The co-creation process of the “Outdoor Air and Health Atlas” project: A tool to support decisions and research

Introduction

According to the World Health Organization (WHO), about 4.2 million premature deaths per year are attributable to outdoor air pollution;¹ this mortality is mainly due to exposure to fine particulate matter, which causes cardiovascular and respiratory diseases and cancers.^2,3 

The European Green Deal refers to WHO’s Global Air Quality Guidelines (AQGs), which provide global guidance on thresholds and limits for key air pollutants that cause health risks. In 2019, 99% of the world’s population lived in areas where air quality guidelines were not met.¹

Italy is still among the countries where air pollution levels exceed the WHO guidelines. Moreover, the European Commission decided to refer Italy to the Court of Justice for failing to respect limit values for PM₁₀ and NO₂ and for failing to take appropriate measures to keep exceedance periods as short as possible.

The Italian government has approved the National Plan of Complementary Investments (PNC) (Legislative Decrees 59/2021 and 101/2021), which complements the interventions of the National Recovery and Resilience Plan (PNRR) with national resources. The Plan promoted a new model of territorial health care through the definition of a new systemic institutional arrangement for prevention in health, environment and climate, in line with an integrated approach (one health) and a holistic vision (planetary health).

This is the background of the project “Outdoor air and health ATLAS”, funded by the Italian Ministry of health and conducted by 22 Units in 11 out of 20 Italian Regions; further 4 units joined the project after funding approval. The main aim of the project is the development of a national interactive, open data Atlas which relates air quality data and health outcomes to increase awareness of the effect of air pollution on health and support policy makers, stakeholders, and citizens in the decision-making processes (Figure 1). The Atlas has been planned to be a permanent infrastructure to exploit the potential of the available information.

To achieve these goals, the Atlas’s contents need to be agreed upon by potential users. Therefore, the project adopted a structured co-creation strategy to define the contents and communication solutions. Co-creation is a process that should ensure participation and sharing of decisions in developing a product or a service. It can be used in the public sector, society, business, and universities,⁴ and it is effective in producing more usable services and in reducing the gap between providers and users.⁵

In the development of the Outdoor Air and Health Atlas, several methods were adopted to involve potential users at different steps (Figure 2). Among the others, the Delphi method integrated with meetings involving the scientific and practice community was adopted to agree on the most technical decisions on contents, sources, and methods that should be used for the Atlas. The Delphi technique is a qualitative research approach used to gain consensus through expert opinion on issues and solutions,⁶ especially in situations of limited information availability⁷. For this reason, it has been frequently used in various scientific disciplines, including health care, medicine, engineering, and environmental studies.⁸

The present work describes the co-creation decision-making process adopted to define the contents of the Atlas, in particular the Delphi round surveys and the plenary discussion aimed at collecting the inputs and reaching agreement between experts and potential professional users about the content (i.e., which outcomes, pollutants, and metrics), and methods (i.e., data sources, assumptions and algorithms) that should be used to build the Atlas.

Methods

The Outdoor Air and Health Atlas

While developing the Atlas, the primary goal was to create a system capable of coordinating the generation, processing, interpretation, and utilization of available data on air quality and health outcomes to support decisions grounded in transparent and trustworthy scientific evidence. This will be reached by realizing Atlas reporting national maps of the distribution of the exposure to the most important air pollutants and quantifying their effects on health. The Atlas should allow for predicting the effects on health outcomes of different scenarios of air pollution reduction.

The project involves 26 partners, 22 included in the initial project submitted to the Ministry of Health and 4 who joined the project later (see Figure S1, online Supplementary Materials) and it is coordinated by the Emilia-Romagna Region (Department of prevention and public health) and the Azienda Unità Sanitaria Locale – Istituto di Ricovero e Cura a Carattere Scientifico (ASL-IRCCS) Unit of Reggio Emilia.

It includes six specific aims, which are strictly related to each other and summarized in Figure 1. Objective 1 is the coordination of the project and aims to establish regional control rooms for interpreting data and results in support of the governance of air pollution impact assessment on human health. Building the national Atlas is the core of the project and is specifically referred as objective 2 in the project structure. All the other objectives are functional to the development and dissemination of the Atlas. Objective 3 focus on the research and development of the system, identifying new outcomes and pollutants to be included in the Atlas in future. Objective 5 seeks to enrich longitudinal studies on the socioeconomic determinants of health, carried out by the Italian National Institute of Statistics (Istat) by integrating environmental exposure data linked to subjects’ place of residence.^9,10 This integration applies to two key cohorts: the first includes all Italian residents recorded in the 2011 Population and Housing Census (59,227,313 individuals), who are being followed up for migration, vital status, and causes of death; the second cohort is drawn from the National Health Interview Survey (editions 2000, 2005, and 2013), each consisting of a nationally representative sample of approximately 120,000 individuals, with detailed information on lifestyle factors, socioeconomic conditions, and physical and mental health status. All three survey editions have been individually linked to health data and are currently being followed up for mortality and hospital admission.^11-13 Objective 4 aims to evaluate potential counterfactual scenarios and analyse exemplary situations in the context of real life (Case studies) to assess the usability and usefulness of the Atlas. Finally, objective 6 focuses on training the potential users and disseminating the Atlas.

The co-creative process

General methodology

A co-creation process was adopted to define the contents of the Atlas. This process includes three main actions to involve different users and stakeholders and to address various aspects of the Atlas:

1. The consultation of the scientific and practice communities of healthcare workers and environmental protection operators, which is the topic of this report and will be detailed below. This consultation aimed at defining the most technical part, including data sources, exposures and outcomes, measures, dose-response functions, and counterfactual scenarios. It used the Delphi method described below. To carry out this consultation, the coordinators administered a Delphi survey through online questionnaires and plenary meetings in order to reach consensus on the most controversial topics, the coordinators conducted a Delphi survey with online questionnaires and plenary meetings to reach an agreement on controversial topics;

2. The involvement of stakeholders and potential users through interviews and focus groups to identify their information needs, improve clarity, select relevant counterfactual scenarios, and define the interactive features of the maps and the most useful summary statistics;

3. The presentation of the prototype Atlas to regional control rooms involved in the project which include local level policy makers and technicians, to identify the most useful outputs and define how the Atlas should evolve and be updated. 

Points 2 and 3 are not the object of this report.

Delphi method

To reach consensus among the experts and decision makers involved in the project, a Delphi method was adopted. Each Delphi round consisted of a survey and a subsequent meeting among the participants, to discuss the results, reframe the questions, and, eventually, add new items (Figure 2). The Delphi survey was concluded in 2 rounds since, during the in-person meeting after the second round, the experts reached an agreement on the relevant questions or decided that it should be discussed in the following project in-person meeting. Google Forms were used for the administration of the survey, as it is a tool that allows anonymised responses and quickly provides a summary of the results. In the first round, held in April 2023, the questionnaire was administered to all 22 working units participating in the objective 2, asking for an individual answer. Participants included public health operators, environmental epidemiology researchers, environmental exposure researchers, environmental protection operators, decision makers involved in environment protection and public health institutions. After the first round, a meeting was carried out in June 2023, where individual answers were analysed and discussed within the objective 2 working group. In the second round, which took place in July 2023, the call was extended to other 4 institutions differently involved in the project (Istat, Regional Environmental Protection Agencies – ARPAs – and two Regions not involved in the objective 2 working group), requiring 1 answer per unit. A plenary session meeting was held in Milan in September 2023 with the participation of all the units of the Atlas project and with the involvement of a consultant of the National Environmental Protection Agency (ISPRA) to discuss critical answers and technical issues. 

The questions administered in the first round were developed by a small group of investigators involved in objective 2 and the coordinators of objectives 1, 3, 4, and 5. The main topics covered by the questions were data sources, territorial disaggregation, possible pollutants and health outcomes (mortality and hospitalization), attributable fractions calculation, and counterfactual pollution reduction scenarios against which to evaluate the possible benefits of policies. A list of pollutants to include in the Delphi was selected a-priori based on previous studies recently conducted in Italy on the topic, such as the projects  ‘European Study of Cohorts for Air Pollution Effects’ (ESCAPE),^14,15 ‘Effects of Low-Level Air Pollution: A Study in Europe’ (ELAPSE)¹⁶, ‘Big Data in Epidemiologia ambiEntale e occuPazionale’ (BEEP)¹⁷, and ‘Uso di BIG data per la valutazione degli Effetti sanitari acuti e cronici dell’inquinamento atmosferico nella Popolazione Italiana’ (BIGEPI)¹⁸. 

The pre-selected pollutants were PM_2.5, PM₁₀, O₃, and NO₂, as these are the pollutants for which there is a dose-response function with a high level of certainty for consolidated outcomes (i.e., easily detectable throughout the territory). For each proposed pollutant in the Delphi, participants were asked to rate the importance to include them in the Atlas. The range is based on a minimum defined as ‘low importance’ and a maximum defined as ‘high importance’.

Mortality and hospitalization outcomes were pre-selected a-priori considering the dose-response function criteria adopted for exposure. The pre-selected outcomes were initially organized in a matrix and each region included in the project was asked to complete a checklist based on the data availability. The completed checklists were subsequently reviewed to assess outcomes for which data sources were available in all Regions to be included in the first Delphi round. The chosen outcomes were then cross-verified with findings from previous major studies.^16,19-21 For each proposed  health outcome, Delphi participants were asked to express a personal judgement about the certainty of the causal link (sufficient/possible) between the pollutant and health outcome, the public health significance (high/low), the accuracy of available information in routine databases (acceptable/non-acceptable), and a synthesis of the opportunity to include it among the maps of the Atlas (yes/no).

The need to further discuss specific topics during the meetings was evaluated according to the level of agreement in the answers. If the majority opted for the same answer, the agreement was considered to have been reached. In case of discordant answers, it was considered appropriate to discuss the topic during the post-survey meetings.  

Three further in-person meetings were held involving all the working units of the Atlas project and external advisors to gather feedback on the Atlas plan. The first meeting took place in Rome in March 2024, focusing on the choice of consolidated outcomes and indicators to include in the Atlas, the level of pollutants to consider for calculating the proportion of attributable deaths, and the counterfactual scenarios to define impact assessments. In the second meeting, held in Riccione in April 2024, the discussion focused on how to present the maps in the Atlas and how to communicate the results. In the discussion, experts in infographics, communication, and co-creation were involved. Finally, in Palermo, in September 2024, the selected outcomes were validated with the results of a recent systematic review about the impact of air pollution on non-accidental mortality in a session where the authors of the review were invited.^22,23 The group considered other outcomes such as asthma in children and adults, for which the systematic review assessed a dose response function, and decided to conduct a detailed analysis of the available sources information to evaluate their consistency with those used in original studies and decide whether they could be included in the Atlas. 

Results

The two Delphi rounds reached a participation rate of 73% (16 answers on 22 working units) and 85% (22 answers on 26 working units and stakeholders), respectively, including project working units and external experts.

Data sources

Concerning the outcomes, the decision was to use aggregated data rather than anonymised individual data to mitigate privacy and security concerns. Data on all-cause mortality and cause-specific mortality will be provided by Istat, while hospitalization data (hospital discharge records, HDR) will be provided by the Ministry of Health.

To reduce data granularity and address the issue of random fluctuations, the application of smoothing techniques to the maps was recommended. The experts’ suggestions focused on Bayesian models in general (21%) or more specifically on the Kernel, Besag-York-Mollié (BYM) technique (32%).²⁴ Following the plenary session, the working units solved the smoothing issue by using provincial-level data except for municipalities with a population of more than 500,000 overcoming the need to use one of the two models described above.

Reliable spatiotemporal estimates of particle exposure, including PM₁₀, PM_2.5, and PM_10-2.5 (coarse particles, i.e., those between 2.5 and 10 microns), were developed for Italy on a 1-km² grid covering the years 2013 to 2015 using the Random Forest (RF) machine learning approach.²⁵ These estimates were later updated to cover the years from 2016 to 2019 and expanded to include NO₂ and O3.²⁶ Since 2020, the estimates have been updated annually using the same modelling approach. The used predictors include parameters derived from atmospheric ensemble model outputs, satellite atmospheric observations, land use data, and meteorological data. 

Exposures and outcomes

The survey highlighted a high degree of agreement to include PM_2.5 and PM₁₀ in the Atlas, while there were discrepancies about the inclusion of NO₂ and O₃ (Figure 3). During the post-survey meetings, the group decided to include the attributable fraction estimates only for PM_2.5 and NO₂. PM₁₀ was excluded to avoid wrong interpretations, i.e., double counting, due to providing both attributable deaths to PM_2.5 and PM₁₀. The exclusion of O₃ was mostly due to the difficulties in rendering, with a unique intuitive visualization, PM_2.5 and NO₂ exposure, and the seasonal exposure to O₃.

Short-term and long-term outcomes were considered for both mortality and hospitalisation. For long-term outcomes, the survey showed a high level of agreement on natural causes, cardiovascular, cerebrovascular, and respiratory diseases, and lung cancer (Figure 4). The same result was obtained for short-term outcomes, except for lung cancer (Figure 5).

Moreover, asthma and neonatal outcomes were considered as potential additional outcomes to include in the Atlas based on their causal association with air pollution and public health relevance. However, for neonatal outcomes, the systematic reviews did not estimate a quantifiable dose-response function applicable in the Atlas maps, and therefore, they were excluded from the Atlas. For asthma, the working group analysed the case definitions used in the primary evidence included in the systematic reviews to estimate the dose response and found that all the studies used ad hoc measurements or included primary care visits among the endpoints, while no study showed a clear association with hospitalization only or mortality. Thus, the asthma-related outcomes, for which a dose response function was available, were not included in the Atlas, because they did not satisfy the criterion of accurate information in the available data sources. Moreover, the plenary session in Rome emphasized a general interest in prioritizing mortality over hospitalization outcomes.

The group discussed also the issue of how to obtain reliable estimates of the attributable fraction of short-term outcomes using the average annual mean concentrations of pollutants. Nevertheless, no final decision was reached about this point.

Counterfactual scenarios

Two models for calculating the impact with counterfactual scenarios were proposed to define the fraction of attributable deaths: theoretical counterfactual (i.e., the number of avoidable deaths if pollution levels were within WHO thresholds, applied to the attributable fraction of actual deaths) and hypothetical counterfactual (i.e., the number of deaths that would occur if pollution levels were reduced to European union regulation thresholds by a specified future date). 

The survey identified the combination of the two strategies for counterfactual calculation with a 53% adhesion rate. Following the meeting in Milan, the working group agreed to calculate the fraction of attributable deaths based on the reference levels set at 5 µg/m³ for PM_2.5 and 10 µg/m³ for NO₂ by the European environment agency, following the WHO limits.²⁷ During a plenary discussion in Rome, the possibility of conducting a sensitivity analysis with a lower counterfactual scenario, set at 2.5 µg/m3 for PM_2.5, was considered but ultimately not included as a final approach. Further decisions included providing a set of predefined scenarios with pollutant levels and the corresponding timeframes required to reach these thresholds and allowing users to set their own scenarios.

Maps and infographics

The plenary meeting discussion defined the maps to enclose into the Atlas:

• Population maps including density and fragility indicators (age over 65 years, low education level, i.e., less than secondary school degree or a cohort specific classification,²⁸ prevalence of at least one chronic disease according to the National Health Interview survey);
• Pollution concentration maps (PM₁₀, PM_2.5, NO₂, and O₃);
• Health outcomes maps including mortality (all-cause mortality, cardiovascular and respiratory disease, lung cancer) and hospitalization (all-cause hospitalization, cardiovascular and respiratory disease);
• Population-weighted exposure (PWE);
• Attributable death fraction maps.

Only PM_2.5 and NO₂ will be considered when calculating PWE and attributable deaths. The group discussed the issue of partial overlapping in the estimated effects of the two pollutants, with the consequent risk of double counting attributable deaths. The proposed solution was to include a warning about this problem, but not try the production of a combined estimate. 

Discussion

Scientific evidence and data availability were the primary criteria used to select both diseases and age groups for inclusion. No outcomes were excluded based on lack of relevance to public health. Regarding scientific evidence, expert judgment aligned with the findings of the systematic reviews used by the Atlas to estimate the dose-response function²¹ and with the most recent WHO report on health risk due to air pollution in Europe³. Moreover, the availability of information has led to a reduction in the number of outcomes that can be included. Specifically, outcomes must be quantifiable using data sources that are available nationwide (HDR and death certificates). To ensure consistency, the Atlas will include outcomes for which the original studies used administrative data on admissions or mortality to estimate dose/response concentrations. This criterion led to the exclusion of asthma in children and chronic obstructive pulmonary disease (COPD) in adults. 

Regarding the pollutants, PM_2.5 and NO₂ were selected due to their nationwide availability, their dose-response functions related to most outcomes, and the absence of seasonal variations in the data. The main identified issue was the partial overlap in the outcomes attributed to both pollutants.  Therefore, the estimates obtained separately cannot be summed up without considering the influence of one on the other. The group considered the consequent risk of double-counting the attributable deaths as a critical communication issue. Nevertheless, the group did not find any suitable solution for producing estimates of the attributable deaths with bi-pollutant models nor did decide to produce synthetic indicators of PM_2.5 and NO₂ exposure. This topic will require further research and elaboration.

The Google Forms questionnaire facilitated responses from professionals across the entire national territory, favouring good participation rates. Between the first and the second round, there were not many differences in the choices of outcomes and exposures to be included. This stability was likely due to a high level of agreement between the experts and decision-makers included in the panel. It is important to underline that the decision-makers who participated were also well-informed on technical issues. Moreover, the consensus achieved during the post-survey discussion may have played a significant role.

Finally, the Delphi method was not overly time-consuming for individual experts, who reported no difficulties in completing the survey despite the detailed questions regarding content and methods. The participation in the Delphi was relatively high, compared with similar experiences and provided a wide diversity of angles).²⁹ On the other hand, the process impacted on the overall Atlas production time making it longer. However, this may not be seen only as a disadvantage, as it facilitated communication among different stakeholders, leading to more informed and shared decision-making.

Co-creation in the development of public services is an emerging area of interest, and the European Commission is investing in it,^30,31 although it remains underrepresented in literature. The construction of the Atlas includes a co-creation process involving various actions and methods. As shown in Figure 2, a multiple strategy has attempted to orient the choices that must be based on both scientific knowledge and understanding of lay users’ information needs. The Delphi method, complemented by frequent remote and in-person plenary meetings, was specifically chosen to involve experts, while lay users, unfamiliar with technical issues, were involved to define the information needs and the best communication modalities through focus groups and interviews. The involvement of expert stakeholders and the high participation rate in the Delphi survey instill confidence in the appropriateness and soundness of this decision.

Furthermore, the in-person meeting favoured the interaction with stakeholders and experts not directly involved in the project; in particular, the one held in Riccione in 2024 was a public event. This interaction enriched the co-creation process. The public presentations sparked active and constructive debates between the audience and the speakers. Additionally, the alignment between the results from the Delphi process and the new systematic review in plenary sessions facilitated an open debate about the inputs received from the external advisors. Nevertheless, the process did not arrive to identify viable solutions for all the scientific questions. Indeed, the group decided to prioritize both short- and long-term effects of air pollution on health, but no practical solution was identified to estimate the short-term effects starting from the chosen exposure data. The explicit endorsement of the method by the sponsor further affirmed its validity as a process. Moving forward, the Atlas project will engage more users, both citizens and members of the control rooms, in an ongoing dialogue with the experts and the web designers (Figure 2).

This co-creation experience underscores the challenges of engaging both experts and lay users in defining public health services. While the effectiveness of co-creation in producing impactful interventions and tools is acknowledged by many authors,^30,32 these cannot be separated from scientific evidence. Therefore, determining which topics should be discussed exclusively by experts and which should involve lay users represents the main challenge in implementing co-creation in this field.³³

Although users were not directly involved, the Delphi method could potentially be functional in the co-creation process and effective in reducing the gap between providers and users. Indeed, the heterogeneity of the decision-makers and their varying experiences with the public were reflected in their responses and the resulting decisions. 

Conclusions

The co-creation process adopted in this study facilitated broad participation and agreement on methods and contents among different units, experts, stakeholders, and external advisors, enabling informed technical decisions regarding the topics, sources, and methods to be included in the Atlas. At the same time, it has maintained consistency with the best available scientific evidence.

Conflicts of interest: none declared.

Funding: Project implemented with the technical and financial support of the Ministry of health – PNC PREV-A-2022-12376. The Italian government has approved the National Plan of Complementary Investments – PNC (Legislative Decrees 59/2021 and 101/2021), which complements the interventions of the National Recovery and Resilience Plan (PNRR) with national resources. This study is part of the specific investment (E.1) “Health, Environment, Biodiversity and Climate”, whose goal is to establish a new model of territorial health care through the definition of a new systemic institutional arrangement for prevention in health, environment, and climate, in line with an integrated approach (one health) and a holistic vision (planetary health). The investment E.1, our focus area (Area A), covers priority programmes for the health system and it is structured into eight projects. Our project, titled “Outdoor air and health: an integrated Atlas to support decisions and research”, applied to project No. 2, “Prevention and reduction of health risks related to outdoor air pollution” received funding by the Ministry of health-PNC PREV-A-2022-12376981 (Air and Health Atlas Study Group, investment E.1 of the PNC in collaboration with Ministry of health).

References

1. World Health Organization. Ambient (outdoor) air pollution. Key facts. 24.10.2024. Available from: https://www.who.int/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health (last accessed: 17.07.2025).
2. Boogaard H, Patton AP, Atkinson RW, et al. Long-term exposure to traffic-related air pollution and selected health outcomes: A systematic review and meta-analysis. Environ Int 2022;164:107262. doi: 10.1016/j.envint.2022.107262
3. World Health Organization, European Region. 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. Copenhagen: WHO Regional Office for Europe; 2025. Available from: https://iris.who.int/server/api/core/bitstreams/19bdf746-06de-4f1b-90ba-f8d53989936f/content (last accessed: 12.01.2026).
4. Eckhardt J, Kaletka C, Krüger D, Maldonado-Mariscal K, Schulz AC. Ecosystems of Co-Creation. Front Sociol 2021;6:642289. doi: 10.3389/fsoc.2021.642289
5. Agnello DM, Loisel QEA, An Q, et al. Establishing a Health CASCADE-Curated Open-Access Database to Consolidate Knowledge About Co-Creation: Novel Artificial Intelligence-Assisted Methodology Based on Systematic Reviews. J Med Internet Res 2023;25:e45059. doi: 10.2196/45059
6. McPherson S, Reese C, Wendler MC. Methodology Update: Delphi Studies. Nurs Res 2018;67(5):404-10. doi: 10.1097/NNR.0000000000000297
7. Skulmoski GJ, Hartman FT, Krahn J. The Delphi Method for Graduate Research. J Inf Technol Educ 2007;6(1):1-21. doi: 10.28945/199
8. Beiderbeck D, Frevel N, von der Gracht HA, Schmidt SL, Schweitzer VM. Preparing, conducting, and analyzing Delphi surveys: Cross-disciplinary practices, new directions, and advancements. MethodsX 2021;8:101401. doi:10.1016/j.mex.2021.101401
9. Alicandro G, Grande E, Sebastiani G, Frova L. The 2011 Italian census cohort for the study of socioeconomic inequality in mortality. Int J Public Health 2020;65(5):693-97. doi: 10.1007/s00038-020-01375-6
10. Petrelli A, Ventura M, Di Napoli A, Pappagallo M, Simeoni S, Frova L. Socioeconomic inequalities in avoidable mortality in Italy: results from a nationwide longitudinal cohort. BMC Public Health 2024;24(1):757. doi: 10.1186/s12889-024-18205-6
11. Marinacci C, Grippo F, Pappagallo M, et al. Social inequalities in total and cause-specific mortality of a sample of the Italian population, from 1999 to 2007. Eur J Public Health 2013;23(4):582-87. doi: 10.1093/eurpub/cks184
12. Gandini M, Scarinzi C, Bande S, et al. LIFE Med Hiss: An innovative cohort design for public health. MethodsX 2018;6:82-91. doi: 10.1016/j.mex.2018.12.007
13. Gandini M, Scarinzi C, Bande S, et al. Long term effect of air pollution on incident hospital admissions: Results from the Italian Longitudinal Study within LIFE MED HISS project. Environ Int 2018;121(Pt 2):1087-97. doi: 10.1016/j.envint.2018.10.020
14. Eeftens M, Beelen R, de Hoogh K, et al. Development of Land Use Regression models for PM(2.5), PM(2.5) absorbance, PM(10) and PM(coarse) in 20 European study areas; results of the ESCAPE project. Environ Sci Technol 2012;46(20):11195-205. doi: 10.1021/es301948k
15. Wang M, Beelen R, Basagana X, et al. Evaluation of land use regression models for NO2 and particulate matter in 20 European study areas: the ESCAPE project. Environ Sci Technol 2013;47(9):4357-64. doi:10.1021/es305129t
16. Stafoggia M, Oftedal B, Chen J, et al. Long-term exposure to low ambient air pollution concentrations and mortality among 28 million people: results from seven large European within the ELAPSE project. Lancet Planet Health 2022;6(1):e9-18. doi: 10.1016/S2542-5196(21)00277-1
17. Gariazzo C, Carlino G, Silibello C, et al. Impact of different exposure models and spatial resolution on the long-term effects of air pollution. Environ Res 2021;192:110351. doi: 10.1016/j.envres.2020.110351
18. Maio S, Gariazzo C, Stafoggia M, et al. BIGEPI project: environmental and health data. Epidemiol Prev 2023;47(6) Suppl 3:8-18. Italian. doi: 10.19191/EP23.6.S3.003
19. Beelen R, Raaschou-Nielsen O, Stafoggia M, et al. Effects of long-term exposure to air pollution on natural-cause mortality: an analysis of 22 European cohorts within the multicentre ESCAPE project. Lancet 2014;383(9919):785-95. doi: 10.1016/S0140-6736(13)62158-3
20. Dimakopoulou K, Samoli E, Beelen R, et al. Air pollution and nonmalignant respiratory mortality in 16 cohorts within the ESCAPE project. Am J Respir Crit Care Med 2014;189(6):684-96. doi: 10.1164/rccm.201310-1777OC
21. Federica Nobile F, Caranci N, Strippoli E, et al. Association between long-term exposure to air pollution and cause-specific mortality within five Italian longitudinal metropolitan studies. Atmosferic Enviroment  2024;339:120873. doi: 10.1016/j.atmosenv.2024.120873
22. Boogaard H, Samoli E, Patton AP, et al. Long-term exposure to traffic-related air pollution and non-accidental mortality: A systematic review and meta-analysis. Environ Int 2023;176:107916. doi: 10.1016/j.envint.2023.107916
23. Boogaard H, Atkinson RW, Brook JR, et al. Evidence Synthesis of Observational Studies in Environmental Health: Lessons Learned from a Systematic Review on Traffic-Related Air Pollution. Environ Health Perspect 2023;131(11):115002. doi: 10.1289/EHP11532
24. Besag J, York J, Mollié A. Bayesian image restoration, with two applications in spatial statistics. Ann Inst Statist Math 1991;43(1):1-20. doi: 10.1007/BF00116466
25. Stafoggia M, Bellander T, Bucci S, et al. Estimation of daily PM10 and PM2.5 concentrations in Italy, 2013-2015, using a spatiotemporal land-use random-forest model. Environ Int 2019;124:170-79. doi: 10.1016/j.envint.2019.01.016
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. World Health Organization. WHO global air quality guidelines. Particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. Geneva: World Health Organization; 2021. Available from: https://iris.who.int/bitstream/handle/10665/345329/9789240034228-eng.pdf (last accessed: 17.07.2025).
28. Petrelli A, Di Napoli A, Sebastiani G, et al. Italian Atlas of mortality inequalities by education level. Epidemiol Prev 2019;43(1) Suppl 1:1-120. doi: 10.19191/EP19.1.S1.002
29. Schifano J, Niederberger M. How Delphi studies in the health sciences find consensus: a scoping review. Syst Rev 2025;14(1):14. doi: 10.1186/s13643-024-02738-3
30. Matti C, Rissola G, Martinez P, et al. Co-creation for policy: Participatory methodologies to structure multi-stakeholder policymaking processes. In: Matti C, Rissola G (eds). Cordis Europa. Project ID 770356. 2023. Luxembourg: Publications Office of the European Union, 2022. Available from: https://bit.ly/3nS3FVp (last accessed: 05.04.2026).
31. Turku University of Applied Sciences. CoSIE - Co-creation of Service Innovations in Europe. 2023. Available from: https://cosie.turkuamk.fi/arkisto/index.html (last accessed: 05.04.2026).
32. Batalden M, Batalden P, Margolis P, et al. Coproduction of healthcare service. BMJ Qual Saf 2016;25(7):509-17. doi: 10.1136/bmjqs-2015-004315
33. Giorgi Rossi P, Ferrari F, Amarri S, et al. Describing the Process and Tools Adopted to Cocreate a Smartphone App for Obesity Prevention in Childhood: Mixed Method Study. JMIR Mhealth Uhealth 2020;8(6):e16165. doi: 10.2196/16165