- Email: [email protected]
Monetary burden of health impacts of air pollution in Mumbai, India: Implications for public health policy
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
available at www.sciencedirect.com
Public Health journal homepage: www.elsevier.com/puhe
Original Research
Monetary burden of health impacts of air pollution in Mumbai, India: Implications for public health policy A.M. Patankar a,*, P.L. Trivedi b a b
K.J. Somaiya Institute of Management Studies and Research, Vidyavihar (East), Mumbai 400077, India Department of Humanities and Social Sciences, Indian Institute of Technology, Bombay, India
article info
summary
Article history:
Objectives: Mumbai, a mega city with a population of more than 12 million, is experiencing
Received 20 January 2010
acute air pollution due to commercial activity, a boom in construction and vehicular traffic.
Received in revised form
This study was undertaken to investigate the link between air pollution and health impacts
20 October 2010
for Mumbai, and estimate the monetary burden of these impacts.
Accepted 10 November 2010
Study design: Cross-sectional data were subjected to logistic regression to analyse the link between air pollution and health impacts, and the cost of illness approach was used to measure the monetary burden of these impacts.
Keywords:
Methods: Data collected by the Environmental Pollution Research Centre at King Edward
Air pollution
Memorial Hospital in Mumbai were analysed using logistic regression to investigate the
Morbidity
link between air pollution and morbidity impacts. The monetary burden of morbidity was
Logistic regression
estimated through the cost of illness approach. For this purpose, information on treatment
Concentrationeresponse
costs and foregone earnings due to illness was obtained through the household survey and
coefficients
interviews with medical practitioners.
Cost of illness
Results: Particulate matter (PM10) and nitrogen dioxide (NO2) emerged as the critical pollutants for a range of health impacts, including symptoms such as cough, breathlessness, wheezing and cold, and illnesses such as allergic rhinitis and chronic obstructive pulmonary disease (COPD). This study developed the concentrationeresponse coefficients for these health impacts. The total monetary burden of these impacts, including personal burden, government expenditure and societal cost, is estimated at 4522.96 million Indian Rupees (INR) or US$ 113.08 million for a 50-mg/m3 increase in PM10, and INR 8723.59 million or US$ 218.10 million for a similar increase in NO2. Conclusions: The estimated monetary burden of health impacts associated with air pollution in Mumbai mainly comprises out-of-pocket expenses of city residents. These expenses form a sizable proportion of the annual income of individuals, particularly those belonging to poor households. These findings have implications for public health policy, particularly accessibility and affordability of health care for poor households in Mumbai. The study provides a rationale for strengthening the public health services in the city to make them more accessible to poor households, especially those living in the slums of Mumbai. ª 2010 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ91 98205 01438; fax: þ91 22 25157219. E-mail address: [email protected] (A.M. Patankar). 0033-3506/$ e see front matter ª 2010 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2010.11.009
158
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Introduction Urban air pollution is a matter of concern among policy makers, organizations, donor agencies and people across the world due to its adverse consequences for human health and the environment. The dynamics of economic growth, industrialization and urbanization coupled with population growth are primarily responsible for air pollution. In urban areas, air pollution is a serious problem due to the concentration of people, both as perpetrators and victims of pollution. Worldwide, it is estimated that urban air pollution causes 800,000 deaths annually and the loss of 4.6 million life-years, twothirds of which occur in the developing countries of Asia alone.1 Air pollutants such as particulate matter, sulphur oxides, nitrogen oxides, ozone and carbon monoxide are responsible for a range of acute and chronic morbid effects and premature mortality in urban areas. Pollutants such as greenhouse gases (e.g. carbon dioxide, methane, nitrous oxide and chlorofluorocarbons) have transboundary impacts such as ozone depletion and global warming which affect the health and livelihood of millions of people, particularly in the poor and developing countries. Air pollution is, thus, a global ‘public bad’ that imposes massive economic and social costs at local, regional and global levels. In developing countries, millions of urban residents are exposed to acute levels of air pollution.2 The emerging economies of China, India and the Latin American countries are in a transition phase with increased life expectancy, changes in lifestyles and influx of population into urban areas. Cities such as Mumbai, Delhi, Beijing, Shanghai, Manila, Bangkok, Mexico City and Santiago are facing acute air pollution due to industrial activity, population growth, construction booms for housing and infrastructure, increased vehicular traffic, congested streets, poorly maintained vehicles, limited access to clean fuel and lack of effective control programmes.1,3,4 Mumbai is one of the largest mega cities in the world with a population of more than 12 million. The city has a sizable industrial, commercial and trading base, and plays host to pharmaceutical, thermal power and fertilizer companies, along with oil refineries and engineering, electronic and electrical businesses. The average per-capita income in the city is 69,000 Indian Rupees (INR) or US$ 1725 (2004e2005 prices),c which is three times that of the national per-capita income.5 In recent years, industries have closed down or moved out of Mumbai, making way for large residential complexes, thus leading to rapid expansion of suburbs and changing the land use pattern. Service sector activities, particularly information-technology-enabled services, have also increased sizably, contributing to more vehicular traffic and a boom in construction. Given the economic pre-eminence of Mumbai, the fact that more than 50% of the city’s residents still live in slums and face a life stricken by poverty, malnutrition and unhygienic surroundings cannot be ignored. The authors felt the need to examine the health consequences of air pollution for Mumbai from the point of view of planning for better public
c
Conversion to US$ done at the exchange rate 1 US$ ¼ INR 40.
health, higher productivity and better quality of life for the residents. This empirical study was undertaken to examine the relationship between air pollution and respiratory and cardiovascular symptoms and diseases for the residents of Mumbai. The study developed concentrationeresponse (CR) coefficientsd based on locally obtained data, and estimated the monetary costs of the health impacts caused by increases in air pollution in the city. The findings of this study have wider policy implications for pollution control programmes and provision of public healthcare facilities in Mumbai.
Methods Epidemiological studies conducted in different parts of the world have demonstrated a wide range of acute and chronic health impacts due to air pollution.4 Some notable examples are studies conducted in Jakarta,6 Santiago,7,8 Mexico City,9 Shanghai,10 Taiwan11e13 and China.14 Similar studies have also been carried out in India in Mumbai,15e18 Delhi,19e21 Hyderabad22 and Chandigarh.23 The studies conducted in Mumbai have either examined the health effects of air pollution on the basis of CR functions extrapolated from other studies, or have investigated the prevalence of respiratory morbidity but not its causal link with air pollution. The main objective of this empirical study was to examine the relationship between air pollution and morbidity in Mumbai by developing CR functions and estimating the monetary burden of health effects. This study focused on morbidity effects alone and not premature mortality related to air pollution, since morbidity is felt at regular and short-term changes in air pollution, unlike mortality which is a rare and extreme occurrence. Similarly, this study was only concerned with the human health effects of air pollution. Welfare effects (e.g. effects on property and vegetation) and ecological effects were not included in the scope of the study.
Data sources A dataset provided by the Environmental Pollution Research Centre (EPRC), attached to the Department of Chest Medicine at King Edward Memorial (KEM) Hospital in Mumbai, was used to examine the relationship between air pollution and respiratory and cardiovascular morbidity in Mumbai. The EPRC team carried out a survey in six areas of Mumbai (Worli, Borivali, Khar, Bhandup, Andheri and Maravali) between February 2003 and April 2004 to create a representative environmental health profile of the residents. More than 1800 individuals residing within 1 km of the air quality monitoring stations located in the six areas were administered the questionnaire. The survey gathered information on socio-economic profile, occupation, travel pattern, and d
CR function is the relationship between air pollution and human health effects determined through epidemiological studies. It empirically explains the variations in the number of cases of illness or death in the population for changes in the ambient concentrations of air pollutants and other explanatory factors such as age, gender, diet, smoking habit, occupation, etc.
159
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
proxies for indoor and outdoor exposure to air pollution. The health condition of the respondents and the prevalence of symptoms and illnesses were diagnosed or confirmed by the medical team during the survey. A subset of 1542 respondents from the EPRC survey was selected for this study. This included individuals above 11 years of age who were administered the entire questionnaire by the EPRC team. The health effects diagnosed in the survey were symptoms (e.g. cough, phlegm, breathlessness, wheezing and cold) and diseases [e.g. bronchitis, asthma, upper respiratory tract infections (URTI), cardiac, other chest illnesses, allergic rhinitis and chronic obstructive pulmonary disease (COPD)]. Diagnosis of acute symptoms was based on the health condition prevailing at the time of the survey, as well as the reported frequency of the symptoms during the year prior to the survey. Air pollutants included in the analysis as explanatory variables were particulate matter less than 10 mm in size (PM10), sulpur dioxide (SO2), nitrogen dioxide (NO2) and ammonia (NH3). These pollutants are regularly monitored by the roof-top monitors in the six areas (selected in the EPRC survey) by the Air Pollution Monitoring and Research Laboratory (APMRL) of the Municipal Corporation of Greater Mumbai, and the monthly averages are reported.e The APMRL records the levels of suspended particulate matter (SPM) in Mumbai. PM10 was calculated as 57% of SPM in Mumbai, as the literature on air pollution indicates the range of PM10 in SPM varies from 45% to 70%.6 The pollution records were taken from the APMRL to obtain the levels of pollutants for the analysis, as indicated in Table 1, based on the months in which the EPRC survey was carried out. Other explanatory variables selected from the EPRC data and used as controlling factors in all the variants of the regression model were age, gender, smoking habit, distance travelled to place of work, occupation, width of road adjacent to residence, presence of polluting industry near residence, hours spent in kitchen, quality of kitchen ventilation and type of the cooking fuel used. For estimating the monetary costs of the health impacts related to air pollution, a survey of 150 households was carried out by the researchers in five areas of Mumbai (Borivali, Andheri, Dadar, Chembur and Colaba) to obtain information on the cost of treatment and lost wages due to restricted and
e
In Mumbai, ambient air pollution levels fluctuate depending on seasonal and climatic factors. Generally, air pollution levels are low during the monsoon season (JuneeSeptember) and are quite high during winter (DecembereJanuary). Levels of pollutants vary significantly between areas, with Borivali recording the lowest levels of pollutants and Maravali recording the highest levels of pollutants. The difference between minimum and maximum monthly averages of pollutants within areas is also quite large during the year. For instance, the SPM levels in Maravali range from 150 mg/m3 to 650 mg/m3. In Borivali, the fluctuations in monthly averages are from 55 mg/m3 to 288 mg/m3. NH3 is not a criteria pollutant, but it is monitored in Mumbai due to the high levels experienced in parts of the city due to proximity to a fertilizer factory. There are large fluctuations in NH3 levels from 33 mg/m3 to 1200 mg/m3 in Maravali during the year.
Table 1 e Average monthly concentration of pollutants during survey period (in mg/m3). Pollutant
Name of ward Worli Borivali Khar Bhandup Andheri Maravali
SPM PM10a SO2 NO2 NH3
269 153 39 51 29
55 31 9 24 28
427 243 23 83 43
342 195 31 64 48
293 167 22 55 48
579 330 39 127 149
SPM, suspended particulate matter; PM10, particulate matter less than 10 mm in size. Source: APMRL, MCGM (2008). a PM10 is calculated as 57% of SPM based on the average range found in the literature (45e70%).
total inactivity days related to the considered symptoms and diseases.f Similar information was also obtained from doctors in public and private hospitals, and general practitioners in the city. Information obtained from households and medical practitioners was used to calculate the total monetary costs of health impacts associated with changes in air pollution levels in Mumbai.
Research methods The cross-sectional survey design used by the EPRC team influenced the choice of regression model employed in this study. Logistic regression is generally used when the dependent variable is binary or dichotomous. The regression model was used to test the null hypothesis which proposed no significant relationship between air pollution and respiratory and cardiovascular morbidity in Mumbai. The dependent variables in the model were symptoms and diseases. All the regression model variants used the controlling factors (i.e. age, gender, smoking habit, etc.) as outlined earlier. The data allowed the authors to test the incidence or occurrence of a particular symptom or disease against air pollution levels across different areas. The regression output yielded log odds ratios, which were used to predict changes in the odds of occurrence of a symptom or disease due to changes in air pollution levels. The coefficients corresponding to the significant odds ratios were used for further analysis as the CR coefficients in the study. Similar methodology can be found in cross-sectional studies conducted in Beijing,24 Guangzhou, Wuhan and Lanzhou,25 Benxi,26 Taiwan12,13 and four cities in China.27 Using the CR coefficients, changes in the health effects for given changes in air pollution concentrations were estimated using the US Environmental Protection Agency methodology.6 f
The areas selected for this survey differ from the areas originally targeted by the EPRC survey. Since household income is one of the most important factors in Mumbai to determine the choice of healthcare service (public vs private), areas were targeted depending on the baseline information available about the socioeconomic status of households. For instance, households belonging to lower income strata were targeted in areas of Chembur and Colaba. Households in Dadar and Borivali were predominantly middle-income families and those targeted in Andheri belonged to higher income strata.
160
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Monetary costs of the health effects were estimated using the cost of illness approach, which comprises the cost of treatment for illnesses and lost earnings due to missed work days or restricted activity days. Similar methodology has been used for estimating the monetary burden of health effects in studies conducted in Mexico City9,28 and Santiago.29 For this study, the household survey was carried out by the researchers in five areas of Mumbai to obtain information on the cost of treatment, including cost of medicines, equipment and consultation charges of doctors, as well as restricted and total inactivity days due to illness. The households were asked to give details about the type of treatment, service provider, cost of treatment, restricted activity days, total inactivity days, and loss of wages or salary per day for each symptom and disease listed in the questionnaire. Furthermore, a survey questionnaire was administered to the medical practitioners engaged in public and private health facilities, as well as general practitioners in the city. This survey obtained information on the outpatient and inpatient costs of treatment in public and private healthcare facilities, type of treatment, days of hospitalization required for symptoms and illnesses, days of restricted and total inactivity, and average number of missed work days for each symptom and disease listed in the questionnaire. The proportion of people using public and private healthcare services, the average wages of such individuals, and the symptoms and diseases requiring hospitalization were based on information obtained from the EPRC data and previous studies. A summary of the range of estimates obtained for symptoms and diseases is given in Table 2.
Results The cross-sectional survey design used by the EPRC team influenced the choice of regression model employed in this
study. The dependent variables, presence (or absence) of cardiovascular and respiratory symptoms and diseases, were used in the logistic regression framework to examine their relationship with the explanatory variables. In order to examine the link between air pollutants and morbidity outcomes (symptoms and diseases), the odds ratios for the four pollutants (PM10, NO2, SO2 and NH3) were obtained using multipollutant variants of the regression model. PM10 and NO2 were not considered together in the same variant of the model due to their high correlation. However, PM10 and NO2 were considered separately in different variants of the regression model as their sources in Mumbai are different. PM10 mainly has domestic or area sources, including construction activity, and NO2 is produced by vehicular and industrial emissions.8 The results of the regression model were obtained and interpreted on the basis of the significance tests used in the empirical literature in this research arena.30 The regression output showed statistically significant relationships between air pollutants and specific health outcomes, thus invalidating the null hypothesis. The regression results yielded the log odds ratios, which were used to predict changes in the odds of occurrence of a symptom or disease due to changes in air pollution levels. The coefficients corresponding to the significant odds ratios were the CR coefficients used in the study to estimate changes in the health effects for given changes in air pollution concentrations, similar to the methodology used in studies conducted in Taiwan.12,13
Estimation of health effects This study found a significant relationship between PM10 and NO2 with symptoms such as cough, breathlessness, wheezing and cold, and illnesses such as allergic rhinitis and COPD. SO2 was found to be significantly associated with cardiac ailments and other chest illnesses, and NH3 was related to phlegm and
Table 2 e Range of estimates obtained for symptoms and diseases. Nature of illness Symptoms Cough Phlegm Breathlessness Wheezing Cold Diseases Bronchitis Asthma URTI Allergic rhinitis COPD
Cost of treatment Daily cost of treatment in Days required hospital (in Rs)a in hospital (non-hospital) a (in Rs)
Days of restricted activity
Days of total inactivity
Frequency per person per yeard
200e700 200e700 300e700 300e700 100
e e 1000e3000 1000e3000 e
e e 4 4 e
2 2 2 2 1
1 1 2e4b 2e4b e
3 2 1e2c 1e2c 8
250e1000 250e1000 100e1000 100e1000
1000e3000 1000e3000 e e
6 6 e e
8 8 2 2
6 6 1 1
1e2c 1e2c 2 2
1000e5000
5000e10000
7
10
5e7b
1e2c
URTI, upper respiratory tract infection; COPD, chronic obstructive pulmonary disease. Data obtained from household survey and doctors. a Lower values in the range are the costs of treatment in public OPDs and public hospitals, and higher values are the costs of consulting specialist physicians or private hospitals. b Lower value shows total inactivity days without hospitalization and upper value shows inactivity days with hospitalization. c Lower values show the frequency of hospitalization per year and upper values show the frequency of outpatient treatment per year. d Frequency indicates the number of times a person suffers from the symptom or disease per year.
161
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Table 3 e Concentrationeresponse (CR) coefficients and incidence of health effects for Mumbai. Health effect Symptoms Cough Phlegm Breathlessness Wheezing Cold Diseases Cardiac Other chest illnesses Allergic rhinitis COPD
Pollutant
CR coefficient (%)a,b
Change in health effect (%)c
Estimated incidence per 1 lakh populationd
Estimated incidence for Mumbaie
PM10 NO2 NH3 PM10 NO2 PM10 NO2 PM10 NO2
0.007 0.021 0.012 0.009 0.028 0.006 0.020 0.006 0.018
0.35 1.05 1.20 0.45 1.40 0.30 1.00 0.30 0.90
350 1050 1200 450 1400 300 1000 300 900
42,000 126,000 144,000 54,000 168,000 36,000 120,000 36,000 108,000
SO2 SO2 NH3 PM10 NO2 PM10 NO2
0.118 0.162 0.024 0.014 0.046 0.014 0.023
2.36 3.24 2.40 0.70 2.30 0.70 1.15
2360 3240 2400 700 2300 700 1150
283,200 388,800 288,000 84,000 276,000 84,000 138,000
COPD, chronic obstructive pulmonary disease; PM10, particulate matter less than 10 mm in size. a The coefficient represents percentage change in the prevalence rate of morbidity outcome per mg/m3 change in pollution concentration. b Coefficients obtained in the regression model controlling for factors such as age, gender, smoking habit, distance travelled to place of work, occupation, width of road adjacent to the residence, presence of polluting industry near residence, hours spent in kitchen, quality of kitchen ventilation and type of cooking fuel used. c Incidence of health effects estimated for 50-mg/m3 increase in PM10, 50-mg/m3 increase in NO2, 20-mg/m3 increase in SO2 and 100-mg/m3 increase in NH3. d 1 lakh ¼ 100,000. e Incidence for Mumbai city calculated for the population of 12 million as per 2001 Census.
other chest illnesses. The CR coefficients corresponding to significant odds ratios were used to estimate the incidence of health effects for given changes in air pollution concentrations,g as shown in Table 3.
Estimation of monetary burden The figures shown in Table 2 were used to estimate the monetary burden of the health outcomes. Certain assumptions were made for estimating the health costs. For instance, based on the EPRC data and empirical findings,31e33 it was assumed that 25% of city residents use public health facilities and have a daily wage of approximately INR 200. For the remaining city residents, the daily wage was assumed to be approximately INR 500 and they were assumed to use private health facilities. Based on the information gathered from medical practitioners, it was assumed that 20% of cases of breathlessness and wheezing and 50% of cases of COPD would require hospitalization. Table 4 shows estimates of the incidence and total monetary costs of illness for Mumbai. The monetary estimates shown in Table 4 indicate out-ofpocket expenses of individuals, which include the cost of g The incidence of health effects was calculated based on the US Environmental Protection Agency methodology using the following equation: DHjp ¼ bjp POPj DAp, where DHjp is the change in health effect j due to pollutant p, bjp is the CR coefficient for health effect j related to pollutant p, POPj is population at risk of health effect j, and DAp is change in the concentration of pollutant p.
treatment and foregone earnings (lost wages) due to illness. Table 5 shows the out-of-pocket expenses on treatment of health outcomes as a proportion of annual personal income. In order to determine the total monetary burden of illness caused by air pollution, personal expenses towards cost of treatment and foregone earnings, government expenditure incurred in public healthcare facilities, and societal costs due to loss of productivity were considered. The total burden of illness estimated here was, thus, the sum of personal costs, government or public expenditure, and societal costs as seen in Table 6. Government or public expenditure on health care was only relevant for cases treated in public dispensaries or hospitals, and hence was considered likewise. The cost to society is the loss of productivity, which was assumed to be equivalent to the lost wages on account of restricted activity and total inactivity days due to illness. The monetary burden of health impacts is likely to increase in future as the cost of treatment and wages increase. Therefore, the authors attempted to estimate the future trends in monetary burden, assuming an increase in the prices of medicines and consultation charges by 9% per annum, and an increase in wages by 5% per annum based on Consumer Price Index trends.34 Government expenditure in public healthcare facilities has remained more or less stagnant in percentage terms over the years in Maharashtra state and Mumbai, and per-capita expenditure has actually decreased.32 The authors assumed that this trend will continue in future. They further assumed that Mumbai’s population will increase at a rate of 1.87% per annum based on past growth rates.35 Given these assumptions, future trends in
162
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Table 4 e Incidence and total cost of illness for Mumbai (in million Indian Rupees). Nature of illness Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD
Pollutant
Estimated incidence in Mumbaia
People using public health care
Total cost in public health care
People using private health care
Total cost in private health care
Estimated total burden for Mumbai
PM10 NO2 NH3 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2
42,000 12,600 144,000 54,000 168,000 36,000 120,000 36,000 108,000 84,000 276,000 84,000 138,000
10,500 31,500 36,000 13,500 42,000 9000 30,000 9000 27,000 21,000 69,000 21,000 34,500
6.30 18.90 21.60 23.22 72.24 15.48 51.60 1.80 5.40 10.50 34.50 424.20 696.90
31,500 94,500 108,000 40,500 126,000 27,000 90,000 27,000 81,000 63,000 207,000 63,000 103,500
53.55 160.65 183.60 188.73 587.16 125.82 419.40 9.45 28.35 126.00 414.00 2709.00 4450.00
59.85 179.55 205.20 211.95 659.40 141.30 471.00 11.25 33.75 136.50 448.50 3133.20 5147.40
COPD, chronic obstructive pulmonary disease; PM10, particulate matter less than 10 mm in size. a Incidence for Mumbai city calculated for the population of 12 million as per 2001 Census, and incidence estimated for 50-mg/m3 increase in PM10, 50-mg/m3 increase in NO2 and 100-mg/m3 increase in NH3.
the monetary burden of illness due to air pollution were estimated as depicted in Fig. 1.
as allergic rhinitis and COPD. The results also indicate that NO2 has relatively greater health impacts in terms of incidence and costs. The major source of NO2 in Mumbai is dieseland gasoline-based vehicles. There has been a substantial increase in the number of vehicles in Mumbai over the last decade. With expanding suburbs, rapid development in neighbouring districts, growing affluence and availability of affordable personal vehicles, this number is going to increase manifold in the near future. Thus, programmes and initiatives are urgently required to control the NO2 levels. Substantial economic costs are associated with the increase in incidence of these health outcomes with increases
Discussion The findings of this study have revealed a significant relationship between air pollution and morbidity outcomes for Mumbai. Among the four pollutants considered in the study, PM10 and NO2 emerged as the most significant pollutants and have a significant relationship with symptoms such as cough, breathlessness, wheezing and cold, and diseases such
Table 5 e Out-of-pocket expenses as proportion of annual personal income. Nature of illness
Cost of illness without hospitalization Cost of illness (in INR)
Public healthcare facilities Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD Private healthcare facilities Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD
Cost of illness with hospitalization
Frequency per yeara
Cost as % of annual incomeb
Cost of illness (in INR)
Frequency per person per year
600 600 900 900 200 500 3000
3 2 2 2 8 2 2
3.0 2.0 3.0 3.0 2.7 1.7 10.0
e e 5000 5000 e e 37400
e e 1 1 e e 1
62.3
1700 1700 2200 2200 350 2000 10,000
3 2 2 2 8 2 2
3.4 2.3 2.9 2.9 1.9 2.7 13.3
e e 14500 14500 e e 76000
e e 1 1 e e 1
e e 9.7 9.7 e e 50.7
COPD, chronic obstructive pulmonary disease; INR, Indian Rupees. a Frequency based on the information obtained from doctors. b Annual income of those in public health care is INR 60,000 and those in private health care is INR 150,000.
Cost as % of annual income
8.3 8.3
163
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Table 6 e Total burden of illness for Mumbai (in million Indian Rupees). Nature of illness Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD
Pollutant
Incidence for Mumbaia
Total burden in public health services
Total burden in private health service
Total burden of disease for Mumbai
Total burden in million US$b
PM10 NO2 NH3 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2
42,000 12,600 144,000 54,000 168,000 36,000 120,000 36,000 108,000 84,000 276,000 84,000 138,000
11.55 34.65 39.60 34.70 107.94 23.13 77.10 3.60 10.80 21.00 69.00 476.18 782.29
85.05 255.15 291.60 269.73 839.16 179.82 599.40 16.20 48.60 189.00 621.00 3213.00 5278.00
96.60 289.80 331.20 304.43 947.10 202.95 676.50 19.80 59.40 210.00 690.00 3689.18 6060.79
2.42 7.25 8.28 7.61 23.68 5.07 16.91 0.50 1.49 5.25 17.25 92.23 151.52
COPD, chronic obstructive pulmonary disease; PM10, particulate matter less than 10 mm in size. a Total incidence of health effects is estimated for the 12 million population of Mumbai as per 2001 Census figures for 50-mg/m3 increase in PM10, 50-mg/m3 increase in NO2 and 100-mg/m3 increase in NH3. b 1 US$ ¼ 40 Indian Rupees.
in PM10 and NO2. If personal costs, governmental costs and societal costs are included, the total monetary burden escalates further. However, it should be noted that the costs calculated in this study are for morbidity effects alone. If the mortality effects of the pollutants are also taken into account, the monetary burden is likely to increase substantially. Hence, the costs estimated here provide conservative lower bound estimates of the costs of air pollution for Mumbai city, and the actual costs are expected to be much larger. Most of the monetary burden comprises out-of-pocket expenses borne by the suffering individuals and families. As the estimates show, these expenses represent a significant proportion of the annual income of people, particularly for those belonging to poor households. There are, thus, tangible costs of health impacts which can no longer be ignored by policy makers and citizens. The huge burden on individuals also reflects the inadequacy of the public health facilities available in the city. The findings of this study call for strengthening of the public health infrastructure to cater for the health needs of those who cannot afford expensive treatment.
The personal costs of health impacts are mainly paid through individuals’ incomes. There is no system of private health insurance or social insurance which can adequately protect individuals from such costs. Currently, the coverage of the health insurance schemes is negligible. There is no social insurance scheme in place in the entire country. Thus, the health consequences are entirely faced by individuals and families themselves, sometimes through the sale of assets or loans to fund the medical expenses. Efforts are urgently needed to put in place a system of health insurance which eases the burden of disease on people. A significant issue emerging from this study is the stagnant health expenditure of the Government and its consequences for the population. There is an acute need to strengthen the public healthcare system in Mumbai in terms of accessibility, affordability and quality in order to help poor households. This assumes more relevance since half of Mumbai’s population lives in slums with greater exposure to all types of pollution, and a life stricken by poverty, malnutrition and unhygienic living conditions.
18000 16000 14000
Million INR
12000 10000
8000 6000 4000 2000 0 2008
2009
2010
Burden due to PM10
2011
2012
Burden due to NO2
2013
2014
2015
Total burden of illness
Fig. 1 e Future trends in burden of illness for Mumbai city. INR, Indian Rupees; PM10, particulate matter less than 10 mm in size.
164
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Ethical approval None sought.
Funding None declared.
Competing interests None declared.
Acknowledgments The authors wish to thank Dr. Amita Athavale, Head of the Department of Chest Medicine and EPRC at KEM Hospital, Mumbai for providing the dataset, without which this study would not have been feasible.
references
1. UNDP/World Bank. Health impacts of outdoor air pollution. Urban air pollution. South Asia Urban Air Quality Management Briefing Note No. 11. UNDP/World Bank ESMAP; 2003. 2. WDI. World development indicators. Washington, DC: World Bank; 2006. 3. Romieu I, Hernandez-Avila M. Air pollution and health in developing countries: a review of epidemiological evidence. In: McGranahan G, Murray F, editors. Air pollution and health in rapidly developing countries. Earthscan; 2003. 4. HEI. Health effects of outdoor air pollution in developing countries of Asia: a literature review. Special Report 15. HEI International Scientific Oversight Committee. Boston, MA: Health Effects Institute; 2004. 5. Economic survey 2006e07. Available at. Government of Maharashtra, http://www.maharashtra.gov.in/; 2006 (last accessed 15/08/2008). 6. Ostro BD. Estimating the health effects of air pollution: a methodology with application to Jakarta. Washington, DC: World Bank; 1994. Policy Research Working Paper 1301. 7. Ostro BD, Sanchez JM, Aranda C, Eskeland GS. Air pollution and mortality: results from Santiago, Chile. Policy Research Working Paper No. 1453. Washington, DC: World Bank; 1995. 8. Ostro BD, Eskeland GS, Feyzioglu T, Sanchez JM. Air pollution and health effects: a study of respiratory illness among children in Santiago, Chile. Washington, DC: World Bank; 1998. 9. Mexico Air Quality Management Team. Improving air quality in metropolitan Mexico city: an economic valuation. Policy Research Working Paper No. 2785. Washington, DC: World Bank; 2002. 10. Li J, Guttikunda SK, Carmichael GR, Streets DG, Chang YS, Fung V. Quantifying the human health benefits of curbing air pollution in Shanghai. J Environ Manag 2003;70:49e62. 11. Alberini A, Krupnick A. Air pollution and acute respiratory illness: evidence from Taiwan and Los Angeles. Am J Agric Econ 1997;79:1620e4. 12. Chen PC, Lai YM, Wang JD, Yang CY, Hwang JS, Kuo HW, et al. Adverse effect of air pollution on respiratory health of primary school children in Taiwan. Env Health Persp 1998;106:331e5. 13. Hwang BF, Jaakkola JK, Lee YL, Lin YC, Guo YL. Relation between air pollution and allergic rhinitis in Taiwanese school children. Respir Res 2006;7:23.
14. Bank World. Cost of pollution in China: economic estimates of physical damages. Washington, DC: World Bank; 2007. 15. Kamat SR, Godkhindi KD, Shah BW, Mehta AK, Varsha N, Gregrat J, et al. Prospective 3 year study of health morbidity in relation to air pollution in Bombay, India. Bombay Air Pollution e Health Study. MCGM, KEM Hospital; 1980. 16. Kamat SR, Patil JD, Gregrat J, Dalal N, Deshpande JM, Hardikar P. Air pollution related respiratory morbidity in central and north-eastern Bombay. J Assoc Physic India 1992; 40:588e93. 17. World Bank Technical Paper No. 381. In: Shah J, Nagpal T, editors. Urban air quality management strategy in Asia: Greater Mumbai report. Washington, DC: World Bank; 1997. 18. Joseph AE, Sawant AD, Srivastava A. PM10 and its impacts on health e a case study in Mumbai. Int J Environ Health Res 2003; 13:207e14. 19. Cropper ML, Simon NB, Alberini A, Sharma PK. The health effects of air pollution in Delhi, India. Policy Research Working Paper 1860. Washington, DC: World Bank; 1997. 20. Chhabra SK, Chhabra P, Rajpal S, Gupta RK. Ambient air pollution and chronic respiratory morbidity in Delhi. Arch Environ Health 2001;56:58e64. 21. Agarwal KS, Mughal MZ, Upadhyay P, Berry JL, Mawer EB, Puliyel JM. The impact of atmospheric pollution on vitamin d status of infants and toddlers in Delhi, India. Arch Dis Child 2002;87:111e3. 22. Kumar KS, Prasad CE, Balakrishna N, Rao KV, Reddy PUM. Respiratory symptoms and spirometric observations in relation to atmospheric pollutants in a sample of urban population. Asia Pacif J Publ Health 2000;12:58e64. 23. Gupta D, Boffetta P, Gaborieau V, Jindal SK. Risk factors of lung cancer in Chandigarh, India. Ind J Med Res 2001;113: 142e50. 24. Xu X, Wang L. associations of indoor and outdoor particulate level with chronic respiratory illness. Am Rev Respir Dis 1993; 148:1516e22. 25. Jhang J, Hu W, Wei F, Wu G, Chang W-L, Chapman R. Long-term changes in air pollution and health implications in four Chinese cities. Energy Sustainable Dev 2005;IX:67e76. 26. Jin LB, Qin Y, Xu Z. Relationship between air pollution and acute and chronic respiratory disease in Benxi, China. Chin J Environ Sci 2000;17:268e70. 27. Zhang J, Song H, Tong S, Li L, Liu B, Wang L. Ambient sulfate concentration and chronic disease mortality in Beijing. Sci Total Environ 2000;262:63e71. 28. Bell M, Davis D, Gouveia N, Borja-Aburto V, Cifuentes L. The avoidable health effects of air pollution in three Latin American cities: Santiago, Sao Paulo and Mexico City. Environ Res 2006;100:431e40. 29. Maharashtra Pollution Control Board. Report on environmental status of Mumbai region. Maharashtra Pollution Control Board. Government of Maharashtra; 2005. 30. Hosmer D, Lemeshow S. Applied logistic regression. In: Wiley Series in Probability and Mathematical Statistics. New York: Wiley Interscience; 1989. 31. NSSO. Morbidity and treatment of ailments. Report No. 44. NSSO; 1998. 32. Government of Maharashtra. Maharashtra human development report 2002. Government of Maharashtra; 2002. 33. World Bank. Urban poverty and transport: the case study of Mumbai. Washington, DC: World Bank; 2003. 34. CSO. Ministry of statistics and programme implementation. Available at. Government of India, http://www.mospi.nic.in/; 2008 (last accessed 20/08/2008). 35. World Gazetteer. India: largest cities and towns and statistics of their populations. World Gazetteer; 2008.
available at www.sciencedirect.com
Public Health journal homepage: www.elsevier.com/puhe
Original Research
Monetary burden of health impacts of air pollution in Mumbai, India: Implications for public health policy A.M. Patankar a,*, P.L. Trivedi b a b
K.J. Somaiya Institute of Management Studies and Research, Vidyavihar (East), Mumbai 400077, India Department of Humanities and Social Sciences, Indian Institute of Technology, Bombay, India
article info
summary
Article history:
Objectives: Mumbai, a mega city with a population of more than 12 million, is experiencing
Received 20 January 2010
acute air pollution due to commercial activity, a boom in construction and vehicular traffic.
Received in revised form
This study was undertaken to investigate the link between air pollution and health impacts
20 October 2010
for Mumbai, and estimate the monetary burden of these impacts.
Accepted 10 November 2010
Study design: Cross-sectional data were subjected to logistic regression to analyse the link between air pollution and health impacts, and the cost of illness approach was used to measure the monetary burden of these impacts.
Keywords:
Methods: Data collected by the Environmental Pollution Research Centre at King Edward
Air pollution
Memorial Hospital in Mumbai were analysed using logistic regression to investigate the
Morbidity
link between air pollution and morbidity impacts. The monetary burden of morbidity was
Logistic regression
estimated through the cost of illness approach. For this purpose, information on treatment
Concentrationeresponse
costs and foregone earnings due to illness was obtained through the household survey and
coefficients
interviews with medical practitioners.
Cost of illness
Results: Particulate matter (PM10) and nitrogen dioxide (NO2) emerged as the critical pollutants for a range of health impacts, including symptoms such as cough, breathlessness, wheezing and cold, and illnesses such as allergic rhinitis and chronic obstructive pulmonary disease (COPD). This study developed the concentrationeresponse coefficients for these health impacts. The total monetary burden of these impacts, including personal burden, government expenditure and societal cost, is estimated at 4522.96 million Indian Rupees (INR) or US$ 113.08 million for a 50-mg/m3 increase in PM10, and INR 8723.59 million or US$ 218.10 million for a similar increase in NO2. Conclusions: The estimated monetary burden of health impacts associated with air pollution in Mumbai mainly comprises out-of-pocket expenses of city residents. These expenses form a sizable proportion of the annual income of individuals, particularly those belonging to poor households. These findings have implications for public health policy, particularly accessibility and affordability of health care for poor households in Mumbai. The study provides a rationale for strengthening the public health services in the city to make them more accessible to poor households, especially those living in the slums of Mumbai. ª 2010 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ91 98205 01438; fax: þ91 22 25157219. E-mail address: [email protected] (A.M. Patankar). 0033-3506/$ e see front matter ª 2010 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2010.11.009
158
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Introduction Urban air pollution is a matter of concern among policy makers, organizations, donor agencies and people across the world due to its adverse consequences for human health and the environment. The dynamics of economic growth, industrialization and urbanization coupled with population growth are primarily responsible for air pollution. In urban areas, air pollution is a serious problem due to the concentration of people, both as perpetrators and victims of pollution. Worldwide, it is estimated that urban air pollution causes 800,000 deaths annually and the loss of 4.6 million life-years, twothirds of which occur in the developing countries of Asia alone.1 Air pollutants such as particulate matter, sulphur oxides, nitrogen oxides, ozone and carbon monoxide are responsible for a range of acute and chronic morbid effects and premature mortality in urban areas. Pollutants such as greenhouse gases (e.g. carbon dioxide, methane, nitrous oxide and chlorofluorocarbons) have transboundary impacts such as ozone depletion and global warming which affect the health and livelihood of millions of people, particularly in the poor and developing countries. Air pollution is, thus, a global ‘public bad’ that imposes massive economic and social costs at local, regional and global levels. In developing countries, millions of urban residents are exposed to acute levels of air pollution.2 The emerging economies of China, India and the Latin American countries are in a transition phase with increased life expectancy, changes in lifestyles and influx of population into urban areas. Cities such as Mumbai, Delhi, Beijing, Shanghai, Manila, Bangkok, Mexico City and Santiago are facing acute air pollution due to industrial activity, population growth, construction booms for housing and infrastructure, increased vehicular traffic, congested streets, poorly maintained vehicles, limited access to clean fuel and lack of effective control programmes.1,3,4 Mumbai is one of the largest mega cities in the world with a population of more than 12 million. The city has a sizable industrial, commercial and trading base, and plays host to pharmaceutical, thermal power and fertilizer companies, along with oil refineries and engineering, electronic and electrical businesses. The average per-capita income in the city is 69,000 Indian Rupees (INR) or US$ 1725 (2004e2005 prices),c which is three times that of the national per-capita income.5 In recent years, industries have closed down or moved out of Mumbai, making way for large residential complexes, thus leading to rapid expansion of suburbs and changing the land use pattern. Service sector activities, particularly information-technology-enabled services, have also increased sizably, contributing to more vehicular traffic and a boom in construction. Given the economic pre-eminence of Mumbai, the fact that more than 50% of the city’s residents still live in slums and face a life stricken by poverty, malnutrition and unhygienic surroundings cannot be ignored. The authors felt the need to examine the health consequences of air pollution for Mumbai from the point of view of planning for better public
c
Conversion to US$ done at the exchange rate 1 US$ ¼ INR 40.
health, higher productivity and better quality of life for the residents. This empirical study was undertaken to examine the relationship between air pollution and respiratory and cardiovascular symptoms and diseases for the residents of Mumbai. The study developed concentrationeresponse (CR) coefficientsd based on locally obtained data, and estimated the monetary costs of the health impacts caused by increases in air pollution in the city. The findings of this study have wider policy implications for pollution control programmes and provision of public healthcare facilities in Mumbai.
Methods Epidemiological studies conducted in different parts of the world have demonstrated a wide range of acute and chronic health impacts due to air pollution.4 Some notable examples are studies conducted in Jakarta,6 Santiago,7,8 Mexico City,9 Shanghai,10 Taiwan11e13 and China.14 Similar studies have also been carried out in India in Mumbai,15e18 Delhi,19e21 Hyderabad22 and Chandigarh.23 The studies conducted in Mumbai have either examined the health effects of air pollution on the basis of CR functions extrapolated from other studies, or have investigated the prevalence of respiratory morbidity but not its causal link with air pollution. The main objective of this empirical study was to examine the relationship between air pollution and morbidity in Mumbai by developing CR functions and estimating the monetary burden of health effects. This study focused on morbidity effects alone and not premature mortality related to air pollution, since morbidity is felt at regular and short-term changes in air pollution, unlike mortality which is a rare and extreme occurrence. Similarly, this study was only concerned with the human health effects of air pollution. Welfare effects (e.g. effects on property and vegetation) and ecological effects were not included in the scope of the study.
Data sources A dataset provided by the Environmental Pollution Research Centre (EPRC), attached to the Department of Chest Medicine at King Edward Memorial (KEM) Hospital in Mumbai, was used to examine the relationship between air pollution and respiratory and cardiovascular morbidity in Mumbai. The EPRC team carried out a survey in six areas of Mumbai (Worli, Borivali, Khar, Bhandup, Andheri and Maravali) between February 2003 and April 2004 to create a representative environmental health profile of the residents. More than 1800 individuals residing within 1 km of the air quality monitoring stations located in the six areas were administered the questionnaire. The survey gathered information on socio-economic profile, occupation, travel pattern, and d
CR function is the relationship between air pollution and human health effects determined through epidemiological studies. It empirically explains the variations in the number of cases of illness or death in the population for changes in the ambient concentrations of air pollutants and other explanatory factors such as age, gender, diet, smoking habit, occupation, etc.
159
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
proxies for indoor and outdoor exposure to air pollution. The health condition of the respondents and the prevalence of symptoms and illnesses were diagnosed or confirmed by the medical team during the survey. A subset of 1542 respondents from the EPRC survey was selected for this study. This included individuals above 11 years of age who were administered the entire questionnaire by the EPRC team. The health effects diagnosed in the survey were symptoms (e.g. cough, phlegm, breathlessness, wheezing and cold) and diseases [e.g. bronchitis, asthma, upper respiratory tract infections (URTI), cardiac, other chest illnesses, allergic rhinitis and chronic obstructive pulmonary disease (COPD)]. Diagnosis of acute symptoms was based on the health condition prevailing at the time of the survey, as well as the reported frequency of the symptoms during the year prior to the survey. Air pollutants included in the analysis as explanatory variables were particulate matter less than 10 mm in size (PM10), sulpur dioxide (SO2), nitrogen dioxide (NO2) and ammonia (NH3). These pollutants are regularly monitored by the roof-top monitors in the six areas (selected in the EPRC survey) by the Air Pollution Monitoring and Research Laboratory (APMRL) of the Municipal Corporation of Greater Mumbai, and the monthly averages are reported.e The APMRL records the levels of suspended particulate matter (SPM) in Mumbai. PM10 was calculated as 57% of SPM in Mumbai, as the literature on air pollution indicates the range of PM10 in SPM varies from 45% to 70%.6 The pollution records were taken from the APMRL to obtain the levels of pollutants for the analysis, as indicated in Table 1, based on the months in which the EPRC survey was carried out. Other explanatory variables selected from the EPRC data and used as controlling factors in all the variants of the regression model were age, gender, smoking habit, distance travelled to place of work, occupation, width of road adjacent to residence, presence of polluting industry near residence, hours spent in kitchen, quality of kitchen ventilation and type of the cooking fuel used. For estimating the monetary costs of the health impacts related to air pollution, a survey of 150 households was carried out by the researchers in five areas of Mumbai (Borivali, Andheri, Dadar, Chembur and Colaba) to obtain information on the cost of treatment and lost wages due to restricted and
e
In Mumbai, ambient air pollution levels fluctuate depending on seasonal and climatic factors. Generally, air pollution levels are low during the monsoon season (JuneeSeptember) and are quite high during winter (DecembereJanuary). Levels of pollutants vary significantly between areas, with Borivali recording the lowest levels of pollutants and Maravali recording the highest levels of pollutants. The difference between minimum and maximum monthly averages of pollutants within areas is also quite large during the year. For instance, the SPM levels in Maravali range from 150 mg/m3 to 650 mg/m3. In Borivali, the fluctuations in monthly averages are from 55 mg/m3 to 288 mg/m3. NH3 is not a criteria pollutant, but it is monitored in Mumbai due to the high levels experienced in parts of the city due to proximity to a fertilizer factory. There are large fluctuations in NH3 levels from 33 mg/m3 to 1200 mg/m3 in Maravali during the year.
Table 1 e Average monthly concentration of pollutants during survey period (in mg/m3). Pollutant
Name of ward Worli Borivali Khar Bhandup Andheri Maravali
SPM PM10a SO2 NO2 NH3
269 153 39 51 29
55 31 9 24 28
427 243 23 83 43
342 195 31 64 48
293 167 22 55 48
579 330 39 127 149
SPM, suspended particulate matter; PM10, particulate matter less than 10 mm in size. Source: APMRL, MCGM (2008). a PM10 is calculated as 57% of SPM based on the average range found in the literature (45e70%).
total inactivity days related to the considered symptoms and diseases.f Similar information was also obtained from doctors in public and private hospitals, and general practitioners in the city. Information obtained from households and medical practitioners was used to calculate the total monetary costs of health impacts associated with changes in air pollution levels in Mumbai.
Research methods The cross-sectional survey design used by the EPRC team influenced the choice of regression model employed in this study. Logistic regression is generally used when the dependent variable is binary or dichotomous. The regression model was used to test the null hypothesis which proposed no significant relationship between air pollution and respiratory and cardiovascular morbidity in Mumbai. The dependent variables in the model were symptoms and diseases. All the regression model variants used the controlling factors (i.e. age, gender, smoking habit, etc.) as outlined earlier. The data allowed the authors to test the incidence or occurrence of a particular symptom or disease against air pollution levels across different areas. The regression output yielded log odds ratios, which were used to predict changes in the odds of occurrence of a symptom or disease due to changes in air pollution levels. The coefficients corresponding to the significant odds ratios were used for further analysis as the CR coefficients in the study. Similar methodology can be found in cross-sectional studies conducted in Beijing,24 Guangzhou, Wuhan and Lanzhou,25 Benxi,26 Taiwan12,13 and four cities in China.27 Using the CR coefficients, changes in the health effects for given changes in air pollution concentrations were estimated using the US Environmental Protection Agency methodology.6 f
The areas selected for this survey differ from the areas originally targeted by the EPRC survey. Since household income is one of the most important factors in Mumbai to determine the choice of healthcare service (public vs private), areas were targeted depending on the baseline information available about the socioeconomic status of households. For instance, households belonging to lower income strata were targeted in areas of Chembur and Colaba. Households in Dadar and Borivali were predominantly middle-income families and those targeted in Andheri belonged to higher income strata.
160
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Monetary costs of the health effects were estimated using the cost of illness approach, which comprises the cost of treatment for illnesses and lost earnings due to missed work days or restricted activity days. Similar methodology has been used for estimating the monetary burden of health effects in studies conducted in Mexico City9,28 and Santiago.29 For this study, the household survey was carried out by the researchers in five areas of Mumbai to obtain information on the cost of treatment, including cost of medicines, equipment and consultation charges of doctors, as well as restricted and total inactivity days due to illness. The households were asked to give details about the type of treatment, service provider, cost of treatment, restricted activity days, total inactivity days, and loss of wages or salary per day for each symptom and disease listed in the questionnaire. Furthermore, a survey questionnaire was administered to the medical practitioners engaged in public and private health facilities, as well as general practitioners in the city. This survey obtained information on the outpatient and inpatient costs of treatment in public and private healthcare facilities, type of treatment, days of hospitalization required for symptoms and illnesses, days of restricted and total inactivity, and average number of missed work days for each symptom and disease listed in the questionnaire. The proportion of people using public and private healthcare services, the average wages of such individuals, and the symptoms and diseases requiring hospitalization were based on information obtained from the EPRC data and previous studies. A summary of the range of estimates obtained for symptoms and diseases is given in Table 2.
Results The cross-sectional survey design used by the EPRC team influenced the choice of regression model employed in this
study. The dependent variables, presence (or absence) of cardiovascular and respiratory symptoms and diseases, were used in the logistic regression framework to examine their relationship with the explanatory variables. In order to examine the link between air pollutants and morbidity outcomes (symptoms and diseases), the odds ratios for the four pollutants (PM10, NO2, SO2 and NH3) were obtained using multipollutant variants of the regression model. PM10 and NO2 were not considered together in the same variant of the model due to their high correlation. However, PM10 and NO2 were considered separately in different variants of the regression model as their sources in Mumbai are different. PM10 mainly has domestic or area sources, including construction activity, and NO2 is produced by vehicular and industrial emissions.8 The results of the regression model were obtained and interpreted on the basis of the significance tests used in the empirical literature in this research arena.30 The regression output showed statistically significant relationships between air pollutants and specific health outcomes, thus invalidating the null hypothesis. The regression results yielded the log odds ratios, which were used to predict changes in the odds of occurrence of a symptom or disease due to changes in air pollution levels. The coefficients corresponding to the significant odds ratios were the CR coefficients used in the study to estimate changes in the health effects for given changes in air pollution concentrations, similar to the methodology used in studies conducted in Taiwan.12,13
Estimation of health effects This study found a significant relationship between PM10 and NO2 with symptoms such as cough, breathlessness, wheezing and cold, and illnesses such as allergic rhinitis and COPD. SO2 was found to be significantly associated with cardiac ailments and other chest illnesses, and NH3 was related to phlegm and
Table 2 e Range of estimates obtained for symptoms and diseases. Nature of illness Symptoms Cough Phlegm Breathlessness Wheezing Cold Diseases Bronchitis Asthma URTI Allergic rhinitis COPD
Cost of treatment Daily cost of treatment in Days required hospital (in Rs)a in hospital (non-hospital) a (in Rs)
Days of restricted activity
Days of total inactivity
Frequency per person per yeard
200e700 200e700 300e700 300e700 100
e e 1000e3000 1000e3000 e
e e 4 4 e
2 2 2 2 1
1 1 2e4b 2e4b e
3 2 1e2c 1e2c 8
250e1000 250e1000 100e1000 100e1000
1000e3000 1000e3000 e e
6 6 e e
8 8 2 2
6 6 1 1
1e2c 1e2c 2 2
1000e5000
5000e10000
7
10
5e7b
1e2c
URTI, upper respiratory tract infection; COPD, chronic obstructive pulmonary disease. Data obtained from household survey and doctors. a Lower values in the range are the costs of treatment in public OPDs and public hospitals, and higher values are the costs of consulting specialist physicians or private hospitals. b Lower value shows total inactivity days without hospitalization and upper value shows inactivity days with hospitalization. c Lower values show the frequency of hospitalization per year and upper values show the frequency of outpatient treatment per year. d Frequency indicates the number of times a person suffers from the symptom or disease per year.
161
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Table 3 e Concentrationeresponse (CR) coefficients and incidence of health effects for Mumbai. Health effect Symptoms Cough Phlegm Breathlessness Wheezing Cold Diseases Cardiac Other chest illnesses Allergic rhinitis COPD
Pollutant
CR coefficient (%)a,b
Change in health effect (%)c
Estimated incidence per 1 lakh populationd
Estimated incidence for Mumbaie
PM10 NO2 NH3 PM10 NO2 PM10 NO2 PM10 NO2
0.007 0.021 0.012 0.009 0.028 0.006 0.020 0.006 0.018
0.35 1.05 1.20 0.45 1.40 0.30 1.00 0.30 0.90
350 1050 1200 450 1400 300 1000 300 900
42,000 126,000 144,000 54,000 168,000 36,000 120,000 36,000 108,000
SO2 SO2 NH3 PM10 NO2 PM10 NO2
0.118 0.162 0.024 0.014 0.046 0.014 0.023
2.36 3.24 2.40 0.70 2.30 0.70 1.15
2360 3240 2400 700 2300 700 1150
283,200 388,800 288,000 84,000 276,000 84,000 138,000
COPD, chronic obstructive pulmonary disease; PM10, particulate matter less than 10 mm in size. a The coefficient represents percentage change in the prevalence rate of morbidity outcome per mg/m3 change in pollution concentration. b Coefficients obtained in the regression model controlling for factors such as age, gender, smoking habit, distance travelled to place of work, occupation, width of road adjacent to the residence, presence of polluting industry near residence, hours spent in kitchen, quality of kitchen ventilation and type of cooking fuel used. c Incidence of health effects estimated for 50-mg/m3 increase in PM10, 50-mg/m3 increase in NO2, 20-mg/m3 increase in SO2 and 100-mg/m3 increase in NH3. d 1 lakh ¼ 100,000. e Incidence for Mumbai city calculated for the population of 12 million as per 2001 Census.
other chest illnesses. The CR coefficients corresponding to significant odds ratios were used to estimate the incidence of health effects for given changes in air pollution concentrations,g as shown in Table 3.
Estimation of monetary burden The figures shown in Table 2 were used to estimate the monetary burden of the health outcomes. Certain assumptions were made for estimating the health costs. For instance, based on the EPRC data and empirical findings,31e33 it was assumed that 25% of city residents use public health facilities and have a daily wage of approximately INR 200. For the remaining city residents, the daily wage was assumed to be approximately INR 500 and they were assumed to use private health facilities. Based on the information gathered from medical practitioners, it was assumed that 20% of cases of breathlessness and wheezing and 50% of cases of COPD would require hospitalization. Table 4 shows estimates of the incidence and total monetary costs of illness for Mumbai. The monetary estimates shown in Table 4 indicate out-ofpocket expenses of individuals, which include the cost of g The incidence of health effects was calculated based on the US Environmental Protection Agency methodology using the following equation: DHjp ¼ bjp POPj DAp, where DHjp is the change in health effect j due to pollutant p, bjp is the CR coefficient for health effect j related to pollutant p, POPj is population at risk of health effect j, and DAp is change in the concentration of pollutant p.
treatment and foregone earnings (lost wages) due to illness. Table 5 shows the out-of-pocket expenses on treatment of health outcomes as a proportion of annual personal income. In order to determine the total monetary burden of illness caused by air pollution, personal expenses towards cost of treatment and foregone earnings, government expenditure incurred in public healthcare facilities, and societal costs due to loss of productivity were considered. The total burden of illness estimated here was, thus, the sum of personal costs, government or public expenditure, and societal costs as seen in Table 6. Government or public expenditure on health care was only relevant for cases treated in public dispensaries or hospitals, and hence was considered likewise. The cost to society is the loss of productivity, which was assumed to be equivalent to the lost wages on account of restricted activity and total inactivity days due to illness. The monetary burden of health impacts is likely to increase in future as the cost of treatment and wages increase. Therefore, the authors attempted to estimate the future trends in monetary burden, assuming an increase in the prices of medicines and consultation charges by 9% per annum, and an increase in wages by 5% per annum based on Consumer Price Index trends.34 Government expenditure in public healthcare facilities has remained more or less stagnant in percentage terms over the years in Maharashtra state and Mumbai, and per-capita expenditure has actually decreased.32 The authors assumed that this trend will continue in future. They further assumed that Mumbai’s population will increase at a rate of 1.87% per annum based on past growth rates.35 Given these assumptions, future trends in
162
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Table 4 e Incidence and total cost of illness for Mumbai (in million Indian Rupees). Nature of illness Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD
Pollutant
Estimated incidence in Mumbaia
People using public health care
Total cost in public health care
People using private health care
Total cost in private health care
Estimated total burden for Mumbai
PM10 NO2 NH3 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2
42,000 12,600 144,000 54,000 168,000 36,000 120,000 36,000 108,000 84,000 276,000 84,000 138,000
10,500 31,500 36,000 13,500 42,000 9000 30,000 9000 27,000 21,000 69,000 21,000 34,500
6.30 18.90 21.60 23.22 72.24 15.48 51.60 1.80 5.40 10.50 34.50 424.20 696.90
31,500 94,500 108,000 40,500 126,000 27,000 90,000 27,000 81,000 63,000 207,000 63,000 103,500
53.55 160.65 183.60 188.73 587.16 125.82 419.40 9.45 28.35 126.00 414.00 2709.00 4450.00
59.85 179.55 205.20 211.95 659.40 141.30 471.00 11.25 33.75 136.50 448.50 3133.20 5147.40
COPD, chronic obstructive pulmonary disease; PM10, particulate matter less than 10 mm in size. a Incidence for Mumbai city calculated for the population of 12 million as per 2001 Census, and incidence estimated for 50-mg/m3 increase in PM10, 50-mg/m3 increase in NO2 and 100-mg/m3 increase in NH3.
the monetary burden of illness due to air pollution were estimated as depicted in Fig. 1.
as allergic rhinitis and COPD. The results also indicate that NO2 has relatively greater health impacts in terms of incidence and costs. The major source of NO2 in Mumbai is dieseland gasoline-based vehicles. There has been a substantial increase in the number of vehicles in Mumbai over the last decade. With expanding suburbs, rapid development in neighbouring districts, growing affluence and availability of affordable personal vehicles, this number is going to increase manifold in the near future. Thus, programmes and initiatives are urgently required to control the NO2 levels. Substantial economic costs are associated with the increase in incidence of these health outcomes with increases
Discussion The findings of this study have revealed a significant relationship between air pollution and morbidity outcomes for Mumbai. Among the four pollutants considered in the study, PM10 and NO2 emerged as the most significant pollutants and have a significant relationship with symptoms such as cough, breathlessness, wheezing and cold, and diseases such
Table 5 e Out-of-pocket expenses as proportion of annual personal income. Nature of illness
Cost of illness without hospitalization Cost of illness (in INR)
Public healthcare facilities Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD Private healthcare facilities Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD
Cost of illness with hospitalization
Frequency per yeara
Cost as % of annual incomeb
Cost of illness (in INR)
Frequency per person per year
600 600 900 900 200 500 3000
3 2 2 2 8 2 2
3.0 2.0 3.0 3.0 2.7 1.7 10.0
e e 5000 5000 e e 37400
e e 1 1 e e 1
62.3
1700 1700 2200 2200 350 2000 10,000
3 2 2 2 8 2 2
3.4 2.3 2.9 2.9 1.9 2.7 13.3
e e 14500 14500 e e 76000
e e 1 1 e e 1
e e 9.7 9.7 e e 50.7
COPD, chronic obstructive pulmonary disease; INR, Indian Rupees. a Frequency based on the information obtained from doctors. b Annual income of those in public health care is INR 60,000 and those in private health care is INR 150,000.
Cost as % of annual income
8.3 8.3
163
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Table 6 e Total burden of illness for Mumbai (in million Indian Rupees). Nature of illness Cough Phlegm Breathlessness Wheezing Cold Allergic rhinitis COPD
Pollutant
Incidence for Mumbaia
Total burden in public health services
Total burden in private health service
Total burden of disease for Mumbai
Total burden in million US$b
PM10 NO2 NH3 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2 PM10 NO2
42,000 12,600 144,000 54,000 168,000 36,000 120,000 36,000 108,000 84,000 276,000 84,000 138,000
11.55 34.65 39.60 34.70 107.94 23.13 77.10 3.60 10.80 21.00 69.00 476.18 782.29
85.05 255.15 291.60 269.73 839.16 179.82 599.40 16.20 48.60 189.00 621.00 3213.00 5278.00
96.60 289.80 331.20 304.43 947.10 202.95 676.50 19.80 59.40 210.00 690.00 3689.18 6060.79
2.42 7.25 8.28 7.61 23.68 5.07 16.91 0.50 1.49 5.25 17.25 92.23 151.52
COPD, chronic obstructive pulmonary disease; PM10, particulate matter less than 10 mm in size. a Total incidence of health effects is estimated for the 12 million population of Mumbai as per 2001 Census figures for 50-mg/m3 increase in PM10, 50-mg/m3 increase in NO2 and 100-mg/m3 increase in NH3. b 1 US$ ¼ 40 Indian Rupees.
in PM10 and NO2. If personal costs, governmental costs and societal costs are included, the total monetary burden escalates further. However, it should be noted that the costs calculated in this study are for morbidity effects alone. If the mortality effects of the pollutants are also taken into account, the monetary burden is likely to increase substantially. Hence, the costs estimated here provide conservative lower bound estimates of the costs of air pollution for Mumbai city, and the actual costs are expected to be much larger. Most of the monetary burden comprises out-of-pocket expenses borne by the suffering individuals and families. As the estimates show, these expenses represent a significant proportion of the annual income of people, particularly for those belonging to poor households. There are, thus, tangible costs of health impacts which can no longer be ignored by policy makers and citizens. The huge burden on individuals also reflects the inadequacy of the public health facilities available in the city. The findings of this study call for strengthening of the public health infrastructure to cater for the health needs of those who cannot afford expensive treatment.
The personal costs of health impacts are mainly paid through individuals’ incomes. There is no system of private health insurance or social insurance which can adequately protect individuals from such costs. Currently, the coverage of the health insurance schemes is negligible. There is no social insurance scheme in place in the entire country. Thus, the health consequences are entirely faced by individuals and families themselves, sometimes through the sale of assets or loans to fund the medical expenses. Efforts are urgently needed to put in place a system of health insurance which eases the burden of disease on people. A significant issue emerging from this study is the stagnant health expenditure of the Government and its consequences for the population. There is an acute need to strengthen the public healthcare system in Mumbai in terms of accessibility, affordability and quality in order to help poor households. This assumes more relevance since half of Mumbai’s population lives in slums with greater exposure to all types of pollution, and a life stricken by poverty, malnutrition and unhygienic living conditions.
18000 16000 14000
Million INR
12000 10000
8000 6000 4000 2000 0 2008
2009
2010
Burden due to PM10
2011
2012
Burden due to NO2
2013
2014
2015
Total burden of illness
Fig. 1 e Future trends in burden of illness for Mumbai city. INR, Indian Rupees; PM10, particulate matter less than 10 mm in size.
164
p u b l i c h e a l t h 1 2 5 ( 2 0 1 1 ) 1 5 7 e1 6 4
Ethical approval None sought.
Funding None declared.
Competing interests None declared.
Acknowledgments The authors wish to thank Dr. Amita Athavale, Head of the Department of Chest Medicine and EPRC at KEM Hospital, Mumbai for providing the dataset, without which this study would not have been feasible.
references
1. UNDP/World Bank. Health impacts of outdoor air pollution. Urban air pollution. South Asia Urban Air Quality Management Briefing Note No. 11. UNDP/World Bank ESMAP; 2003. 2. WDI. World development indicators. Washington, DC: World Bank; 2006. 3. Romieu I, Hernandez-Avila M. Air pollution and health in developing countries: a review of epidemiological evidence. In: McGranahan G, Murray F, editors. Air pollution and health in rapidly developing countries. Earthscan; 2003. 4. HEI. Health effects of outdoor air pollution in developing countries of Asia: a literature review. Special Report 15. HEI International Scientific Oversight Committee. Boston, MA: Health Effects Institute; 2004. 5. Economic survey 2006e07. Available at. Government of Maharashtra, http://www.maharashtra.gov.in/; 2006 (last accessed 15/08/2008). 6. Ostro BD. Estimating the health effects of air pollution: a methodology with application to Jakarta. Washington, DC: World Bank; 1994. Policy Research Working Paper 1301. 7. Ostro BD, Sanchez JM, Aranda C, Eskeland GS. Air pollution and mortality: results from Santiago, Chile. Policy Research Working Paper No. 1453. Washington, DC: World Bank; 1995. 8. Ostro BD, Eskeland GS, Feyzioglu T, Sanchez JM. Air pollution and health effects: a study of respiratory illness among children in Santiago, Chile. Washington, DC: World Bank; 1998. 9. Mexico Air Quality Management Team. Improving air quality in metropolitan Mexico city: an economic valuation. Policy Research Working Paper No. 2785. Washington, DC: World Bank; 2002. 10. Li J, Guttikunda SK, Carmichael GR, Streets DG, Chang YS, Fung V. Quantifying the human health benefits of curbing air pollution in Shanghai. J Environ Manag 2003;70:49e62. 11. Alberini A, Krupnick A. Air pollution and acute respiratory illness: evidence from Taiwan and Los Angeles. Am J Agric Econ 1997;79:1620e4. 12. Chen PC, Lai YM, Wang JD, Yang CY, Hwang JS, Kuo HW, et al. Adverse effect of air pollution on respiratory health of primary school children in Taiwan. Env Health Persp 1998;106:331e5. 13. Hwang BF, Jaakkola JK, Lee YL, Lin YC, Guo YL. Relation between air pollution and allergic rhinitis in Taiwanese school children. Respir Res 2006;7:23.
14. Bank World. Cost of pollution in China: economic estimates of physical damages. Washington, DC: World Bank; 2007. 15. Kamat SR, Godkhindi KD, Shah BW, Mehta AK, Varsha N, Gregrat J, et al. Prospective 3 year study of health morbidity in relation to air pollution in Bombay, India. Bombay Air Pollution e Health Study. MCGM, KEM Hospital; 1980. 16. Kamat SR, Patil JD, Gregrat J, Dalal N, Deshpande JM, Hardikar P. Air pollution related respiratory morbidity in central and north-eastern Bombay. J Assoc Physic India 1992; 40:588e93. 17. World Bank Technical Paper No. 381. In: Shah J, Nagpal T, editors. Urban air quality management strategy in Asia: Greater Mumbai report. Washington, DC: World Bank; 1997. 18. Joseph AE, Sawant AD, Srivastava A. PM10 and its impacts on health e a case study in Mumbai. Int J Environ Health Res 2003; 13:207e14. 19. Cropper ML, Simon NB, Alberini A, Sharma PK. The health effects of air pollution in Delhi, India. Policy Research Working Paper 1860. Washington, DC: World Bank; 1997. 20. Chhabra SK, Chhabra P, Rajpal S, Gupta RK. Ambient air pollution and chronic respiratory morbidity in Delhi. Arch Environ Health 2001;56:58e64. 21. Agarwal KS, Mughal MZ, Upadhyay P, Berry JL, Mawer EB, Puliyel JM. The impact of atmospheric pollution on vitamin d status of infants and toddlers in Delhi, India. Arch Dis Child 2002;87:111e3. 22. Kumar KS, Prasad CE, Balakrishna N, Rao KV, Reddy PUM. Respiratory symptoms and spirometric observations in relation to atmospheric pollutants in a sample of urban population. Asia Pacif J Publ Health 2000;12:58e64. 23. Gupta D, Boffetta P, Gaborieau V, Jindal SK. Risk factors of lung cancer in Chandigarh, India. Ind J Med Res 2001;113: 142e50. 24. Xu X, Wang L. associations of indoor and outdoor particulate level with chronic respiratory illness. Am Rev Respir Dis 1993; 148:1516e22. 25. Jhang J, Hu W, Wei F, Wu G, Chang W-L, Chapman R. Long-term changes in air pollution and health implications in four Chinese cities. Energy Sustainable Dev 2005;IX:67e76. 26. Jin LB, Qin Y, Xu Z. Relationship between air pollution and acute and chronic respiratory disease in Benxi, China. Chin J Environ Sci 2000;17:268e70. 27. Zhang J, Song H, Tong S, Li L, Liu B, Wang L. Ambient sulfate concentration and chronic disease mortality in Beijing. Sci Total Environ 2000;262:63e71. 28. Bell M, Davis D, Gouveia N, Borja-Aburto V, Cifuentes L. The avoidable health effects of air pollution in three Latin American cities: Santiago, Sao Paulo and Mexico City. Environ Res 2006;100:431e40. 29. Maharashtra Pollution Control Board. Report on environmental status of Mumbai region. Maharashtra Pollution Control Board. Government of Maharashtra; 2005. 30. Hosmer D, Lemeshow S. Applied logistic regression. In: Wiley Series in Probability and Mathematical Statistics. New York: Wiley Interscience; 1989. 31. NSSO. Morbidity and treatment of ailments. Report No. 44. NSSO; 1998. 32. Government of Maharashtra. Maharashtra human development report 2002. Government of Maharashtra; 2002. 33. World Bank. Urban poverty and transport: the case study of Mumbai. Washington, DC: World Bank; 2003. 34. CSO. Ministry of statistics and programme implementation. Available at. Government of India, http://www.mospi.nic.in/; 2008 (last accessed 20/08/2008). 35. World Gazetteer. India: largest cities and towns and statistics of their populations. World Gazetteer; 2008.