ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Cooking Stoves, Indoor Air Pollution and Respiratory Health in Rural Orissa

Indoor air pollution emitted from traditional fuels and cooking stoves is a potentially large health threat in rural regions. This paper reports the results of a survey of traditional stove ownership and health among 2,400 households in rural Orissa. We find a very high incidence of respiratory illness. About one-third of the adults and half of the children in the survey had experienced symptoms of respiratory illness in the 30 days preceding the survey, with 10 per cent of adults and 20 per cent of children experiencing a serious cough. We find a high correlation between using a traditional stove and having symptoms of respiratory illness. We cannot, however, rule out the possibility that the high level of observed respiratory illness is due to other factors that also contribute to a household's decision to use a traditional stove, such as poverty, health preferences and the bargaining power of women in the household.

SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200871Cooking Stoves, Indoor Air Pollution and Respiratory Health in Rural OrissaEsther Duflo, Michael Greenstone, Rema HannaIndoor air pollution emitted from traditional fuels and cooking stoves is a potentially large health threat in rural regions. This paper reports the results of a survey of traditional stove ownership and health among 2,400 households in rural Orissa. We find a very high incidence of respiratory illness. About one-third of the adults and half of the children in the survey had experienced symptoms of respiratory illness in the 30 days preceding the survey, with 10 per cent of adults and 20 per cent of children experiencing a serious cough. We find a high correlation between using a traditional stove and having symptoms of respiratory illness. We cannot, however, rule out the possibility that the high level of observed respiratory illness is due to other factors that also contribute to a household’s decision to use a traditional stove, such as poverty, health preferences and the bargaining power of women in the household. Esther Duflo (eduflo@mit.edu) is at the Massachusetts Institute of Technology and the Abdul Latif Jameel Poverty Action Lab. Michael Greenstone (mgreenst@mit.edu) is at MIT, J-PAL and the Brookings Institution. Rema Hanna (rema_hanna@ksg.harvard.edu) is now at the Kennedy School of Government, Harvard University and with J-PAL.Over 72 per cent of all households in India and 90 per cent of households in the country’s poorer, rural areas use traditional solid fuels, such as crop residue, cow-dung and firewood, to meet their cooking needs [Census of India 2001]. The burning of solid fuels indoors in open fires or traditional cooking stoves (‘chulhas’) results in high levels of toxic pollutants in the kitchen area. As such, the use of these fuels is considered a major risk factor for lung cancer as well as cardiovascular and respiratory disease [WHO 2002]. This paper reports on a new sur-vey – in a rural area of the state of Orissa – that was designed to increase understanding of what types of households use tradi-tional cooking fuels and to understand the correlations among fuel use, pollution levels and respiratory health.The available evidence suggests that the indoor air pollution (IAP) from biomass fuels and traditional cooking stoves may be a serious health threat, particularly to women and young children who spend a considerable amount of time near the cooking stove. The smoke from burning solid fuels produces many pollutants, including particulate matter (PM) and carbon monoxide (CO) that have been shown to be highly toxic in animal studies and associated with increased rates of infant mortality [Chay and Greenstone 2003a and 2003b].1 The emissions rates of pollutants from traditional stoves are extremely high. For example, Smith (2000) reports that mean 24-hour PM10 concentration from solid-fuel-using households in India sometimes exceeds 2,000 μg/m3, where PM10 refers to particulate matter with a diameter of less than or equal to 10μm; these particles are widely believed to pose the greatest health problems. By compa-rison, the United States (US) Environmental Protection Agency’s (EPA) standard for an acceptable annual 24-hour average of PM10 in theUS is 50 micrograms per cubic metre (μg/m3) [EPA 2006]. The reported concentrations in India also look ex-tremely high when compared to theEPA’s other PM10standard, which deems more than one exceedance per year of 150 μg/m3 in a 24-hour period to be unacceptable.What does this mean for public health? The World Health Report 2002 of the WHO estimates that the percentage of the national burden of disease due to solid fuel use is 3.5 per cent in India. Exposure to indoor air pollutants is a major risk factor for chronic obstructive pulmonary disease (COPD) in adults and acute lower respiratoryinfections(ARI) among young children [WHO 2002]. It is estimated that indoor air pollution accounts for one-third of ARI cases and that ARI accounts for up to 20 per cent of deaths among children under the age of five. This means that,aftercontaminated water, solid fuels – used by over half of
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly72theworld’spopulation – are the most important environmental causeofdisease [Bruce et al 2006]. Despite the importance of impact of IAP on the health of women and children, there is very little systematic evidence on the usage patterns of traditional cooking fuels and stoves in India. There is even less reliable field survey data on either the associations between traditional cooking fuels and stoves and respiratory health or the causal effects of cleaner stoves on health. This paper reports on our attempt to fill these important gaps in the literature. We analyse data from the largest field survey of stove use and respiratory health ever carried out in India.1 The Chulha Survey The survey, conducted between January and July 2006, is a joint collaboration by the non-governmental organisation (NGO) Gram Vikas and the Chennai-based Centre for Microfinance at the Institute for Financial Management and Research (CMF).2 It covered 2,357 households across 40 villages in the districts of Ganjam and Gajapati in Orissa. The sample was drawn from the villages in which Gram Vikas operates. In this respect, it is not strictly representative of all of India. However, to the extent that the households surveyed are rural and poor, it is representative of the typical users of traditional fuels in India. To our knowledge, this dataset is the largest and most comprehensive one available, combining a high-quality household socio-economic survey; a physical check-up; recall-based health data; reliable information on stoveavailability, stove use and fuel consumption; and economic productivity and well-being data for children and adults.The survey consisted of three modules, consecutively adminis-tered on each household visit. First, the CMF field officers admin-istered the comprehensive household survey, which covered household composition (size and members’ ages, sexes, and rela-tionships to the head), demographics (education levels, caste, religion), economics (income flows, indebtedness, wealth) and consumption (particularly related to health and fuel). The ques-tionson wealth and income mimic the National Sample Survey ofIndia – a nationally representative survey conducted every five years – allowing us to see where the sampled households fall in the income distribution of rural India. To assess household exposure to indoor air pollution, the survey also covered stove ownership by type, stove use, housing construction and fuel use.Next, CMF field officers interviewed each household member individually (for children aged 14 and below, the child’s primary care-giver completed the child survey). The interviews covered the individual’s exposure to stoves, including stove type, dura-tion of exposure and frequency. Questions included the number of meals the individual cooked that week, time spent cooking and time spent near the stove. Next, CMF field officers collected data on the self-reported health status of each survey respondent. Respondents were asked a series of recall questions on symptoms experienced in the last 30 days (coughs or flu, cough with blood, etc) and the severity of the symptoms. To further gauge respira-tory function, respondents were asked how difficult it was for them to perform common activities (work in the field for one day, walk 200 metres, walk five kilometres, draw water from a well, climb a small hill, lift or carry heavy objects such as a 5-kg bag, etc). Finally, to gauge the productivity effects of respiratory heath,CMF collected data on school attendance for children and employment status and time-use patterns for adults over the last 24 hours.After the survey, the field officers administered a short physical health examination in order to develop a richer portrait of the respondent’s general and respiratory health status. They obtained biometric data (height, weight and arm circumference) and then conducted two tests designed to gauge actual respiratory func-tioning and exposure to smoke. First, following the RESPIRE study in Guatemala [Smith et al 2006],CMF measuredCO in exhaled breath with a micro medical CO monitor. The presence of CO in breath is a biomarker of recent exposure to air pollution from biomass fuel combustion, making it a good proxy for individual exposure to smoke from cooking stoves in rural settings where there are few other sources, such as combustion in vehicle engines or industry. Second, the field officers conducted a spirometry test, which is designed to gauge respiratory health by measuring how much air the lungs can hold and how well the respiratory system can move air in and out of the lungs. The tests were conducted using guidelines from the American Association for Respiratory Care. The following measures were recorded: (1) FEV1, the forced expiratory volume in the first second; (2) FVC, the forced vital capacity; (3) PEF, the peak expiratory flow; (4)MMEF, the maximum midexpiratory flow; and (5) FEV1 per cent, the FEV1 expressed as a percentage of the total volume. Each individual was tested up to seven times, until at least two FEV1 readings were within 100mls (or 5 per cent) of each other. In contrast to peak flow tests, spirometry readings can be used to diagnose obstructive lung disorders (such asCOPD and asthma) and restrictive lung disorders. Further, they provide the only measurements of lung function that are comparable across individuals [Beers, Berkow et al 1999]. 2 Demographics and Household Stove UseTable 1 presents the basic statistics on demographics from the sample. Households in the study tend to be quite disadvantaged. Per capita consumption is roughly Rs 500 per month, with about one-third of households living on less than $1 per day (in per capita at purchasing power parity). Only half of the households have electricity and a similar fraction have a literate head of household. A large percentage of households come from the traditionally disadvantaged minority groups of India: about a quarter of house-holds belong to the sche-duled caste and a little over 10 per cent belong to the tribal communities. In Table 2 (p 73), we provide sample statistics on household stoveuse.Column 1 provides these statistics for the full sample, whilecolumns 2 and 3 provide these statistics for those above andbelow the median income in the sample (Rs 400), respec-tively.Households typically cook 14 meals a week (about two meals a day), regardless of their income status. To cook these meals, most households – about 94 per cent – use a traditional stove withbiomass fuel as their primary cooking stove. Almost always, Table 1: Household DemographicsMonthly per capita expenditure 495.71Head of household literate 51.9%Household has electricity 46.9%Scheduled caste 27.0%Scheduled tribe 12.6%
69 77 66 86 20 41
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly74once in the last week. We first present results on general health –namely, average height and body mass index (BMI). Height is often seen as an indicator of long-run health and nutritional sta-tus, whileBMI is an indicator of short-run health and nutritional status. Both indicators suggest that this study population tends to be in poor overall health. Males in the sample are roughly 160 cm in height (below the Indian average of 167). On average, males in the sample have a BMI of 19.62. A person withBMI less than 19 is clas-sified as malnourished; about half of the males in this sample fall into this category. The women in the sample tend to be shorter (about 151 cm) and also have alowBMI aver-age of approximately 19.The self-reported health data in-dicate a substantially high preva-lence of respiratory disease in both males and females. About one-third of all males had a cough in the last 30 days, with 9 per cent of all males (or about one-third of those who had a cough) stating that they had a cough that lasted more than two weeks. Forty-three per cent of all males had a cold or flu in the last 30 days, with 10 per cent of all males stating that they had a cold or flu that lasted more than two weeks. The incidence rates are virtually identical for women, re-gardless of whether or not they had participated in cooking ac-tivities in the last five days. These self-reported replies to the health questions reflect genuine health problems because house-holds report that they devote roughly 17 per cent of their con-sumption expenditures to healthcare. While this number seems high, a survey of a poor population in the state of Rajasthan also found a high ratio of health expenditures to overall expenditures [Banerjee, Deaton, Duflo 2004]. The activity index measures the number of activities the indi-vidual indicated he or she would be unable to do or unable to do without help (self-reported). Activities include: work in the field for one day, walk 200 metres, walk five kilometres, draw water from a well, climb a small hill, lift or carry heavy objects (for exam-ple, a 5-kg bag), routine housework such as cleaning or cooking, stand up from sitting on the floor, bow, squat or kneel, routine daily activities (dressing, bathing or using the toilet) and routine house-work such as cleaning or cooking. All responses were entirely self-reported – the respondents were not asked to demonstrate their ability or inability to perform them. As such, the activity index is a good measure of the respondent’s perception of his or her own health. This may be important since other household members’ perceptions of how healthy an individual is could be a factor in their allocation of household healthcare consumption and of housework – some of which increases exposure to pollutants. Women have a higher activity index than men, indicating that they report having more difficulties performing these basic phys-ical activities. Women who do not cook have the highest activity index, suggesting that women who report being indisposed to physical activity are given leave from household responsibilities. This is supported by the fact that women who do not cook tend to be older (on average, they tend to be 11 years older than women who cook) and tend to be the mother of the head of the household. Table 4 also shows the objective measure of CO exposure and respiratory health taken during the physical examination. The CO breath test used measures the concentration in parts per million (ppm) of CO in the lung alveolar – a good proxy for exposure toIAP. In the absence of high exposure to IAP, concentrations in the range of 0 to 6 ppm would indicate that the person is a non-smoker, while 7 to 10 ppm, 11 to 20 ppm, and greater than 20 ppm would indicate that the person is a light, regular or heavy smoker, respectively [Jarvis et al 1980; Jarvis et al 1986]. Men had higherCO breath test readings than the women: 52 per cent of men had CO readings above 6 and 33 per cent had readings above 10; for women these figures were 44 and 21 per cent, respectively. Since women spend more time cooking than men, we would expect the women to show higher smoke exposure. However, men’s higher exposure to smoke could be due to the higher prevalence of tobacco smoking among men; we could not, however, determine whether this was true, as few individuals in the survey admitted to smoking cigarettes or beedis. Among the women, those who do not cook have lower smoke exposure than those who do: 45 per cent of women who cook regularly had a CO reading greater than six, compared to 40 per cent of those who do not cook. This difference is significant at the 1 per cent level. While this could be due to the fact that women who do not cook spend less time around the stove, there could also be alternative explanations: women who do not cook may be less physically able (as indicated by the activity index) and there-fore, choose to smoke less. Finally, Table 4 shows the results of the spirometry test, which measure individuals’ respiratory function. The FEV1 is the volume of air expelled in the first second of a forced expiration starting from full inspiration and the FVC is the maximum volume of air in liters that can be forcibly and rapidly exhaled. The FEV1/FVC is the ratio of the two measures. The average FEV1/FVC was 83 per cent for men and 85 per cent for women. These values are similar to the 85.78 per cent reported in a recent study in Guatemala [Smith et al 2006].The results of the spirometer can be used to diagnose “obstruc-tive respiratory disease” and “moderate to severe restrictive respiratory disease” [Mannio et al 2003; Barreiro and Perillo 2004]. A lower FEV1/predicted FEV1 (typically below 80 per cent) combined with a lower FEV1/FVC (typically below 70-75 per cent) indicates a higher probability of an obstructive respiratory disorder.4 We Table 4: Health Status of Adults Females Cooked At Did Not Least Once Cook at in the Last All in the Males AllWeekLastWeek (1) (2) (3) (4)Height 161.97150.88151.28150.50BMI 19.62 19.39 19.12 20.23Cough in the last 30 days 0.33 0.31 0.31 0.32Cough last more than two weeks 0.09 0.09 0.08 0.10Cold or flu in the last 30 days 0.43 0.45 0.45 0.44Cold or flu last more than twoweeks 0.10 0.10 0.10 0.09Activity index 0.49 0.89 0.81 1.96% with CO reading above 6 ppm 0.52 0.44 0.45 0.40% with CO reading above 10 ppm 0.33 0.21 0.21 0.19FEV1/FVC 0.830.850.860.84FEV1/predicted FEV1 0.70 0.69 0.69 0.69Obstructive respiratory disease 0.12 0.10 0.09 0.12Moderate to severe restrictive respiratorydisease 0.14 0.14 0.13 0.19
SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200875find that about 12 per cent of males and 10 per cent of females can be classified as having obstructive respiratory disease. About 14 per cent of the population would be classified as hav-ing moderate to severe restrictive respiratory disease.5 Interest-ingly, though, women who did not cook in the last week have a higher probability of experiencing moderate to severe restrictive respiratory disease.3.2 Child HealthTable 5 reports on children aged 14 and under. It is evident that the incidence of respiratory symptoms is also high among them: 50 per cent of all children had a cough in the last 30 days and 40 per cent of these (or 20 per cent of the sample) had had a serious cough. These illnesses translated into high healthcare expendi-tures: about 40 per cent of the children saw a healthcare provider in the last 30 days. Children aged 10 to 14 years were also give theCO breath test. About 37 per cent of these children had CO read-ings greater than 6 ppm and 18 per cent had readings greater than 10 ppm. Chil-dren’s exposure toCO appeared to be lower than that of the adults but their CO readings were still extremely high for a group that includes very few smokers. 3.3 Household Stove Use and HealthTable 6 presents suggestive evidence that indicates that clean stove use is associated with lower CO exposure and better health for women (panel A) and children (panel B). The table presents the results of a linear probability model, with the dependent variables noted in the row headings. We report the coefficient and associ-ated standard error for an indicator variable for using a clean stove. We also report the mean of the dependent variable, the R-squared statistic and the number of observations. All regressions include controls for age, scheduled caste, number of household members, an indicator variable for land-ownership, household per capita expenditures, and gender (for the children). We additionally control for height and weight in col-umns 1, 2 and 3 when the outcome of interest is obtained from the spirometry or CO breath tests. The regressions for adults also include an indicator variable for whether the individual is a self-reported smoker. We find a strong negative correlation between using a clean stove and having a highCO reading. Using a clean stove is associ-ated with a 7 percentage-point decrease in the probability of having aCO reading greater than 6. Given that about 44 per cent of women have a CO reading over 6, this corresponds to a 15 per cent difference (column 1). The results are larger for children under 14, with an 8 percentage-point or 23 per cent difference, but this difference is not statistically significant. Using a clean stove as the primary stove reduces the probability that a child has a CO reading above 10 by 8 percentage points. Given that 18 per cent of kids have aCO reading over 10, this corresponds to a 45 per cent difference (column 1).There is little evidence that using a clean stove affects the probability of having an obstructive respiratory disease (column 3). However, we find that using a clean stove is associ-ated with a lower probability of having a cold or flu in the last 30 days (column 4), of having a cough in the last 30 days (column 5) and of having a serious cough in the last 30 days (column 6). As shownin column 7, there is not a statistically significant rela-tionshipbetween using a clean stove and the activity index.Of course, these correlations between stove use and CO expo-sure and health status do not indicate a causal effect of clean stoves: households that use a clean stove may also be healthier because they are richer (recall Table 2) or more health-conscious. Nevertheless, the findings in this section indicate that there is substantial scope for improvement in the respiratory health status of these households and that reducingCO exposure through the introduction of clean stoves might be an effective channel to achieve this improvement. 4 Conclusions and Future ResearchOur survey provides evidence on stove use and health in Orissa. We show that indoor air pollution is indeed a signifi-cant health threatinrural areas where households rely on traditional chulhas for their cooking needs. We find a high incidence of respiratory illness: about one-third of all adults and half of all children experi-enced symptoms of respira-tory illness in the 30 days prior to the survey, with 10 per cent of adults and 20 per cent of children experi-encing a serious cough. There is a strong correla-tion between using a stove with cleaner fuels and hav-ing better respiratory health, suggesting that the use of traditional stoves may indeed be a culprit behind these high levels of respiratory disease. However, because the choice of stove use is correlated with other factors that affect health (such as income levels and empowerment of women), we cannot fully Table 5: Health Status of ChildrenVariable MeanBMI 19.19Cough in the last 30 days 0.50Serious cough in the last 30 days 0.21% with healthcare visit in the last 30 days 0.42% with CO reading above 6 ppm 0.37% with CO reading above 10 ppm 0.18Table 6: Associations between Stove Use, CO Exposure and Health CO>6 CO>10 Obstructive Cold or Cough in Serious Activity Respiratory Flu in the the Last Cough Index Disease Last 30 Days 30 Days (1) (2) (3) (4) (5) (6) (7)A Women Clean -0.071 -0.06 -0.013 -0.085 -0.07 -0.032 -0.075 (0.036)* (0.027)** (0.022)(0.032)***(0.028)** (0.016)** (0.101) Mean 0.44 0.21 0.10 0.45 0.31 0.09 0.88 N 2859 2859 2673 3734 3733 3732 3736 R2 0.000.000.030.010.020.020.26B Children Clean -0.088 -0.082 -0.094 -0.044 (0.065) (0.044)* (0.030)*** (0.023)* Mean 0.37 0.18 0.5 0.21 N 713 713 4589 4588 R2 0.040.020.030.04(1) For women, a serious cough is defined as a cough lasting more than two weeks. For children, parents were asked whether the cough was “serious”. (2) Activity Index measures the number of activities the individual indicated he would be unable to do or unable to do without help. Activities include: work in the field for one day, walk 200 metres, walk five kilometres, draw water from a well, climb a small hill, lift or carry heavy objects (e g, a 5-kg bag), routine housework such as cleaning or cooking, stand up from sitting on the floor, bow, squat or kneel, and routine daily activities such as dressing, bathing or using the toilet.

Dear reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Comments

(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top