Costs of attributable burden disease to PM_2.5 ambient air pollution exposure in Medellín, Colombia, 2010–2016

Introduction

The impacts of air pollution on morbidity and mortality indicators at the global, regional, and local levels have been extensively documented. The most relevant information, coming from empirical research, establishes the adverse effects of pollutants PM10, PM2.5, and ozone on acute and chronic events, particularly in children and older adults.¹ These impacts have been measured in the last decade in terms of the burden of attributable disease to air pollution, which involves the measurement of disability-adjusted life years (DALYs), that is, the sum of potential years of life lost due to premature death and years lived with disability.²

At the international level, there are estimates from the Institute for Health Metrics and Evaluation (IHME) of the University of Washington, whose annual reports establish the DALYs attributable to ambient air pollution, specifically related to exposure to PM2.5. In its 2017 report, from a sample of 195 countries, the IHME identified air pollution as one of the main causes of the global burden of disease; for the year 2015 the IHME calculated that long-term exposure to air pollution by fine particulate matter (PM2.5) caused 4.2 million deaths and 103.1 million years of healthy life lost, accounting for 7.6% of total global mortality.²

Seen in this way, air pollution has become a public health problem, increasingly involving the population living in large cities around the world. This problem also manifests itself in an economic burden that compromises significant society resources, related to disease care and productivity losses due to mortality and premature disability attributable to air pollution.

At the global level, for the year 2013, the economic burden of exposure to ambient air pollution by PM2.5 was established at 3.5 billion dollars, strictly associated with the loss of well-being due to premature mortality. In terms of productivity losses, the burden amounted to 144 billion dollars a year. These productivity losses were most costly in South Asia and Sub-Saharan Africa regions, which comprised 0.39% and 0.23% of the gross domestic product (GDP) in 2013, respectively. In Latin America and the Caribbean, where the effects appear to be relatively lower, this economic burden represented 0.09% of GDP in 2013.³

In Colombia, as can be seen below, few studies have documented the economic burden derived from the impact on morbidity and mortality caused by air pollution. All of this even though exceptional levels of air pollution have been registered in cities like Medellín, with daily and annual concentrations that exceed the reference values established in the respective guide of the World Health Organization (WHO).⁴ Indeed, for the period 2010–2016, in this city, the levels of PM10 and PM2.5 reached daily averages of 68.2% and 35.8 μg/m3.¹

At the national level, it is possible to underscore the pioneering research by Larsen, which estimated morbidity and mortality costs attributable to exposure to PM2.5 ambient air pollution, amounting to 2275.7 million dollars in 2002.⁵ With some methodological changes, ten years later, the World Bank (WB) updated Larsen’s study, estimating the economic burden at 6319.1 million dollars in 2010.⁶ In 2018, the National Planning Department (DNP, by its Spanish acronym) reported that it had updated these studies, with a nationwide cost measurement of 4449.6 million dollars in 2015, due to morbidity and mortality attributable to urban air pollution.⁷

In a recent study by the National Institute of Health (INS, by its acronym in Spanish) of Colombia, the economic burden due to mortality attributable to exposure to ambient and indoor air pollution by pm2.5 in the country was established at 160 million dollars for the year 2016, which represented about 0.05% of that year’s GDP. This baseline scenario was accompanied by low estimates of 84.5 million dollars and high estimates of 462.9 million dollars.⁸

For the city of Medellín, there are no estimates of the economic burden of attributable morbidity and mortality to air pollution. Based on the Local attributable burden disease to PM2.5 ambient air pollution study in Medellín,⁹ we established the costs of it. We expected that the assessment of the economic burden will provide relevant information for decision makers and analysts of this problem, in the sense of promoting the economic assessment of ongoing actions or potential air pollution containment programs in the city.

Methods

Results

In the study period 2010–2016, Medellín recorded a total of 5,540 potential years of working life lost due to premature deaths attributable to exposure to PM2.5 ambient air pollution. On average, the PYWL were about 791 per year. From this calculation the costs of premature deaths in Medellín were estimated for each event by sex and age group. The data for each year are in millions of 2016 dollars (see Table 1).

The indirect costs of premature mortality, the direct medical costs of disease care, and the total costs of premature morbidity and mortality are presented below.

Discussion

The findings of this research are supported by methodologies validated by experts and international organizations with extensive experience in the areas of health economics and environmental epidemiology. The costing procedures adopted are among the options suggested by the theoretical approaches and the empirical procedures provided by the specialized literature.

This research agrees, in this sense, with the methodological approaches applied in calculating the costs of attributable premature mortality to air pollution, carried out by the WB and the IHME in their 2016 global report,³ which also validate the human capital approach applied in this research.

In this methodological aspect, it also agrees with the study of the national aggregate of mortality attributed to pm2.5 ambient air pollution, carried out in 2018 by the INS⁸; also with international studies, such as the one reported by the Ministry of Social Development of Chile²⁸ and the one reported by Guo et al. in China.²⁹

Regarding the assessment of the morbidity costs, this study provides an added plus compared to what is usually observed in the national and international literature. Much of the reports at the global level can be applied only to the costing of premature mortality, leaving aside the cost of morbidity. The abovementioned INS study did not report the morbidity cost either, focusing exclusively on mortality costs. At the national level, in the study by Larsen,⁵ as well as in its updates of 2014⁶ and 2017,⁷ estimates are made of the two components, that is, premature mortality and morbidity attributed to air pollution.

With regard to morbidity, it would also be necessary to specify the methodological concordances and differences of this study when compared to those at the national and international level, in terms of valuation of morbidity costs. In the study of Medellín, the cost-of-illness approach from the field of health economics was adopted to estimate the morbidity cost, based on information validated by official and governing institutions of the city’s health care system. The national studies already referred to, that is, those by Larsen,⁵ Golub et al.,⁶ and the DNP,⁷ proceed in the same way in the cost of morbidity. The WB and the IHME,^3,22 although they do not advocate the effective measurement of these morbidity costs, they do privilege, in their recommendations, the willingness-to-pay approach.

However, these organizations show the uncertainties and variability in the estimates of morbidity costs with a willingness-to-pay approach. Overall, they recognize the lack of a standard framework for the valuation of these costs and, therefore, they do not measure them in their reports on the economic environmental burden. This is understandable from the perspective of international comparability of estimates, given that for a wide range of countries there may be a lack of data on health expenditures and medical treatment costs. This is compounded by differences between countries in health care systems, contractual forms, and payment and billing schemes between patients, care providers, and public institutions.

This study is the first research in the city of Medellín aimed at estimating the costs of the burden of morbidity and mortality attributed to PM2.5 ambient air pollution. Its realization required a prior and novel effort to estimate the burden of morbidity and mortality attributable to air pollution in this city, for which standard methodologies from the IHME and WHO were adopted.^2,30

According to the findings of this research, in the period 2010–2016, the costs of morbidity and mortality attributed to pm2.5 pollution totaled in the city of Medellín about 281.2 million dollars: 80% of these costs corresponded to the premature mortality component, and the remaining 20% to medical care and productivity losses due to morbidity.

In relative terms, these figures do not differ from the reports of national and international studies. In the abovementioned report of the Organization for Economic Cooperation and Development (OECD),³¹ it is reported that the impacts of morbidity represented 27% of the total costs, although this figure requires caution, since the component of mortality comes from the willingness-to-pay approach. Similar caution should be taken when mentioning the WB and IHME report of 2016,²² which includes studies from Nicaragua and China where morbidity costs may represent 44% of the total morbidity and mortality costs. In a more recent study concerning Egypt, which incorporates a detailed review of the literature, Larsen³² establishes that this proportion would be in the range of 10% to 30% of the total cost.

Experts from the OECD, as well as those from the IHME and the WB, agree in reporting a minimum proportion of 10% for morbidity costs contribution to the total morbidity and mortality costs. Of the three national studies that addressed the valuation of morbidity costs attributed to air pollution, for this component with a cost-of-illness approach and for mortality with a willingness-to-pay approach, Larsen⁵ and Golub et al.⁶ estimated a proportion of 21%. In the most recent one by the DNP,⁷ whose methodological approach adopts the two previous ones, the morbidity costs reported for 2015 represent 13.1% of the total costs of morbidity and mortality attributed to pm2.5 ambient air pollution.

In the period 2010–2016, Medellín registered a slight downward trend in the costs of premature mortality, especially in what corresponds to the group of men; this trend reverts as of 2015, which coincides with the rebound in the number of PYWL. This circumstance may be associated with the PM2.5 pollution episodes that have occurred since that year in the city of Medellín. By then, the daily concentration levels of this particulate matter during the first quarter of 2015 and 2016 reached values higher than three-times the WHO reference levels.¹

The costs of premature mortality concentrated to a greater extent in the group of men (61.1%) and in the extreme age population groups, particularly in adults aged 55 and over (78.5%). In the group of men, more than in that of women, there was an increasing gradient in the costs of premature mortality as age increased. This behavior reflects the combined effects of the wage differential and the ages at which deaths occur. Men earn higher wages and the effects of air pollution fall more heavily on the young and elderly populations.

According to the types of events that are the object of this study, the results of mortality costs should be addressed in greater detail. When measurements are made based on retirement ages, acute events represent about 81.9% of total costs for the period 2010-2016. Upon including ages over 60 in the estimation of these costs, acute events represent 65.4%. In both scenarios, the mortality costs of this type of event are higher when compared to those of chronic events.

It should be noted that, in the group of acute events, the costs of deaths associated with ischemic heart diseases stood out, reaching a participation of 42% with respect to the total costs for the period. This pattern of distribution for premature mortality costs is in line with the findings reported in the INS study⁸ at the national level, with respect to a higher participation of mortality costs due to acute events. That study reveals a participation of nearly 48% for mortality costs associated with ischemic heart diseases in the total costs for 2016.

The costs of morbidity attributed to PM2.5 ambient air pollution in the city of Medellín reached a value of 56.5 million dollars in the period 2010–2016. Of these, 92.7% corresponded to direct costs of health care, and the remaining 7.3% corresponded to costs of lost working days due to sick leaves. These morbidity costs represented nearly 20% of the total costs for the period, and showed, in their component of health care costs, a trend contrary to the costs of premature mortality, with an increasing trend in the first years and a decrease since 2013, which shows a lack of sensitivity to the episodes of PM2.5 ambient air pollution registered in the city in 2015 and 2016.

The consolidated estimate of these costs, seen in its distribution by sex, showed a different behavior from that registered for mortality costs, with a more even distribution between men and women. When analyzed according to the patients’ age, the marked gradient that was observed for mortality costs was not seen, although the pattern of higher incidence in the extreme groups was present, particularly in those aged 55 and over, whose total costs represented about 70.4% in women and 68% in men.

This pattern reflects two situations widely documented in the specialized literature^33,34: on the one hand, the fact that women, compared to men, seek medical care more frequently and adhere more regularly to ongoing care; and on the other hand, there is also evidence that morbidity, especially the chronic kind, due to a strictly biological effect, tends to be more pronounced in the elderly. In fact, in the total health care costs, those corresponding to the care of chronic respiratory diseases represented 43%, while those of acute lower respiratory infections added a participation of 41%.

The costs of morbidity and mortality attributable to PM2.5 air pollution in Medellín totaled, for the entire period, 281.2 million dollars, with an annual average of 40.2 million dollars. These annual costs represented 0.013% of the national GDP in 2016 and 0.091% of the GDP of the Department of Antioquia in the same year. These are relative calculations lower than the proportions estimated in the national studies by Larsen (0.78% of the national GDP),⁵ by Golub et al. (1.12% of the national GDP),⁶ and the DNP (1.5% of the national GDP).⁷

Measurements of the cost of premature mortality assumed from the human capital approach support greater comparability. In effect, these costs, in the case of Medellín, annually represented 0.011% of the national GDP and 0.072% of the departmental GDP. Meanwhile, these mortality costs, with a similar human capital approach, were calculated by the INS⁸ for the national level at 0.05% of the 2016 GDP. It should be noted that these productivity losses due to premature mortality were estimated by the WB and IHME³ at 0.09% of the 2013 GDP in Latin America and the Caribbean.

The foregoing shows consistency in the estimates made in this research for the costs of premature mortality from the human capital approach, but at the same time shows that the valuation of mortality costs with this approach, when compared to the willingness-to-pay approach, underestimates the true environmental costs.

In this sense, it is worth mentioning that the economic burden derived from air pollution goes beyond the costs of morbidity and mortality attributed to exposure to PM2.5. It should be taken into consideration that other pollutants (O3, SOx, NOx, and CO2), as well as their synergistic effects, also negatively impact the health of the population. Additionally, it is evident that several of these pollutants also pose an economic burden in terms of the deterioration of production processes and public and private infrastructure, constituting a hindrance to economic development and the quality of life of the population.

These limitations are typical of the restrictions in the scope of all studies about economic burden attributed to air pollution. Regarding the Medellín study, it is worth mentioning other limitations concerning the information base and the cost estimation methodology, specifically related to the values of the population attributed fraction and the valuation of mortality costs. It is imperative to make progress in epidemiological research, through cohort studies that allow documenting, for the local context, the real attribution of morbidity and mortality to air pollution. Likewise, it is convenient to carry out studies on the statistical value of life related to the risks of mortality due to air pollution at the local or national level.

Regarding the estimation of morbidity costs, this study did not make any progress in the component of indirect costs or production losses incurred by patients and relatives due to effects of disease and the search for effective medical care. There was also no information available to value the intangible costs associated with having a disease. Studies on disease should be encouraged for events specifically associated with air pollution, incorporating local information on sick leaves, reduction of productive capacity, and willingness to pay to avoid suffering and pain associated with those illnesses.

Nonetheless, beyond the above limitations, this study offers a valid informative basis on the costs of morbidity and mortality attributed to PM2.5 ambient air pollution in the city of Medellín. In this sense, the results may constitute supporting material to establish priorities in public policy decisions regarding the general objectives of local development and the policy options aimed at containing air pollution in the city. The valuation of these costs will facilitate the economic evaluation, either cost-effectiveness or cost-benefit, of those policy actions that involve the health and environment sectors.

Data availability