (McArthur, 1947).

In contrast to such studies, today when addressing public health questions development policy forums commonly treat the relationships between technology and the environment, and between the environment and disease transmission in a surprisingly simplistic manner: In each case traditional technology, health, and environmental policy research has focused on understanding and controlling isolated sections of the natural environment in a mechanistic manner.  It is all too common to find malaria treated via a particular drug distribution program, only to experience a dramatic increase in drug resistant strains the next season. Similarly, malnutrition is often addressed by presenting a particular improved crop variety to a community that also lacks water resources for irrigation.  In three successive years, the World Bank issued its influential annual World Development Report titled Development and the Environment (1992), Investing in Health (1993), and Infrastructure for Development (1994) on the above-mentioned topics.  The three reports individually contain a number of important points, but could have been written by three different organizations: each stands in almost complete isolation from the lessons, problems, and policy options set forth in the other two.  In fact, the basic connections among these topics should be obvious: community health requires sanitation infrastructure; building roads and new settlements alters forests and wetlands; deforestation and erosion often result in reduced agricultural yields; and so forth.

In the specific case of household environment, consistently overlooked is the fact that the outdoor and indoor environments of rural communities are complex ecological-health systems which require integrated management.  The presence and density of various rodents and pests around and inside the house, for example, is influenced by the proximity of cattle, the storage of firewood or food, and smoke from burning of wood or dung (NRC, 1991).  In this complex ecological-economic system, it is not surprising that human activity - especially the introduction of new technology - and human health are tightly connected with the local environment.  The integrated management strategy undoubtedly calls not only for the bridging of the traditional scientific fields which deal with the various aspects of the human and natural environments but also for a new reservoir of knowledge which studies such domains as inherently connected phenomena.

Geographical Focus

Nowhere is the need for reliable, non-polluting and locally managed sources of energy more pressing than in Africa, where over 60% of national energy supplies, and 90% of household energy needs are met with biomass fuels (Bradley, 1991).  The dependence on biomass contributes to a range of economic, health, and environmental problems as human activity is embedded in, and interrelated with, a complex ecological system among the rural communities of African countries (Johnson and Anderson, 1988; Kjekshus, 1996).  Demand for wood impacts the natural ecosystem through fuel harvesting (often over one ton per person per year) and disruption of natural nutrient cycling.  Indoor air pollution from traditional cooking methods has serious health implications.  High levels of wood smoke exposure - often 10 to 20 times the recommended exposure limits - have been reported in studies from many developing nations (Kammen et al., 1999).  This, in turn, has been linked to acute respiratory infection (ARI), in particular pneumonia, as well as eye infections and burns (Ezzati et al., 1999).  ARI is the leading health hazard to children in developing nations, and in the general population results in an estimated 4.3 million deaths per year (Smith, 1993).

Research Objectives and Policy Applications

The specific objectives of this research and its policy applications are as follows:

Household choice of energy technology and lessons for technology transfer: The design, dissemination, and management of new technologies is fundamental to the objectives of economic and human advancement in developing nations, and to our ability to conserve the environment.  Renewable energy technologies -  such as photovoltaics, large and small-scale wind energy technologies, biomass systems - are deemed as particularly important in the latter context.  At the same time little predictive theoretical or empirical understanding has emerged from the various approaches to "technology transfer studies" in the context of industrialized or developing nations (Kammen, 1995).  Such an understanding, however, is vital to any efforts that wish generalize and apply the results of specific pilot projects to a wide range of development planning.  We studied the various factors that influence the household choice of technology along the following themes: economic factors versus cultural forces; scale of decision making: community level, household level, and intra-household (in particular gender issues).

The work has been carried out in collaboration with Ekero Jiko Sales, a community based development group, to introduce locally designed and manufactured ceramic cookstoves. Most improved (high efficiency and low emission) cookstoves use a ceramic liner to retain heat, as opposed to traditional 3-stone cookers (see below) which have few insulating properties. Improved cookstoves are designed for both of the common fuels used in Kenyan society: firewood and charcoal.

Traditional 3 stone cooker

The impacts of energy technology on indoor air pollution: We studied the impact of the various forms of household energy technology on the level of indoor air pollution resulting from smoke emission, and any improvements that may come about as an outcome of the introduction of improved (high-efficiency-low-emission) ceramic cookstoves. . Unlike numerous studies that have focused on the efficiency and emissions of improved cookstoves in laboratory environment only, emissions under the conditions of operation by actual users was considered by conducting day-long (14-hour) monitoring of pollution levels at different points inside the house. Mean daily suspended particulate levels of 1000 - 5000 ((mg/m3) were found to be common among those households which use firewood, with maximum levels as high as 200,000 ((mg/m3). During this period wood consumption was also monitored as were the activities of all the members of the house, especially those related to cooking. The first set of results from this analysis indicate that improved stoves (both wood and charcoal) reduce the average pollution level but there is still a large region of overlap between improved stoves and open fire.

Cooking smoke emanating from the roof of a house

Indoor air pollution and health: A central focus of this research has been to establish the quantitative relationship between the level of various toxicants in smoke, in particular suspended particulates and carbon monoxide, and respiratory and eye infections. This "dose-response" relationship is fundamental to a detailed assessment of the cost-benefit trade off expected from various technological interventions, a framework which has been largely lacking from current approaches to managing technology transfer and training programs.

Research Site

The research is based at the 55,000 acre Mpala Research Centre/Ranch in Laikipia District, Kenya, which is administered by Princeton University, the Smithsonian Institution, the Kenya Wildlife Service, and the National Museums of Kenya.  Activities at the research centre have a high profile with not only the local community groups but also the large land holders in Laikipia and the national government.

Mpala Research Centre and Mpala ranch together comprise two villages which house the staff of the research centre and the ranch, and more than five bomas – compounds which house pastoralist households who take care of cattle in traditional homesteads.  The large degree of population stability at Mpala provides a rare condition for the collection of panel data (Deaton, 1995).  Further, the large matrix of houses and occupations at Mpala provides an unmatched opportunity for comparative research.

Research Methodology

We employed a variety of qualitative and quantitative research methods in my study.  A summary of data collection schemes for establishing the specific health-environment interaction are as follows:

Indoor air pollution and exposure to smoke: Emissions measurements will take place with commercially available PersonalDataRam (for Total Suspended Particulate (TSP)) and ENEREC combustion analyzer (for carbon monoxide and carbon dioxide).  We conduct day-long monitoring of pollution level at different points inside the house for a representative number of households based on fuel type, stove type, and characteristics of the house in each air pollution category.  During this period we also monitor wood consumption and the activities of all the members of the house especially as related to cooking.  The air pollution and time-activity information together provide a profile of the person’s exposure to smoke during the day (exposure is dependent on how much smoke there is in the house and how much time the individual spends indoors).

Health data: In collaboration with Dr. A.W. Muriithi, from the Department of Pediatrics and Acute Respiratory (ARI) Program at the University of Nairobi Medical School, we have set up a data collection scheme to monitor household health with emphasis on acute respiratory infection (ARI), eye infection, and skin disease.  We have provided training for two community nurses from Nanyuki District Hospital on the World Health Organization protocols for diagnosing ARI.  They visit the communities at Mpala on weekly basis to collect clinical health records and provide basic medicine to the local communities.


Summit Foundation, Compton Foundation, ASAL Development Programme Laikipia District, SSRC, CIS, CRS.



Back to the Top

Selected Abstracts

Ezzati, M., Mbinda, B.M., and Kammen, D.M. (2000) "Comparison of Emissions and Residential Exposure from Traditional and Improved Cookstoves in Kenya," Environmental Science and Technology (ES&T), 34 (2), p. 578-583.

Tables and Figures (PDF).

Suspended particulate matter and carbon emissions from combustion of biomass, in addition to their environmental consequences, have been causally associated with the incidence of respiratory and eye infections.  Improved stoves offer the potential for emission reduction.  We compare the emissions of suspended particulate matter and carbon monoxide from traditional and improved biofuel stoves in Kenya under the actual conditions of household use.  Data for analysis is from 137 14-hour days of continuous real-time emission concentration monitoring in a total of 38 households.  Our analysis shows that improved (ceramic) wood-burning stoves reduce daily average suspended particulate matter concentration by 48% (1822 mg/m3; 95% C.I. 663 - 2982) during the active burning period and by 77% (1034 mg/m3; 95% C.I. 466 - 1346) during the smouldering phase.  Ceramic stoves also reduce the median and the 75th and 95th percentiles of daily emission concentration during burning period and the 95th percentile during the smouldering phase, and therefore shift the overall emission profile downwards.  Improved charcoal-burning stoves also offer reductions in indoor air pollution compared to the traditional metal stove, but these are not statistically significant.  The greatest reduction in emission concentration is achieved as a result of transition from wood to charcoal where mean emission concentrations drop by 87% (3035 mg/m3; 95% C.I. 2356 - 3500) during burning period and by 92% (1121 mg/m3; 95% C.I. 626 - 1216) when smouldering, as well as large reductions in the median and 75th and 95th percentiles.  These results indicate that transition to charcoal, followed by the use improved wood stoves, are viable options for reduction of human exposure to indoor air pollution in many developing nations.

Keywords: Household energy technology, biomass combustion, indoor air pollution, suspended particulate matter, improved cookstoves, field study.

Ezzati, M., Saleh, H., and Kammen, D. M. (2000) "The Critical Role of Microenvironments in Pollutant Exposure and Impact: Individual Behaviour and Indoor Air Pollution from Biomass Combustion in Kenya," Environmental Health Perspectives, submitted.

Acute and chronic respiratory infections as a result of exposure to indoor particulate matter are the leading cause of morbidity and mortality in developing countries.  Efforts to develop effective intervention strategies and detailed quantification of the exposure-response relationship for indoor suspended particulate matter require accurate estimates of exposure.  We use continuous monitoring of indoor air pollution and individual time-activity budget data to construct detailed profiles of exposure for 342 individuals in 55 households in rural Kenya.  Data for analysis is from 210 14-hour days of continuous real-time monitoring of PM10 concentration as well as the location and activities of household members, supplemented by data on spatial dispersion of pollution and interviews.  Young and adult women are found not only to have the highest absolute exposure to particulate matter from biomass combustion (4016 mg/m3 and 7616 mg/m3 daily average exposure concentrations respectively), but also the largest exposure relative to that of males in the same age group (2.7 and 6.3 times respectively).  Exposure during brief high-intensity emission "episodes" accounts for 38% to 63% of the total exposure of household members who take part in cooking and 0% - 15% for those who do not.  Simple models that neglect the spatial distribution of pollution within the home, intense emission episodes, and activity patterns underestimate exposure by 35% - 80% for different demographic sub-groups, resulting in inaccurate and biased estimations.  Health and intervention impact studies should therefore consider the critical role of exposure patterns, including the short periods of intense emission, in detail to avoid spurious assessments of risks and benefits.

Keywords: Biomass Combustion, Indoor Air Pollution, Suspended Particulate Matter, Exposure Assessment, Field Study, Public Health, Household Energy, Africa.

Ezzati, M., Singer, B.H., and Kammen D. M. (1999) "Towards an Integrated Framework for Development-Environment Policy: The Dynamics of Environmental Kuznets Curves", Submitted to Ecological Economics.

Environmental Kuznets curves have become the focus of increasing attention in assessing the relationship between economic growth and environmental change.  But little is known on how the interaction between ecological and economic factors may result in an EKC or other qualitatively different outcomes.  We introduce a restricted class of vector valued differential equations for representing the evolution of the socioeconomic and environmental phenomena which influence or are influenced by the process of economic development.  Our model includes not only the complete path of each economic or environmental variable over time but also the highly critical interactions among multiple variables.  We demonstrate analytically and numerically that among a multiplicity of possible outcomes, an inverted-U pattern can be obtained only under very specific circumstances.  Further, such an outcome requires attention to the multiple factors which form the economic-environmental system, rather than a single dominant one.  This dynamic and multivariable formulation provides a tool for discriminating among various proposed development plans on the basis of their outcome, and for retrospective analysis of project impact and payoff.

Keywords: Development policy analysis and forecasting, environmental change, economic growth, environmental Kuznets curves, differential equations, system dynamics.

Ezzati, M., D. M. Kammen, and B. H. Singer (1999) "The Health Impacts of Exposure to Indoor Air Pollution from Biofuel Stoves in Rural Kenya," The Proceedings of Indoor Air 99: the 8th International Conference on Indoor Air Quality and Climate; Edinburgh, Scotland; August 1999, 3, 130-135.

We use individual exposure profiles and time-series health data to study the relationship between exposure to suspended particulate matter and the incidence of acute respiratory infection (ARI), eye infection, and headache.  For adults, we find significantly higher incidence of ARI (p < 0.01), eye infection (p < 0.05), and headache (p < 0.01) among females.  Illness incidence rate, especially among adults, is significantly related to average exposure; a 1000 mg/m3 increase in average TSP concentration would result in a 3% increase in the number of illness cases among adults and 4% increase among children.  This relationship disappears when we account for the gender of the individual.  This connection is attributed to cooking duties, performed entirely by women who are exposed to high concentrations within short periods of time.  This intensity of exposure is likely to be as important as average exposure  in determining health risk associated with indoor air pollution.

Kammen, D. M., M. Ezzati, and B.M. Mbinda (1999) "The Determinants of Exposure to Indoor Air Pollution from Biofuel Stoves in Rural Kenya," The Proceedings of Indoor Air 99: the 8th International Conference on Indoor Air Quality and Climate; Edinburgh, Scotland; August 1999, 3, 171-176.

We integrate individual time-activity budget data and continuous monitoring of indoor air quality to construct personal profiles for exposure to suspended particulates resulting from biofuel combustion in an array of stove-fuel combinations in rural Kenya.  The analysis shows that on average the improved (ceramic) wood stoves reduce the level of suspended particulate emission by approximately 50% from that of traditional open fire (p =  0.07).  For individuals over 5 years of age, average personal exposure is correlated with the gender of the household member, with women having higher exposure than men (p = 0.01).  We attribute this to the extra time spent inside the house by female members of the household (p = 0.01).  This result suggests the need for a multiplicity of strategies for reduction of exposure to indoor air pollution, based on both reduction of emissions from stoves and modifications to the structure of the cooking area or the time spent near stove.

Saleh, H., M. Ezzati, D. M. Kammen, and B.H. Singer (1999) "The Health Impacts of Exposure to Indoor Air Pollution from Biofuel Stoves in Rural Kenya: A Behaviour Based Analysis," Submitted to Environmental Science and Policy.

Suspended particulate matter from combustion of biomass has been causally associated with the incidence of respiratory and eye infections.  We integrate individual time-activity budget data and continuous monitoring of indoor air pollution to construct personal profiles for exposure to suspended particulates.  This analysis shows that pollution level varies considerably (by more than an order of magnitude) within even small rural homes.  Exposure is therefore highly influenced by the location and activities of household members, themselves determined by gender and age. We use the individual exposure profiles and time-series health data to study the relationship between exposure to suspended particulate matter and the incidence of acute respiratory infection (ARI), eye infection, and headache.  Combining physiological and statistical evidence, we infer that there is a logarithmic relationship between exposure to suspended particulate matter and the probability of diagnosis with ARI, eye infection, or headache (p-values in the range 0.006 - 0.15 for different ailments).  Each logarithmic increase in exposure results in an additional 1%, 1.2%, and 3% probability of being diagnosed with eye infection, ARI, and one of these three illnesses respectively.  The implication of a logarithmic relationship is that to achieve significant improvements in health, exposure (and hence pollution levels) must decrease considerably.

Keywords: Indoor Air Pollution, Suspended Particulate Matter, Household Energy Technology, Biomass Combustion, Acute Respiratory Infection, Rural Public Health.
Back to the Top

Last updated 4/2/2000

RAEL Home |  Mission |  About RAEL |  Lab Members  
Projects |  Publications |  RAEL News  
Public Outreach |  Support RAEL |  Contact Us