The “Outdoor Air and Health Atlas”: Methods for estimating the health impact of air pollution and their application on a national scale

Introduction

According to the Global Burden of Disease Study, in 2023, outdoor particulate (PM2.5) air pollution accounted for 4.9 million risk-attributable deaths (95%CI 3.9-6.1 deaths) and 123.5 million attributable Disability-Adjusted Life Years (DALYs) (95%CI 101:4-147.4), making it the leading risk factor for mortality, and the second one for associated DALYs, after low birthweight.¹ Adverse health outcomes include premature deaths due to cardiovascular disease (including stroke and ischaemic coronary artery disease), respiratory disease and lung and bronchial neoplasms.^2,3 Additionally, emerging evidence suggests links to metabolic conditions such as type 2 diabetes, reproductive outcomes, and neurological and neuropsychiatric disorders.^4-9

Normative authorities such as the World Health Organization (WHO) and countries within the European Union (EU) have recognised the significant health burden of outdoor air pollution and have therefore implemented monitoring systems and progressively stricter regulations on particulate matter, including limits based on annual mean concentrations and daily threshold values.^10-13

Air pollution shows a spatially heterogeneous distribution, reflecting the interplay of multiple emission sources and environmental pressures – predominantly anthropogenic in origin – including urbanisation, industrial activities, livestock farming, agriculture, biomass burning, and the presence of transport and logistics hubs and infrastructure. This great deal of sources is responsible of varying effects on air quality. 

In 2024, for the first time, all air quality monitoring stations belonging to the network of the Italian Institute for Environmental Protection and Research (ISPRA) registered yearly mean concentrations of PM_2.5 not exceeding the threshold 25 μg/m³,¹⁴ which is the current limit value set by national legislation^11-13; on the other one hand, 70% of the monitoring stations measured a concentration level above 10 μg/m³, that is the threshold of the incoming new EU Directive, and almost all stations (312 on 319) exceeded the WHO recommended value of 5 μg/m³ ^10-14.

Nearly the total Italian population (99.7%) remains exposed to average PM_2.5 concentrations exceeding the WHO guideline value (5 μg/m³).15 Possible download trends can be observed considering that the population-weighted exposure nationwide decreased from 15 μg/m³ in 2016 to 12 μg/m³ in 2024, with a proportion of the population exposed to PM_2.5 levels above the threshold established by the new EU Directive (10 μg/m³) from 93% in 2016 to 83% in 2024.¹⁵ 

Regarding NO₂, annual mean concentrations in 2024 were 17% lower than the average for the preceding decade (2014-2023) and a clear decrease was observed at 89% of monitoring stations. In 2024, annual mean concentrations exceeding current national limit value of 40 μg/m³ ^11-13 were observed just at 1.3%;16 the limit value introduced by the new EU Directive (20 μg/m³) was respected by 66% of monitoring stations, while only 16% of them reported concentrations below the WHO recommended value of 10 μg/m³.¹⁶

In 2024, the 77% of the population was exposed to NO₂ concentrations above the WHO guideline value (10 μg/m³), with an average population-weighted exposure of 11 μg/m³. Between 2016 and 2024, exposure levels showed an overall reduction, with the proportion of the population exposed to NO₂ levels exceeding 20 μg/m³ declined from 55% in 2016 to 31% in 2024.¹⁷

Efforts to improve air quality nationwide led to the establishment of a consortium of participating regions and institutions involved in the ‘ATLAS – Outdoor air and health: an integrated atlas to support decisions and research’ project (Air and Health Atlas).

The Air and Health Atlas, funded by The Italian Ministry of Health, is set to provide an integration of routinely measured data and model-derived data on environmental, sociodemographic and health aspects of the population of interest, and will elaborate them with a set of consolidated analytical methodologies to produce methodologically accurate estimates of the current impact of air pollution on human health and how pollution reducing policies could diminish this impact in the future.

Complete information about population exposure, health outcomes, interventions, vulnerable populations, knowledge gaps, and emerging research challenges (such as unconventional pollutants, non-consolidated health outcomes, and environmental justice issues) have to be effectively delivered in appropriate, context-relevant, and correctly interpretable way to inform both public and decision makers about the effects of air quality strategies oriented to the mitigation of health outcomes correlated to air pollution.

Based on well-established methods for exposure estimation, concentration-response functions (CRFs), counterfactual scenario selection, and attributable case calculation, the Atlas project reviewed and integrated the methodological standards and best practices currently recommended by leading environmental and public health agencies, including the European Environment Agency (EEA)^18,19 and the WHO^10,20,21.

This paper describes the methodology developed for the Atlas to produce a national-level burden of disease estimation for selected health outcomes, detailing the procedures adopted for data collection, selection, processing, and presentation across key analytical domains, including population characteristics, environmental measures, exposure estimation, and health outcome identification.

Materials and methods

This section outlines the methodology used to select, retrieve, and elaborate data included in the Atlas.

The health outcomes were selected after a literature review,^2,3,22 consecutive Delphi rounds among project participants and relevant stakeholders, and finally with public discussion in plenary sessions.

For a detailed description of the whole process, see Davoli et al.²³ In brief, the three main criteria for including a health outcome into the study project were: 1. certainty of evidence for the concentration-response function linking air pollution to the health outcome; 2. public relevance of the health outcome; 3. availability of nation-wide data sources supporting a robust evaluation.

Similarly, for the pollutants to be included a literature review informed an initial list that has been elaborated by the expert panel through Delphi rounds and discussed in plenary sessions. The criteria were the availability of an accurate information source for the whole country, the certainty and strength of association with health outcomes, and the suitability for being mapped.

Similarly, the selection of pollutants was carried out through a comprehensive literature review that produced a preliminary list of candidates that was double-checked by the components of an expert panel through consecutive Delphi rounds and finally discussed in public plenary sessions. The selection criteria for inclusion of pollutants were the availability of accurate nation-wide information sources, including an accurate spatiotemporal model, the certainty and the relative strength of association with the health outcomes being studied, and the possibility of being represented on a geographical map.

The methodology overview is described in the following sections and summarized in Table 1.

Table 2 presents an overview of temporal structure of data with an example using 2019 as the index year.

Population

National population data was gathered from the Italian National Statistical Institute (Istat) database (available at https://demo.istat.it/).

For each municipality, age- and gender-disaggregated population over 30 years of age for the years 2002-2024 was retrieved. For each year, resident population on January 1st was provided.

Additionally, to perform population-weighted exposure computation, publicly available population counts for each census block were retrieved (territorial base of reference: 2021 Census block)^24,25 from Istat 2021 nationwide Permanent General Census of Population and Housing Survey.

For the 2021 Population Census Survey, 756,376 Census blocks were defined (average population ± SD: Pi = 78 ± 174;

average area Ai = 0.40 km² ± 1.65 km²; median Pi = 7, PiIQR = 67; median area  Ai = 0.019 km², AiIQR = 0.063 km²),

of which 451,300 with at least one inhabitant 

(Pi = 131 ± 209, avarage area Ai = 0.49 km² ± 1.84 km²; median Pi = 46, PiIQR = 141; median area Ai = 0.03 km², AiIQR = 0.09km²).

Environmental data

The entire Italian territory was subdivided into 307,635 1 km x 1 km grid cells. For each cell, annual mean concentration values for PM_2.5 and NO₂ were estimated for the years 2016-2024.

Data for the distribution of environmental concentration of PM_2.5 and NO₂ were produced using an integrated approach that combines measurements assessed through the Italian Institute for Environmental Protection and Research (ISPRA) air quality monitoring station network, satellite data and spatial predictors, through a machine-learning modelling approach (random forest, RF). The RF model is illustrated in Stafoggia et al.26 Briefly, the model is trained on daily measurements and uses multiple spatial (land use, population density, road networks, elevation, impervious surfaces, etc.) and spatiotemporal covariates (dispersion models, satellite retrievals on aerosol optical depth, meteorological variables, vegetation indices) to predict observed concentrations. Once the model is optimized (through cross-validation techniques) it predicts daily mean concentrations of PM_2.5 and NO² in all 307,635 grid cells and days.

Population-weighted exposure

Population-weighted exposure (PWE), a proxy of the average population exposure to pollutant concentrations, was calculated by the overlay of modelled atmospheric pollutant concentrations and residential population density.

Pollutant concentrations were provided as gridded estimation at 1 km × 1 km spatial resolution, whilst population density at the finest spatial granularity were provided by Istat at census block level. The main assumption for PWE calculation is that the residential population is uniformly distributed within each census block.

In order to assign PWE estimate to a given administrative unit (e.g., municipality), the pollutant concentration [C]_l was first calculated for each census block by computing an area-weighted average of the pollutant concentration on the overlapping grid cells (Figure 1).

Technically, the pollutant concentration within each grid cell [C]_l was weighted by the proportion of the area falling within the census block area (Ai).

[1]

To determine the PWE for the study area of interest (i.e., k area containing all k census blocks), we calculated the weighted mean of pollutant concentrations across all k Istat census blocks. The concentration at each block ([C]_l) was then weighted by its corresponding resident population (Pi), ensuring the final estimate reflects the actual human exposure within the administrative boundaries of interest.

[2]

where:

• P_tot is the total resident population within the administrative area of interest;
• [C]_l is the pollutant concentration within each Census block I;
• P_i  is the corresponding resident population.

The output represents the weighted average individual exposure to pollutant attributable to residents within the administrative area of interest.

In the Air and Health Atlas, population-weighted exposure was calculated for each year using total resident population on January 1st, 2021, as described in the section Population, along with Census blocks shape defined by Istat 2021 (Permanent) Population and Housing Census.

PWE was then aggregate for each municipality (local administrative unit, LAU2) of the Italian national territory, as well as for wider administrative levels, such as regions (NUTS2).

Mortality data

Selected outcomes, identified through International Classification of Diseases 10th revision (ICD-10) codes, were:

• Mortality for all natural causes (nontraumatic, nonexternal): ICD-10 codes A00-R99;
• Mortality for all cardiovascular diseases (CVD): ICD-10 codes I00-I99;
• Mortality for all respiratory diseases: ICD-10 codes J00-J99;
• Mortality for trachea, bronchus and lung cancer: ICD-10 codes C33-C34.

Mortality data were retrieved from Istat database of death certificates (‘Indagine su decessi e cause di morte’). Retrieved information were year of death, province, municipality, gender (male or female), age class (5-year classes except for those related to first year of life, 1-4 years and centenarians: 0, 1-4, … 95-99, 100+), initial cause of death as ICD-10 code, from 2016 to 2023.

Mortality data were acquired at the province level, except for municipalities with ≥500,000 residents (Rome, Milan, Naples, Turin, Palermo, Genoa).

For causes of death at risk of identification (i.e., individual identity can be reconducted to, when variable triangulation could result in groups of ≤3 cases),²⁷ cause of death was overwritten as “undefined”.

Annual mortality rates were elaborated as crude values to obtain reliable estimation of expected events and attributable burden.

Attributable cases

Attributable cases (AC) were used to quantify the health impact of air pollution exposure. They represent the share of the total burden of a disease associated with the exposure that can be potentially avoided by removing specific exposure factors. AC are usually calculated as the delta of cases attributable to a counterfactual scenario where baseline exposure was reduced according to some interventions (e.g., application of a new normative guideline).

The formula used to estimate AC is grounded in the standard epidemiological assumptions also used in a well-established health-impact assessment tool, such as the WHO AirQ+ software, which quantifies the number of cases attributable to air pollution based on concentration–response relationships derived from epidemiological evidence.^28,29

AC were calculated at the municipality level by estimating the annual number of deaths attributable to a specific pollutant (e.g., PM_2.5). For each municipality i, AC were derived as the difference between the observed number of deaths estimated under current exposure conditions and the number of deaths that would have occurred under a predefined counterfactual scenario (i.e., in the absence of, or at a reference level of, pollutant exposure).

where:

• T is the counterfactual scenario (pollutant concentration);
• y_i refers to expected deaths in that municipality;
• y_i0 refers to expected deaths in that municipality for the counterfactual scenario;
• x_i is the PWE for that municipality;
• β_i = ln(RR), where RR is the relative risk for 10 μg/m³ of a specific pollutant.

Relative risks (RRs) were derived from concentration-response functions (CRFs) describing the association between pollutant exposure intensity and health outcomes. However, CRF-based excess fractions do not equate to aetiologic fractions, as they cannot distinguish cases where exposure effectively caused the event from those that would have occurred regardless of exposure variation.³⁰

Table 3 reports the last update of CRF for PM_2.5 and NO₂ related to different health outcomes used into the Atlas.²²

Attributable fraction (AF%), defined as proportion of cases in a population that can be attributed to a specific risk factor, is then calculated dividing the AC for the total expected cases in the same population.

For a mathematical treatment of the formula of the attributable cases, refer to Appendix 1 (see online Supplementary Materials). In this paper, we provided attributable death counts relative to population-weighted exposure to PM_2.5² and NO₂³ for each municipality and province.

For each year and each pollutant, PWE and crude outcome rate for the previous year were used to determine year-specific and pollutant-specific attributable events. The rationale for this is ensuring that estimated outcomes (which are aggregated by year) were always temporally subsequent to attributed exposure (Table 2).

For the Atlas project, counterfactual scenarios were selected according to WHO recommended yearly mean concentration of 5 μg/m³.¹⁰ Similarly, for NO₂, selected counterfactual scenarios were WHO’s 10 μg/m³.¹⁰

The computation of attributable cases was defined such that, when exposure reached or fell below the counterfactual level, the attributable count could not be negative (i.e., below the counterfactual, RR = 1 and AD = 0).

Results were also stratified by outcome and year and were presented for each pollutant for the 2017-2024 period.

The Atlas also provides estimates of future year-specific PWE and ACs under different scenarios, including the EU scenario based on the EU new Directive. For future impact assessments, year-specific PWE and ACs were estimated using projections of observed data. Methodological details of the forecasting approach are presented in Appendix 2 (see online Supplementary Materials).

Results

Exposure estimation and burden assessment for attributable mortality for natural causes to PM_2.5 and NO₂, for the year 2019, by region (NUTS2) and geographical partition (NUTS1) are presented in Table 4. In Italy, 59,343 deaths (9.9% of total mortality) were attributable to PM2.5 exposure. The attributable fraction was highest in North-Western Italy (12.8%) and lowest in the Islands (6.0%). A total of 31,189 deaths (5.2% of total mortality) were attributable to NO2 exposure, with attributable fractions ranging from 7.8% in North-West Italy to 2.1% in the Islands.

Figure 2 shows mortality attributable fraction (AF%) for natural causes and PWE for PM2.5 in 2019. The same representation for NO₂ is reported in Figure 3.

Discussion

The Air and Health Atlas has been designed as an online tool that makes health impacts of air pollution easy to explore, visualize, and compare between different scenarios, with the aim to support decision-makers and facilitate communication to the public. The default counterfactual scenarios displayed by the Atlas refer to WHO recommend health-based values for PM_2.5 and NO₂. The Atlas provides estimates of future year-specific PWE and ACs under different scenarios, providing policy-makers and citizens with an overview of potentially avoidable deaths linked to pollutant concentrations exceeding regulatory limits.

The impact assessment at the basis of the Air and Health Atlas is conducted according to a consolidated, scientifically sound methodology,^28,29 making outputs fully comparable with those delivered by supra-national institutions such as the WHO and the EEA. Moreover, thanks to the use of homogeneous and standardised exposure and outcomes data available at the national level, the Atlas allows a national coverage and reliable comparisons between different territories and administrative units, with a high spatial resolution at the municipal level.

Assessment of population exposure to air pollutants and its spatiotemporal distribution represents a fundamental first step in environmental epidemiology when investigating associations between environmental factors and health outcomes. Consequently, considerable effort has been devoted to developing effective methods to assess population exposure at increasingly fine spatial and temporal resolutions.³¹ Within this framework, the calculation of population-weighted exposure represents one of the more practical and widely used methodologies.

One of the first instances of use of population-weighted exposure (PWE) was a nationwide Finnish cohort study that assessed the association between radiation exposure after Chernobyl event fallout and childhood leukaemia incidence.³²

An assumption intrinsic to this calculation of PWE is that, within any section i considered, population is evenly distributed in space, taking into account a certain proportion of the section will likely mean considering an analogous share of the population that resides in that section (e.g., A_ji /A_i ≈ P_ji /P_i).

As long as this assumption is satisfactorily met, the two methods of calculating it are effectively equivalent and produce consistent and informative estimates of exposure. Another assumption intrinsic to this calculation is that exposure occurs exclusively at the place of residence. An alternative approach, known as mobility-integrated exposure assessment, has been proposed to address this limitation by accounting for exposure not only at home, but also in the workplace, during commuting and leisure activities. Evidence from initial studies suggest a high correlation between residential-only (static) estimates – reflecting the fact that participants spent approximately 58% of their time at home – and mobility-integrated (dynamic) estimates. As a result, any bias introduced by relying on residential exposure alone is likely to be modest and largely non-differential, tending toward the null.^33-38

The choice of concentration-response functions (CRF) is crucial to the assessment of disease burden related to an exposure. For the Atlas, the more up-to-date CRF made available by the WHO for the European Region²² were chosen. These CRFs represent the best choice for Italian territory, as they are recommended by the WHO for the European Region. The WHO itself provides the ranges of applicability for these functions, within which the PWE values calculated for Italy fall.

In the Atlas, the assessment of the health burden of air pollution is limited to total natural, circulatory, and respiratory mortality and, therefore, probably underestimated the true burden. Growing evidence shows that air pollution also contributes to morbidity outcomes, such as hospital admissions for respiratory and cardiovascular causes, exacerbations of respiratory and neuropsychiatric disorders and onset of chronic disorders such as asthma, COPD, type 2 diabetes, and dementia.^5,7,39 All these morbidities can significantly impair health and quality of life, leading to disability and imposing high direct and indirect economic costs to individuals and to society as a whole. Unfortunately, information on morbidity outcomes – such as hospitalization rates, drug prescriptions, etc. – was not yet available for the entire national territory with the required level of standardisation; therefore, it could not be included in the health burden estimates.

We are working on the theoretical and logistical basis to expand the Atlas to morbidity by including data on attributable hospitalization in the next future. Direct and indirect economic cost will be also considered (ongoing and registered in PROSPERO).⁴⁰

Mortality rates were available at the province level; then, we attributed that rate to each municipality of the province. Even though this approach may have missed local differences in disease occurrence, we believe it assures more stable and reliable estimates by attenuating random spatial variability. The only exception was made for cities with more than 500,000 inhabitants, where municipal mortality rates were used.

Attributable burden for years 2021 through 2023 were calculated considering mortality for year 2019, because mortality in the 2020-2022 period were heavily influenced by the SARS-CoV-2 pandemic and represent outliers in identified mortality trends.

Whenever COVID-19 or a COVID-19-associated condition is defined as the initial cause of death (Istat death certificate, question 4 “cause of death”, part 1, line 1), an ICD-10 code from U07 to U10 is assigned and was likely excluded by our data retrieval step. Nonetheless, initial misclassification due to the novel institution by WHO of emergency ICD codes for SARS-CoV-2 related illnesses and harvesting effect (mortality displacement due to acute mortality rate increases) have been considered likely as well, thus leading to precautionary exclusion of potentially biased data.

According to Istat, in 2023, mortality rates returned to pre-pandemic levels.⁴¹

On the side of the exposure, the Atlas focuses only on two pollutants: PM_2.5 and NO₂. The discussion led to excluding PM₁₀ to avoid redundancy with PM_2.5.

The Atlas do not consider the impact of ozone; there are essentially two main reasons. Firstly, ozone is a highly reactive and seasonal pollutant, with clear scientific evidence regarding acute health impacts. Only recently (and still with some uncertainties) have risk functions for long-term exposure been established, specifically related to respiratory mortality.²²

Secondly, SOMO³⁵ (the sum of daily maximum 8-hour ozone means over 35 ppb) is the metric recommended by the WHO for assessing the long-term health impacts of ozone. By definition, it measures the cumulative seasonal ozone exposure exceeding the 35 ppb threshold. Being a seasonal, threshold-based metric, comparing it directly to the annual, continuous-exposure impacts of other pollutants in the Atlas can easily create confusion for the final users.

Despite the inclusion of only two pollutants, the issue of double-counting the deaths attributable to air pollution when estimating separately those attributable to PM_2.5 and those to NO₂ remain a clear communication challenge, particularly for an online tool accessible also to the general public. The final decision was not to calculate combined effects of both pollutants, but to explain with a text warning that just summing up the attributable deaths to PM_2.5 and NO₂ would lead to overestimating, but that not all the deaths attributable to NO₂ are included in those attributable to PM_2.5.

The issue of potential double-counting when separately estimating cases attributable to PM_2.5 and NO₂ remains a clear communication challenge, particularly for an online tool accessible to non-experts. Despite this, the final decision was not to calculate the combined effects of both pollutants. Instead, the online tool will warn users that summing the attributable cases for both PM_2.5 and NO₂ would lead to a severe overestimation, while also clarifying that not all cases attributable to NO₂ are captured within those for PM_2.5.

We used a random forest model to estimate the distribution of the pollutants. This method is different from that used by most environment protection agencies in Italy, which use models that generally rely on regional scale dispersion models. The choice was due to the intrinsic heterogeneity in the available information about emissions and sources across different regions (e.g., the lack of standardized estimates covering the entire national territory).

The performance of the random forest model used to estimate pollutant concentrations is guaranteed by the model validation phases, particularly the cross-validation of model estimates against measurements from air quality monitoring stations. Better performance was achieved for PM_2.5 than for NO₂.²⁶ It should nevertheless be considered that, based on the location criteria of the monitoring stations (whose distribution follows population density), the areas with presumably higher model estimation uncertainty coincide with the less populated zones. The PWE, which represents the average exposure of the entire resident population, is therefore constructed using the most reliable data.

Conclusions

The Atlas maps show a strongly uneven distribution of health impact in Italy, with the largest burden concentrated in the Po valley and in some other metropolitan areas. These maps are easily intelligible and highly communicative. By gathering information on exposure and health impact in a unique interactive web platform, the Atlas provides a powerful tool to improve public awareness, support policy decisions, and promote behavioural changes among citizens. Retrospective and prospective assessments provide a clear picture of progress already made in the reduction of exposure and of impacts potentially avoidable in the future, thereby limiting unjustified alarm and fostering a constructive approach to the issue of air pollution.

Finally, despite the fact that the Atlas feed has been highly automated, its updating and evolution to be functional on the web will require dedicated funding beyond the end of the project, to ensure its long-term sustainability.

Conflicts of interests: none declared.

Funding: project carried out with technical support and funded by the Italian Ministry of Health – PNC PREV-A-2022-12376981 Outdoor Air and Health: An Integrated Atlas to Support Decision-Making and Research.

References

1. GBD 2023 Disease and Injury and Risk Factor Collaborators. Burden of 375 diseases and injuries, risk-attributable burden of 88 risk factors, and healthy life expectancy in 204 countries and territories, including 660 subnational locations, 1990-2023: a systematic analysis for the global burden of disease study 2023. Lancet 2025;406(10513):1873-922. doi: 10.1016/S0140-6736(25)01637-X
2. Orellano P, Kasdagli MI, Pérez Velasco R, Samoli E. Long-term exposure to particulate matter and mortality: an update of the WHO Global Air Quality Guidelines systematic review and meta-analysis. Int J Public Health 2024;69:1607683. doi: 10.3389/ijph.2024.1607683
3. Kasdagli MI, Orellano P, Pérez Velasco R, Samoli E. Long-term exposure to nitrogen dioxide and ozone and mortality: update of the WHO air quality guidelines systematic review and meta-analysis. Int J Public Health 2024;69:1607676. doi: 10.3389/ijph.2024.1607676
4. U.S. Environmental Protection Agency. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, Dec 2019). EPA/600/R-19/188. Washington, DC: US EPA; 2019. Available from: https://assessments.epa.gov/isa/document/&deid%3D347534
5. Yang BY, Fan S, Thiering E, et al. Ambient air pollution and diabetes: A systematic review and meta-analysis. Environ Res 2020;180:108817. doi: 10.1016/j.envres.2019.108817
6. Ghosh R, Causey K, Burkat K, Wozniak S, Cohen A, Brauer M. Ambient and household PM2.5 pollution and adverse perinatal outcomes: A meta-regression and analysis of attributable global burden for 204 countries and territories. PLoS Med 2021;18(9):e1003718. doi: 10.1371/journal.pmed.1003718
7. Best Rogowski CB, Bredell C, Shi Y, et al. Long-term air pollution exposure and incident dementia: a systematic review and meta-analysis. Lancet Planet Health 2025;9(7):101266. doi: 10.1016/S2542-5196(25)00118-4
8. Lin LZ, Zhan XL, Jin CY, Liang JH, Jing J, Dong GH. The epidemiological evidence linking exposure to ambient particulate matter with neurodevelopmental disorders: A systematic review and meta-analysis. Environ Res 2022;209:112876. doi: 10.1016/j.envres.2022.112876
9. Han C, Lu Y, Cheng H, Wang C, Chan P. The impact of long-term exposure to ambient air pollution and second-hand smoke on the onset of Parkinson disease: a review and meta-analysis. Public Health 2020;179:100-10. doi: 10.1016/j.puhe.2019.09.020
10. World Health Organization. WHO Global Air Quality Guidelines: Particulate Matter (PM2.5 and PM10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide. Geneva: WHO; 2021. Available from: https://www.who.int/publications/i/item/9789240034228
11. European Parliament and Council. Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on Ambient Air Quality and Cleaner Air for Europe. 2008. Available from: https://eur-lex.europa.eu/eli/dir/2008/50/oj/eng
12. European Parliament and Council. Directive (EU) 2024/2881 of the European Parliament and of the Council of 23 October 2024 on Ambient Air Quality and Cleaner Air for Europe (Recast). 2024. Available from: https://eur-lex.europa.eu/eli/dir/2024/2881/oj/eng
13. Repubblica Italiana. Law Decree No.155 of 13.08.2010. Available from: https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:decreto.legislativo:2010-08-13;155
14. Italian National System for Environmental Protection (SNPA). PM2,5: la situazione nel 2024. Available from: https://www.snpambiente.it/pm25-la-situazione-nel-2024/
15. Italian Institute for Environmental Protection and Research (ISPRA). Esposizione della popolazione agli inquinanti atmosferici outdoor – PM2.5. Rome: ISPRA; 2025. Available from: https://indicatoriambientali.isprambiente.it/it/qualita-dellaria/esposizione-della-popolazione-agli-inquinanti-atmosferici-outdoor-pm2.5
16. Italian National System for Environmental Protection (SNPA). NO2: la situazione nel 2024. Available from: https://www.snpambiente.it/no2-la-situazione-nel-2024/
17. Italian Institute for Environmental Protection and Research (ISPRA). Esposizione della popolazione agli inquinanti atmosferici outdoor – NO2. Rome: ISPRA; 2025. Available from: https://indicatoriambientali.isprambiente.it/it/qualita-dellaria/esposizione-della-popolazione-agli-inquinanti-atmosferici-outdoor-no2
18. Soares J, Horálek J, Ortiz AG, Guerreiro C, Gsella A. EEA/ATNI Report 13/2019: Health Risk Assessment of Air Pollution in Europe. Methodology Description and 2017 Results. 2019. Available from: https://www.eionet.europa.eu/etcs/etc-atni/products/etc-atni-reports/etc-atni-report-13-2019-health-risk-assessment-of-air-pollution-in-europe-methodology-description-and-2017-results
19. Soares J, Plass D, Kienzler S, González Ortiz A, Gsella A, Horálek J. ETC HE Report 2023/7: Version 2; Health Risk Assessment of Air Pollution: Assessing the Environmental Burden of Disease in Europe in 2021. 2024. Available from: https://www.eionet.europa.eu/etcs/etc-he/products/etc-he-products/etc-he-reports/etc-he-report-2023-7-health-risk-assessment-of-air-pollution-assessing-the-environmental-burden-of-disease-in-europe-in-2021
20. World Health Organization. Sustainable Development Goal Indicator 3.9.1: Mortality Attributed to Air Pollution. Geneva: WHO; 2024. Available from: https://iris.who.int/items/3ed76386-f5ef-4807-97eb-a2872c2d6a3f
21. Evangelopoulos D, perez-Velasco R, Walton H, et al. The role of burden of disease assessment in tracking progress towards achieving WHO global air quality guidelines. Int J Public Health 2020;65(8):1455-65. doi: 10.1007/s00038-020-01479-z
22. WHO Regional Office for Europe. Health Risks of Air Pollution in Europe: HRAPIE-2 Project: Updated Guidance on Concentration-Response Functions for Health Risk Assessment of Air Pollution in the WHO European Region. Geneva: WHO; 2025. Available from: https://www.who.int/europe/publications/i/item/9789289062633
23. Davoli E, Mazzoleni E, Ottone M, et al. The co-creation process of the “Outdoor Air and Health Atlas” project: A tool to support decisions and research. Epidemiol Prev 2026;50(4-5) Suppl 1:35-45. doi: 10.19191/EP26.4-5.S1.A952.090
24. Italian National Institute of Stastics. Registro Statistico di Base dei Luoghi. 2024.
25. Repubblica Italiana. Decree of the President of the Italian Republic n.223 of 30.05.1989. Available from: https://www.gazzettaufficiale.it/eli/id/2015/08/14/15G00140/sg
26. Stafoggia M, Cattani G, Ancona C, Ranzi A. Exposure assessment of air pollution in Italy 2016-2019 for future studies on air pollution and COVID-19. Epidemiol Prev 2020;44(5-6) Suppl 2:161-68. doi: 10.19191/EP20.5-6.S2.115
27. Codice di deontologia e di buona condotta per i trattamenti di dati personali per scopi statistici e di ricerca scientifica effettuati nell’ambito del Sistema statistico nazionale. Allegato A3 al Codice in materia di protezione dei dati personali – 12 giugno 2014. Gazzetta Ufficiale No.170 of 24.07.2014.
28. Centre for Environment, Climate Change & Health (BON), Living & Working Environments (LWE). Poland Health Impact Assessment of Air Pollution: Introductory Manual to AirQ+. Copenhagen: WHO Regional Office for Europe; 2020. Available from: https://www.who.int/europe/publications/i/item/WHO-EURO-2020-1557-41308-56210
29. Air Quality, Energy and HEalth (AQE). Health Risk Assessment of Air Pollution. General Principles. Copenhagen: WHO Regional Office for Europe; 2016. Available from: https://www.who.int/publications/i/item/9789289051316
30. Brunekreef B, Miller BG, Hurley JF. The brave new world of lives sacrificed and saved, deaths attributed and avoided. Epidemiology 2007;18(6):785-78. doi: 10.1097/EDE.0b013e3181570d88
31. Caplin A, Ghandehari M, Lim C, Glimcher P, Thurston G. Advancing environmental exposure assessment science to benefit society. Nat Commun 2019;10(1):1236. doi: 10.1038/s41467-019-09155-4
32. Auvinen A, Hakama M, Arvela H, et al. Fallout from Chernobyl and incidence of childhood leukaemia in Finland, 1976-92. BMJ 1994;309(6948):151-54. doi: 10.1136/bmj.309.6948.151
33. Wei L, Kwan MP, Vermeulen R, Helbich M. Measuring environmental exposures in people’s activity space: the need to account for travel modes and exposure decay. J Expo Sci Environ Epidemiol 2023;33(6):954-62. doi: 10.1038/s41370-023-00527-z
34. Wei L, Donaire-Gonzalez D, Helbich M, van Nunen E, Hoek G, Vermeulen RCH. Validity of mobility-based exposure assessment of air pollution: a comparative analysis with home-based exposure assessment. Environ Sci Technol 2024;58(24):10685-95. doi: 10.1021/acs.est.3c10867
35. Ndiaye A, Vienneau D, Flückiger B, et al. Associations between long-term air pollution exposure and mortality and cardiovascular morbidity: a comparison of mobility-integrated and residential-only exposure assessment. Environ Int 2025;198:109387. doi: 10.1016/j.envint.2025.109387
36. Schweizer C, Edwards RD, Bayer-Oglesby L, et al. Indoor time-microenvironment-activity patterns in seven regions of Europe. J Expo Sci Environ Epidemiol 2007;17(2):170-81. doi: 10.1038/sj.jes.7500490
37. de Hoogh K, Flückiger B, Probst-Hemsch N, et al. Comparison of residential and mobility-integrated air pollution exposures from tracking campaigns and agent-based modelling in Switzerland and the Netherlands. J Expo Sci Environ Epidemiol 2026;36(3):469-78. doi: 10.1038/s41370-025-00836-5
38. Evangelopoulos D, Wood D, Barratt B, et al. Exposure measurement error in air-pollution epidemiology and its determinants: results from the MELONS study. Int J Epidemiol 2026;55(1):dyaf214. doi: 10.1093/ije/dyaf214
39. Forastiere F, Spadaro JV, Ancona C, et al. Choices of morbidity outcomes and concentration-response functions for health risk assessment of long-term exposure to air pollution. Environ Epidemiol 2024;8(4):e314. doi: 10.1097/EE9.0000000000000314
40. Forastiere F, Krzyzanowski M, Ancona C, et al. Choices of morbidity outcomes and concentration-response functions for health risk assessment of air pollution: an umbrella review of systematic reviews and meta-analyses. PROSPERO 2023. Available from https://www.crd.york.ac.uk/PROSPERO/view/CRD42023397145
41. Italian National Institute of Statistics. Rapporto SDGs 2024. Informazioni Statistiche per l’agenda 2030 in Italia. Rome: Istat; 2024. Available from: https://www.istat.it/produzione-editoriale/rapporto-sdgs-2024/