Towards a Fuller Assessment of the Economic Benefits of Reducing Air Pollution from Fossil Fuel Combustion: Per-Case Monetary Estimates for Children’s Health Outcomes

Abstract

Background:

Impacts on children’s health are under-represented in benefits assessments of policies related to ambient air quality and climate change. To complement our previous compilation of concentration-response (C-R) functions for a number of children’s health outcomes associated with air pollution, we provide per-case monetary estimates of the same health outcomes.

Objectives:

Our goal was to establish per-case monetary estimates for a suite of prevalent children’s health outcomes (preterm birth, low birth weight, asthma, autism spectrum disorder, attention-deficit/hyperactivity disorder, and IQ reduction) that can be incorporated into benefits assessments of air pollution regulations and climate change mitigation policies.

Methods:

We conducted a systematic review of the literature published between January 1, 2000 and June 30, 2018 to identify relevant economic costs for these six adverse health outcomes in children. We restricted our literature search to studies published in the U.S., with a supplemental consideration of studies from the U.K. and prioritized literature reviews with summary cost estimates and papers that provided lifetime cost of illness estimates.

Results:

Our literature search and evaluation process reviewed 1,065 papers and identified 12 most relevant papers on per-case monetary estimates for preterm birth, low birth weight, asthma, autism spectrum disorder, and attention-deficit/hyperactivity disorder. Details are presented in full. We separately identified estimates of the lost lifetime earnings associated with the loss of a single IQ point. The final per-case cost estimates for each outcome were selected based on the most robust evidence. These estimates range from $23,573 for childhood asthma not persisting into adulthood to $3,109,096 for a case of autism with a concurrent intellectual disability.

Conclusion:

To our knowledge, this is the first time that the child-specific health outcomes of preterm birth, low birth weight, asthma, autism spectrum disorder, attention-deficit/hyperactivity disorder, and IQ reduction have been systematically valued and presented in one place. This is an important addition to the body of health-related valuation literature as these outcomes have substantial economic costs that are not considered in most assessments of the benefits of air pollution and climate mitigation policies. In general, however, the available per-case estimates presented here did not incorporate the broad societal and long-term costs and are likely underestimates. Although our context has been air pollution and climate policies, the per-case monetary estimates presented here can be applied to other environmental exposures. Fuller assessments of health benefits to children and their corresponding economic gains will improve decision-making on environmental policy.

I. Introduction

There is growing awareness of the adverse impacts of environmental exposures on birth outcomes, neurodevelopment, and health of children over their lifecourse. Such exposures include air pollutants, endocrine disrupting chemicals, and metals such as lead and mercury. Air pollutants, for example, are associated with preterm birth (PTB), low birth weight (LBW), asthma, autism sprectrum disorder (ASD), attention-deficit/hyperactivity disorder (ADHD), and intelligence quotient (IQ) reduction [1]. Disorders such as asthma and ADHD are prevalent in children in the U.S. and have been increasing over time, with asthma having a prevalence of about 8% and ADHD a prevalence of 10% [2],[3]. Importantly, even disorders with lower prevalence, such as ASD, represent a growing public health concern, with about 1 in 60 U.S. children affected [4].

Despite this high burden of childhood illness, benefit-cost assessments of policies and other interventions have been limited in considering impacts in children – both in terms of avoided cases and avoided economic costs. For example, the air pollution-related child health outcomes considered in the Benefits Mapping and Analysis Program Community Edition (BenMAP-CE) of the U.S. EPA, have been limited to acute bronchitis, lower and upper respiratory symptoms, school day loss, and asthma exacerbation [5]. However, including additional children’s health outcomes requires the relevant concentration-response functions and the associated per-case monetary estimates. Perera et al. address the need for concentration-response (C-R) functions with a synthesis of recent evidence from epidemiological studies of health impacts among young children, producing a series of C-R functions for the six ‘new’ outcomes that are all related to notable fossil fuel combustion by-products, namely particulate matter (PM_2.5 and PM₁₀), nitrogen dioxide (NO₂), and polycyclic aromatic hydrocarbons (PAH) [1]. The level of evidence supporting these C-R functions varies by outcome from ‘causal or likely to be causal’ (PTB, LBW, ASD, development of childhood asthma) to ‘suggestive’ (ADHD, IQ reduction).²

To address the second issue of identifying per-case monetary estimates for these same six outcomes, we reviewed the existing literature and used an incidence-based cost of illness (COI) approach. It is important to note that while the per-case estimates we present are discussed in the context of air pollution and climate change, our results are applicable to many other exposures that affect one or more of these child-specific health outcomes. Examples include lead which has been associated with loss of cognitive ability during childhood and perfluorooctanoic acid (PFOA) which has been linked to decreased fetal growth and increased risk of low birth weight [6],[7]. We chose air pollution reduction as our focal application for two main reasons: 1) BenMAP-CE provides existing benefit-cost assessment infrastructure that focuses on ambient air pollution, and 2) C-R functions established by Perera et al. are based on exposure to fossil fuel combustion by-products. We note that BenMAP-CE uses the term ‘unit values’ where we use ‘per-case monetary estimates’.

We expect that the estimates presented here can be used as inputs in benefit-cost assessments that examine pollution reduction policies. In addition. our methodology for identifying per-case monetary estimates can serve a theoretical as well as a practical purpose, providing a framework for considering other outcomes and settings.

II. Methods and Materials

In determining incidence-based per-case estimates for each child-specific health outcome, it is important to include both immediate medical costs and, to the extent possible, long-term and broad societal costs. This imperative partly comes from the growing body of literature that implicates poor health status early in life as a source of critical later-life outcomes, and is consistent with existing guidance for developing incidence-based COI measures for health outcomes [8]. The following section describes how we developed these estimates.

Literature Search

We completed a systematic literature review to support our per-case monetary estimates. We executed our literature search in PubMed (National Library of Medicine, Bethesda, MD, USA) using a search strategy similar to that of Soilly et al., that linked economic assessment keywords to the relevant health outcomes by Boolean connectors [9]. Below we provide an example query for preterm birth (PTB).

((“Costs and Cost Analysis” OR “Health Care Costs” OR “Direct Service Costs” OR “Hospital Costs” OR “Drug Costs” OR “Cost of Illness”[MeSH Terms])) AND (“Premature Birth” OR “Preterm”[MeSH Terms])

We restricted searches to articles from the U.S., with a supplemental consideration of studies from the U.K., that were published in English between January 1, 2000 and June 30, 2018. We included U.K. studies because, as for autism spectrum disorder, incidence-based COI estimates from the U.S. and U.K. have been remarkably similar despite differences in the respective healthcare systems [10].

We screened the initial set of titles and abstracts to identify studies for further review based on the generalizability of their results. We then prioritized literature reviews with summary cost estimates and papers that provided lifetime COI estimates. If an appropriate review paper was identified for a given outcome, we narrowed the search to focus on studies published after the review. If successive papers were identified, they were checked for consistency and/or new valuation insight. An exception was made to this process if an economic assessment published prior to the review paper was heavily cited in the subsequent literature but may not have been cited in the review paper. This occurred in the case of PTB for which the assessment by the Institute of Medicine (IOM) was deemed a valuable cost valuation source for our purposes. We believe that in this case the IOM assessment was not included in the review paper due to the fact that it was published by the National Academy of Sciences rather than in a peer-reviewed journal.

The process detailed above was replicated for low birth weight, asthma, autism spectrum disorder, and attention-deficit/hyperactivity disorder. For the remaining outcome, “Loss of IQ Point”, appropriate MeSH terms were unavailable, and the selected monetary estimates are based on review of all peer-reviewed papers.

Our literature search and evaluation process is summarized in Figure 1. Briefly, we reviewed 1,065 papers and identified 12 of the most relevant papers on per-case monetary estimates for preterm birth, low birth weight, asthma, autism spectrum disorder, and attention-deficit/hyperactivity disorder. For each of these five outcomes, between one and four papers were found that met the search criteria. Table 1 presents references and details on study design for these 12 papers.

Figure 1.

Flowchart.

Table 1.

The initial set of 12 economic valuation papers spanning five of the six focal outcomes

Authors (year)	Study Location(s)	Study Design and/or Sample Size	Date of Cohort
PRETERM BIRTH (PTB)
REVIEW PAPER
Soilly et al. (2014)	USA, UK, Greece, Finland	18 studies published between 1990 and 2011	Multiple
ADDITIONAL PAPERS
Institute of Medicine (2007)	Utah, USA	Cohort N = 11,357	1998–2000
Khan et al. (2015)	East Midlands, England, UK	Cohort N = 2,404	2009–2010
Hall and Greenberg (2016)	Ohio, USA	Cohort N = 1,444	2012
LOW BIRTH WEIGHT (LBW)
Schmitt et al. (2006)	California, USA	Population-based N = 518,704	2000
Russell et al. (2007)	USA	Population-based N = 384,200	2001
ASTHMA
Brandt et al. (2012)	California, USA	Risk assessment and cost estimation through literature search and survey data	N/A
Nurmagambetov et al. (2018)	USA	Regression of Medical Expenditure Panel Survey (MEPS) Data	2008–2013
AUTISM SPECTRUM DISORDER (ASD)
REVIEW PAPER
Buescher et al. (2014)	USA, UK	Literature search conducted in October 2013	N/A
ATTENTION-DEFICIT/HYPERACTIVITY DISORDER (ADHD)
REVIEW PAPER
Pelham et al. (2007)	USA	13 studies published between 2001 and 2007	N/A
ADDITIONAL PAPERS
Telford et al. (2013)	Cardiff, Wales, UK	Cohort N = 143	1998–2002
Gupte-Singh et al. (2017)	USA	Regression of Medical Expenditure Panel Survey (MEPS) Data	2011

Per-Case Monetary Estimates

Methods of monetizing health endpoints were modeled after the methods used by BenMAP-CE [5]. We standardized selected monetary values from the literature to 2015 USD using annual inflation index values. Specifically, we used the ‘Consumer Price Index for All Urban Consumers: Medical Care’ to adjust medical costs; the ‘Consumer Price Index for All Urban Consumers: Tuition, other school fees, and childcare’ to adjust education costs; the ‘Employment Cost Index’ to adjust wages and salaries for productivity costs; and the all-inclusive ‘Consumer Price Index for All Urban Consumers (CPI-U)’ for more general expenses [11]. Estimates from U.K. studies originally expressed in pounds sterling (£) were first converted to an equivalent USD value using the average annual exchange rate for the year of the study. This value was then updated, if needed, to the year 2015 equivalent using the relevant U.S. inflation index.

Certain studies presented annual costs rather than complete costs for a given case over the lifecourse. To integrate these costs into a net present value equivalent, we discounted the annual costs over an appropriate number of years using a 3% discount rate; years before the age of onset were still considered part of the discounting stream but were valued at zero cost. While alternative discount rates could be considered, the 3% value is supported by EPA and BenMAP-CE methodology, as well as selected U.K. studies such as Khan et al. where a comparable 3.5% discount rate was used [5],[12].

III. Results

Details on the per-case monetary estimates drawn from the initially selected 12 papers are presented in full in Table 2. Further information regarding the findings and transformation of these values for each outcome is provided below, with more comprehensive documentation of each transformation located in the Supplemental Materials. Table 2 also includes the estimates of the lost lifetime earnings associated with the loss of a single IQ point. The final selected per-case cost estimates for each outcome based on the current economic evidence are found in Table 3.

Table 2.

Details of per-case monetary estimates from the initial 12 papers plus IQ cost estimates

Study Basis	Endpoint	Cost of Illness (COI) Definition	Unit Value (in 2015$)
PRETERM BIRTH (PTB)
Review Paper
Soilly et al. (2014)a	PTB by weeks’ gestational age (wGA)	medical costs during the first year of life (±SD)	<28 wGA: $173,607 (±56,377.97) 28–31 wGA: $62,822 (±31,166.38) 32–34 wGA: $18,062 (±8,404.27) 35–36 wGA: $3,667 (±514.15)
Additional Papers
Institute of Medicine (2007)	Any PTB	medical costs + special education costs + lost productivity costs, 3% DR	$70,101
Khan et al. (2015)	PTB by weeks’ gestational age (wGA)	medical costs from birth to hospital discharge	32–33 wGA: $17,362 34–36 wGA: $4,736
		societal costs from birth to 24 months of age, 3.5% DR	32–33 wGA: $17,741 34–36 wGA: $6,691
Hall and Greenberg (2016)	PTB by weeks’ gestational age (wGA)	initial hospital costs	<28 wGA: $483,891 28–31 wGA: $183,367 32–36 wGA: $17,652
	Any PTB		$99,882
LOW BIRTH WEIGHT (LBW)
Schmitt et al. (2006)	LBW by birth weight group	hospital costs (for infant, prenatal, and maternal procedures and services)	<500g: $236,218 500–749g: $392,303 750–999g: $345,641 1000–1249g: $219,395 1250–1499g: $137,945 1500–1749g: $92,902 1750–1999g: $59,116 2000–2499g: $20,102
	Any LBW		$66,490
Russell et al. (2007)	LBW among those with slow growth/malnutrition	initial hospital costs	$15,560
ASTHMA
Brandt et al. (2012)	Age 3 onset, no persistence into adulthood	medical costs + lost productivity costs, 3% DR	$47,480
Nurmagambetov et al. (2018)	Age 3 onset, no persistence into adulthood	medical costs + absentee costs, 3% DR	$23,573
	Age 3 onset, persistence into adulthood		$91,954
AUTISM SPECTRUM DISORDER (ASD)
Review Paper
Buescher et al. (2014)	ASDs with intellectual disability	medical costs + non-medical expenditures + special education costs + lost productivity costs, 3% DR	$3,109,096
	ASDs without intellectual disability		$1,805,941
ATTENTION-DEFICIT/HYPERACTIVITY DISORDER (ADHD)
Review Paper
Pelham et al. (2007)	ADHD case persisting from ages 5–17	medical costs + education costs + crime and delinquency costs, 3% DR	$182,045
Additional Papers
Telford et al. (2013)	ADHD case persisting from ages 12–18	medical costs + costs from social and education services, 3% DR	$43,142
Gupte-Singh et al. (2017)	ADHD case persisting from ages 5–17	medical costs, 3% DR	$8,178
LOSS OF AN IQ POINT
Grosse et al. (2002) + Grosse et al. (2007) + Perera et al. (2014)	Lower bound value	lost lifetime earnings, 3% DR	$8,730
	Best estimate value		$11,298
Gould (2009)	Best estimate value	lost lifetime earnings, 3% DR	$21,467

^aYear not provided for the USD values from this paper, so no inflation calculation performed.

Table 3.

Selected per-case monetary estimates

Study Basis	Endpoint	Cost of Illness (COI) Definition	Unit Value (in 2015$)
PRETERM BIRTH (PTB)
Institute of Medicine (2007)	Any PTB	medical costs + special education costs + lost productivity costs, 3% DR	$70,101
TERM LOW BIRTH WEIGHT (TLBW)
Russell et al. (2007)	LBW among those with slow growth/malnutrition	initial hospital costs	$15,560
ASTHMA
Nurmagambetov et al. (2018)	Age 3 onset, no persistence into adulthood	medical costs + absentee costs, 3% DR	$23,573
	Age 3 onset, persistence into adulthood		$91,954
AUTISM SPECTRUM DISORDER (ASD)
Buescher et al. (2014)	ASDs with intellectual disability	medical costs + non-medical expenditures + special education costs + lost productivity costs, 3% DR	$3,109,096
	ASDs without intellectual disability		$1,805,941
ATTENTION-DEFICIT/HYPERACTIVITY DISORDER (ADHD)
Pelham et al. (2007)	ADHD case persisting from ages 5–17	medical costs + education costs + crime and delinquency costs, 3% DR	$182,045
LOSS OF AN IQ POINT
Grosse et al. (2002) + Grosse et al. (2007) + Perera et al. (2014)	Best estimate value	lost lifetime earnings, 3% DR	$11,298
Gould (2009)	Best estimate value	lost lifetime earnings, 3% DR	$21,467

Preterm Birth (PTB)

PTB (also known as premature birth) is one that occurs before the start of the 37^th week of pregnancy [13].

The literature search yielded a review paper by Soilly et al. from 2014 [9]. Although this paper included a total of 18 papers from a range of locations, the authors ultimately provided summary values based on four U.S.-specific studies. These studies were grouped and used in weighted mean calculations due to their similarities in terms of time scale (all were short-term follow-up) and economic evaluation methodology. Four per-case monetary estimates for the direct medical costs during the first year of life were provided for four different categories of prematurity defined in terms of weeks’ gestational age (wGA): extreme prematurity (<28 wGA), early prematurity (28–31 wGA), moderate prematurity (32–34 wGA), and late prematurity (35–36 wGA). Additionally, standard deviation estimates were provided for each of the estimates associated with these categories (see Table 2).

When the literature search was narrowed to look for papers subsequent to this 2014 review, two additional studies related to PTB were found. The first by Khan et al. was specific in two ways: it focused on a cohort in the East Midlands of England and it concentrated on moderate and late preterm birth (32–36 wGA) [12]. While this study was somewhat limited in generalizability, it corroborated the per-case monetary estimates from the 2014 review paper for the moderate prematurity and late prematurity categories (e.g., the estimate for the Soilly et al. 32–34 wGA category was $18,602; for the Khan et al. 32–33 wGA category the estimate was $17,362) (see Table 2). Two COI-based estimates were presented in the Khan et al. study, one encompassing direct medical costs from birth to hospital discharge, and another encompassing societal costs from birth to 24 months of age. However, values do not differ substantially between these two COI definitions, suggesting that societal costs of PTB early in life may be low.

The second post-2014 study on PTB by Hall and Greenberg focused on Hamilton County in Ohio [14]. While it summarized direct medical costs in the form of initial hospital costs, which could be considered similar to direct medical costs during the first year of life, this study produced much larger per-case monetary estimates than the Soilly et al. study. An extremely preterm birth (<28 wGA) totaled $483,891 in Hall and Greenberg versus $173,607 in Soilly et al. (see Table 2). Moreover, the Hall and Greenberg value lies more than five standard deviations above the Soilly et al. value. One benefit of the Hall and Greenberg data, however, is that they allowed for weighted averaging; and a summary value of $99,882 for any preterm birth was calculated.

In the literature published following the Soilly et al. review paper, costs from a 2007 publication by the Institute of Medicine (IOM) were continually cited. A 2016 paper by Frey and Klebanoff stated that this IOM publication “remains perhaps the most comprehensive analysis of the cost of PTB” [15]. For this reason, we prioritized this study. Similar to the Hall and Greenberg paper, it provides a summary value that is applicable to any PTB. Importantly, this summary value includes costs beyond initial hospital costs; it includes medical care costs up to age 5, maternal delivery costs, and early intervention costs, as well as lifetime medical, special education, and lost productivity costs associated with four disabling conditions related to preterm birth [16]. The IOM value further suggests that Hall and Greenberg may be an overestimation of costs; although the IOM estimate incorporates more long-term costs than the Hall and Greenberg estimate, it is lower than the latter overall value ($99,882 versus $70,101 – see Table 2).

The selected estimate presented in Table 3 is the per-case monetary estimate applicable to any given PTB from the IOM publication, selected because it provides the most comprehensive estimate that can be applied to any PTB. In addition, we believe that it is more generalizable than the other PTB studies because “substantial effort was undertaken to stratify the sample and adjust estimates so as to make valid approximation of costs for the nation as a whole” [16].

Low Birth Weight (LBW)

LBW is defined as a live singleton birth weighing less than 2500 grams. There is a distinction between term LBW (TLBW) and preterm LBW (PLBW), although this distinction is not generally made in the economic literature.

We did not identify a suitable review paper for LBW. Instead, we selected per-case monetary estimates from two population-based studies. Schmitt et al. utilized data from a 2000 California birth cohort and presented infant, prenatal, and maternal costs for several birth weight groups, in addition to a summary value for any LBW [17]. Russell et al. studied the 2001 Nationwide Inpatient Sample from the Healthcare Cost and Utilization Project with data on 384,200 births characterized by a preterm birth or low birth weight diagnosis [18]. Both studies focused on hospital costs alone.

Because of the overlap in PTB and LBW costs (PTB infants are likely to be LBW, and vice versa), we attempted to focus on the cost of TLBW. As no studies presented this cost explicitly, we selected the category of slow growth/malnutrition among those with a LBW diagnosis presented in Russell et al. as a proximate equivalent to TLBW; this per-case monetary estimate was $15,560 (see Table 2). Alternatively, as TLBW babies are most likely to be in the birth weight group directly below the normal birth weight threshold (i.e. the 2400–2499g category), we also selected the related Schmitt et al. cost estimate of $20,102 (see Table 2). This is fairly similar to the previous per-case monetary estimate established from the Russell et al. paper, further supporting the idea that those with slow growth/malnutrition among the LBW diagnosis category are roughly equivalent to those who are TLBW.

The selected per-case monetary estimate presented in Table 3 is the estimate from Russell et al., as we believe it most adequately represents the cost of TLBW and favors generalizability, based as it is on national data.

Asthma

Some asthma-related endpoints and their monetary estimates are currently included in BenMAP-CE. These currently documented endpoints relate to exacerbation of existing asthma and include hospital admissions for asthma, emergency room visits for asthma, asthma-related acute symptoms, and chronic asthma specific to adults ages 27–99 [2]. Of particular interest for our review was the lifetime cost for a newly developed case of chronic asthma during childhood. Because of the breadth of literature on asthma and our interest on childhood asthma incidence, we added an additional search term limiting results to ages of children (birth-18 years).

We did not identify an acceptable review paper for incident cases of childhood asthma. Instead, two studies that presented annual costs of asthma per case were selected and discounting was conducted to transform these values to lifetime costs per case. Some assumptions were made about a typical case of asthma for these discounting procedures, including that age of onset occurs at 3 years of age, which is supported by recent trends in the age of diagnosis [19]. Additionally, to take into account the fact that some cases of asthma persist into adulthood while some do not, we present different per-case monetary estimates for these two scenarios. Importantly for these scenarios, studies across populations demonstrate a persistence rate of around 15% for asthma that continues from childhood into adulthood [20], [21]. However, this proportion does not include cases of asthma that remit then relapse later in life.

Brandt et al. presented annual cost estimates for childhood asthma that included direct medical costs and indirect lost productivity costs [22]. Nurmagambetov et al. presented estimates that similarly distinguished between direct medical costs and indirect absentee costs, and included annual cost estimates for both childhood and adult asthma [23]. Nurmagambetov et al. also allowed for calculation of estimates representative of both persistent and non-persistent asthma cases; in addition, it was the most recent paper found related to this outcome and it considered asthma costs nationally. We therefore selected per-case monetary estimates for a case of asthma with onset at age 3 and no persistence into adulthood using both studies and a per-case monetary estimate for a case of asthma with onset at age 3 and persistence into adulthood from the Nurmagambetov et al. study. The final selected per-case monetary estimates presented in Table 3 are those from Nurmagambetov et al.

Autism Spectrum Disorder (ASD)

ASD is a set of developmental disorders defined by impairments in social interaction and communication along with restricted and repetitive patterns of behavior [24]. The costs of ASD can be large and affect the individual, family, and society as a whole. A 2014 review paper by Buescher et al. considered comprehensive costs including residential care, special education, employment support, medical and nonmedical services, and individual and parental productivity losses [10]. They included both U.K. and U.S. costs, although we present only the U.S. per-case monetary estimates. A distinction was made between ASDs with and without a concurrent intellectual disability. The cost for ASD with an intellectual disability is almost twofold that for ASD without ($3,109,096 versus $1,805,941) (see Table 2), further highlighting the spectrum aspect of the disorder in terms of costs.

As per the search criteria, a query of additional papers published after 2014 was completed but no subsequent papers of interest were found for this outcome. For this reason, the selected per-case monetary estimates in Table 3 are from Buescher et al.

Attention-Deficit/Hyperactivity Disorder (ADHD)

ADHD is one of the most common mental disorders affecting children (although it also affects many adults), with symptoms of inattention, hyperactivity, and impulsivity [24].

The literature search resulted in a 2007 review paper by Pelham et al. that provided an annual per individual cost of ADHD in children and adolescents [25]. This paper was the most comprehensive one found in the search; it considered medical, education, and crime and delinquency costs. For our purposes, the assumption was made that ADHD primarily afflicts school-aged children, so discounting occurred to convert the annual cost to the total cost of an ADHD case persisting from ages 5–17.

The literature search was narrowed to the period following the review by Pelham et al. and identified two other studies that met the search criteria. Telford et al. provided annual cost estimates that assessed the economic burden of ADHD in the U.K. and considered medical costs and costs from social and education services [26]. The cohort assessed in that study consisted of participants between the ages of 12 and 18 years, so we defined our cases to be ages 12–18 and our discounting to convert the annual cost to the total cost occurred under this age assumption. The second paper by Gupte-Singh et al. provided annual cost estimates that were the result of a regression analysis and that considered medical costs of ADHD alone [27]. Discounting occurred to convert this annual cost to the total cost of an ADHD case persisting from ages 5–17 so that we could compare the costs of ADHD affecting school-aged children between these two studies.

As expected, the per-case monetary estimate developed from Gupte-Singh et al. that considers medical costs alone is lowest; and the per-case monetary estimate developed from Pelham et al. that includes crime and delinquency costs is the highest ($8,178 versus $182,045 – see Table 2). The large difference between these two costs highlights the important relative role of societal costs such as education and crime and delinquency. However, the Pelham et al. and Gupte-Singh et al. values are not directly comparable to the per-case monetary estimate developed from the Telford et al. study in that they are discounted with different ages of incidence and years of persistence in mind. The estimate derived from Telford et al. may be an underestimation of the true childhood costs of ADHD in that it considers a narrower range of years, mainly focusing on adolescence rather than childhood more fully.

The selected per-case monetary estimate presented in Table 3 is from Pelham et al. as it presented a summary value based on a comprehensive literature search that was most holistic in its definition.

Loss of an IQ Point

A monetary estimate for the cost related to the loss of a single IQ point was derived differently from the per-case monetary estimates for the other health outcomes in this paper in that a formal literature search did not take place. This was due to the fact that intelligence quotient (IQ) does not have an associated MeSH Term.

The monetary estimates for IQ point are instead composites of different peer-reviewed papers. Perera et al. calculated the lost lifetime earnings associated with loss of IQ point from prenatal exposure to airborne PAH and based their methodology on a 2002 paper by Grosse et al. while using an updated, more conservative earnings-IQ slope from a 2007 paper by Grosse [28],[29], [30]. An estimate for loss of an IQ point utilized components of all three of these papers to calculate a cost based on lost lifetime earnings discounted at a rate of 3%; both a best estimate value ($11,298) and a lower bound value were calculated (see Table 3). Another widely cited estimate for loss of an IQ point is $21,467; this value is from Gould (2009) and based on Salkever (1995), Schwartz (1994), and Nevin et al. (2008) [31],[32, 33] [34]. Both monetary estimates are presented in Table 3.

IV. Discussion

Impacts on children’s health are under-represented in many benefits assessments, including those associated with poor ambient air quality. This is significant as the World Health Organization (WHO) estimates that more than 40 percent of the burden of environmentally related disease and more than 88 percent of the burden of climate change is borne by children younger than 5 [35]. Therefore, monetary estimates for child-specific health outcomes are necessary components for comprehensive benefits assessments of clean air and climate mitigation policies.

To our knowledge, this is the first time that the six child-specific health outcomes of PTB, LBW, Asthma, ASD, ADHD, and Loss of IQ Point have been systematically valued and presented in one place. This is an important addition to the body of health-related valuation literature as the costs associated with the six outcomes are substantial. Reducing exposures that are associated with these outcomes is therefore beneficial from both a public health and economic point of view.

As one example of how our per-case monetary estimates can be applied in case studies, we examine the potential economic benefits of avoided preterm births possible from reduction in anthropogenic PM_2.5 in the United States. Globally, an estimated 2.7 million out of 15 million (18%) PTBs every year are associated with anthropogenic PM_2.5 [36]. In the U.S., over 380,000 babies are born preterm every year; 18% of that number is around 68,700 [37]. If we use the per-case monetary estimate from the IOM that is applicable to any PTB case, the cost of these PTBs attributable to anthropogenic PM_2.5 exceeds 4.8 billion dollars. Moreover, if the number of PM_2.5-attributable cases of PTB in the United States were reduced by just 1%, this would result in approximately 267 million dollars of benefits. As discussed below, these estimates are likely underestimates of the true costs of PM_2.5 for many reasons, including that they only address the lifetime costs of four disabling conditions associated with PTB when there are many more long-term morbidities associated with that outcome [16].

Our per-case monetary estimates can also be applied to other case studies of fossil fuel combustion emissions reductions. For example, Casey et al. in a study of retirements of coal and oil power plants in California showed a reduction in preterm birth from 7.0% to 5.1% in the area proximate to these power plants (0–5 km) [38]. This translated to 85 fewer PTBs associated with these power plant closures, with an estimated avoided cost of 5.96 million dollars.

The per-case monetary estimates presented here are important additions to current unit values in the BenMAP-CE system. For example, chronic asthma based on willingness to pay (WTP) for adults ages 27–99 is valued at a total sum of $53,607 when discounted over the lifecourse at a rate of 3% [5]. Our per-case monetary estimate that includes childhood asthma that persists into adulthood, also discounted at 3%, is $91,954. This shows nearly an additional $40,000 included in the cost of lifetime asthma when childhood is considered.

The per-case estimates presented here are likely to be underestimates for a number of reasons. The first is the reliance on COI: as noted by the EPA Guidelines for Preparing Economic Analyses, the COI method generally does not attempt to measure the loss in utility due to pain and suffering [8]. Because such costs are often not included in COI values, the WTP value for avoiding a lifetime of asthma could be much larger than the COI value, widening the gap between the current BenMAP-CE value and the value that includes childhood incidence. Further, as noted above, a number of costs were not factored into our selected per-case monetary estimates. For example, PTB is associated with comorbidities related to neurosensory and cognitive disability, in addition to impacts on the employment behaviors of parents [39]. These long-term costs to both the individual and the parent were not included. Moreover, adverse health outcomes may not necessarily occur in isolation of one another. For example, in children with ADHD, comorbid asthma is associated with increased behavioral and internalizing symptoms [40]. Such co-occurring adverse health outcomes may result in increased personal and societal economic impacts.

More recently, it has been proposed that early-life health shocks influence not only the individual’s lifecourse, but also subsequent generations’ lifetime trajectories. Currie discusses the possibility that poor health in childhood is an important mechanism for intergenerational transmission of educational and economic status [41]. This would create a cyclical process whereby parents with low socioeconomic status have children with poorer health, these children subsequently have poorer education and economic outcomes, and their future children are more likely to experience the same poor health of earlier generations. The per-case monetary estimates in this paper did not attempt to include these intergenerational costs.

Another consideration is that all populations are not equally susceptible to environmental exposures or adverse health outcomes, and either their C-R functions or their health costs may differ. While the per-case monetary estimates presented here are summary values meant to be applicable to any individual, there are disparities that should not be ignored in practice or in policy. For example, asthma prevalence rates in the U.S. are highest among persons of multiple race (14.1%), followed by persons of black race (11.2%) and American Indian or Alaska Native race (9.4%); white persons had a prevalence rate of 7.7% [42]. Another example of the disparity in health outcomes can be seen with PTB. In 2017, the PTB rate among non-Hispanic black women (14%) was about 50% higher than the PTB rate among non-Hispanic white women (9%) [13]. Miranda et al. argue that the roots of these PTB disparities are a combination of host factors (e.g., maternal obesity, maternal comorbidities, and genetic vulnerabilities), social factors (e.g., inadequate access to health care, high unemployment, and high poverty rates), and environmental factors (e.g., poor air quality, metals exposure, and pesticide use) [43]. This example highlights the complex etiology of these outcomes, as well as the environmental justice component of exposure reduction.

Full lifetime costs and the cost disparities between groups should be the focus of further studies. However, we believe that our review of per-case monetary estimates can be useful in such studies and that our methodology can serve as a template when considering additional outcomes and settings.

Since our focus has been on air pollution exposure, we emphasized the important policy implications that the per-case monetary estimates have in case studies related to the reduction of fossil fuel combustion, such as the Regional Greenhouse Gas Initiative (RGGI) and the Transportation and Climate Initiative (TCI). However, the per-case monetary estimates can also be applied to other environmental policies.

V. Conclusion

To our knowledge, this is the first time that the child-specific health outcomes of preterm birth, low birth weight, asthma, autism spectrum disorder, attention-deficit/hyperactivity disorder, and IQ reduction have been systematically valued and presented together. This is an important addition to the body of health-related valuation literature as these outcomes have substantial economic costs that are not considered in most assessments of the benefits of air pollution and climate mitigation policies. This review shows substantial avoided costs associated with prevention of adverse children’s health and developmental outcomes, ranging from $23,573 for a case of childhood asthma not persisting into adulthood to $3,109,096 for an ASD case with a concurrent intellectual disability. In general, however, the available per-case estimates presented here did not incorporate the broad societal and long-term costs and are likely underestimates. Although our context has been air pollution and climate policies, the per-case monetary estimates presented here can serve as a resource for researchers, policy-makers and other stakeholders in assessing benefits to children’s health of these and other policies to address environmental threats. Fuller assessments of health benefits to children and their corresponding economic gains will promote environmental policies that maximize protection of this most vulnerable group.