Different pathways toward net-zero emissions imply diverging health impacts: a health impact assessment study for France

In the transport sector, efforts to achieve carbon neutrality may generate public health co-benefits by promoting physical activity. This study aims to quantify the health impacts related to active transport based on four different scenarios leading France toward carbon neutrality in 2050. The French Agency for Ecological Transition developed four consistent and contrasting scenarios (S1 to S4) achieving carbon neutrality by 2050 as well as a business-as-usual (BAU) scenario that extends our current lifestyles until 2050, without reaching net-zero. For each of these Transitions2050 scenarios, we distributed the mobility demand for walking, cycling and e-cycling across age groups. Relying on the health impact assessment framework, we quantified the impacts of the corresponding physical activity on all-cause mortality. The impact of each of the carbon neutrality scenarios was determined by comparison with estimates from the BAU scenario. In S1 and S2 scenarios, volumes of active transport are projected to increase to fulfill the World Health Organizations (WHO) recommendations by 2050, while they increase slightly in S3 and decrease in S4. S2 scenario reaches the highest levels of health co-benefits, with 494 000 deaths prevented between 2021 and 2050. This would translate into a life expectancy gain of 3.0 months for the general population in 2050, mainly driven by e-bikes. S1 would provide smaller but important health benefits, while these benefits would be modest for S3. On the contrary, S4 implies 52 000 additional deaths as compared to the BAU scenario and a loss of 0.2 month in life expectancy. Different ways to decarbonize mobility in a net-zero perspective may achieve very contrasting public health co-benefits. This study illustrates how the public health dimension may provide a relevant insight in choices of collective transformation toward net-zero societies.


Introduction
The Paris agreement, adopted in 2015, establish the objective of constraining the rise in average global temperature to 2 • C above preindustrial levels (UNFCCC (United nations Framework Convention on Climate Change), 21st session 2015).To achieve this, a substantial reduction of 43% in greenhouse gas (GHG) emissions by 2030 is necessary (IPCC (Intergovernmental Panel on Climate Change) 2023).The sixth report from the Intergovernmental Panel on Climate Change (IPCC) highlights the importance of numerous actions that effectively reduce our carbon footprint while concurrently promoting human health.These health co-benefits manifest across several dimensions such as diet, transport, energy supply or green and blue infrastructures (IPCC (Intergovernmental Panel on Climate Change) 2023).
The transport sector accounted for approximately 15% of anthropogenic GHG globally in 2019 (IPCC (Intergovernmental Panel on Climate Change) 2023).In France, it contributes up to 31% of GHG emissions (HCC (The High Council on Climate) 2021).Amongst the many existing levers to limit these emissions (electrification of the energy supply, increased vehicle occupancy, lower transport demand…), some have the potential to yield health co-benefits.Among those, shifting to active modes of transportation offers the additional advantage of mainly relying on local decisions such as reallocation of existing infrastructures and land use planning (Stappers et al 2023).
Active transport encompasses all uses of cycling and walking as transportation means, for occupational and non-occupational purposes and has indeed been identified as a lever for achieving socio-economic, air quality and public health co-benefits (IPCC (Intergovernmental Panel on Climate Change) 2023).However, physical activity levels worldwide have failed to meet international recommendations, affecting both adolescents and adults (Guthold et al 2018).In high-income countries, 9.3% of overall mortality can be attributed to physical inactivity (Katzmarzyk et al 2021).Insufficient physical activity also incurs substantial monetary costs for public and private healthcare systems and households while leading to decreased productivity (Ding et al 2016).A 2017 study focusing on urban populations estimated that meeting international guidelines on physical activity, air pollution, noise, heat and access to green spaces could prevent 20% of deaths.Among these factors, physical activity yielded the largest proportion of preventable deaths (Mueller et al 2017).
The health benefits associated with active transport may be substantiated by various health impact assessment methods that consider diverse parameters, regardless of the geographical context (Mueller et al 2015).Institutions and organizations have devised prospective pathways outlining the necessary transformations required for the societies to attain the objectives of the Paris agreement.Numerous studies quantify the health impacts of active commuting in prospective scenarios, with a majority conducted in Europe, North America or Oceania, primarily focusing on urban areas (Mueller et al 2015, Rojas-Rueda et al 2016).Nevertheless, to the best of our knowledge, only one study has been conducted at the national level in France, quantifying the impacts of a single scenario largely relying on non-technological levers compared to the baseline levels of 2021 (Barban et al 2022).
In this study, we quantified the health impacts related to physical activity through active transport based on four consistent and contrasting scenarios leading France toward carbon neutrality in 2050.

Methods
We conducted a quantitative health impact assessment study, following the most recent guidelines for modeling and reporting health effects of climate change mitigation actions based on scenarios that have been developed by the French agency for ecological transition (Hess et al 2020).We specifically attributed passenger-kilometers projections across ages, types of active transport and net-zero emission scenarios.The assessment presents health impacts and intangible costs related with physical activity generated by active transport.

The ADEME transition scenarios
In 2021, ADEME (The French Agency for Ecological Transition) released the outcomes of a prospective project Transitions2050, describing four scenarios (S1 to S4).Each of these scenarios is designed to achieve carbon neutrality by 2050 in metropolitan France following distinct trajectories corresponding to different societal choices (figure S1) (The French Agency for Ecological Transition, n.d.a).Additionally, a business-as-usual scenario (BAU) was formulated, representing a projection of our current lifestyle trends until 2050, without achieving carbon neutrality.It partly corresponds to an extension of certain past trends and partly to evolutions already underway.
Each Transitions2050 scenario describes the evolution of the main energy consumption drivers over time, especially the demand for transportation for various modes of transportation (table 1 & table S1).Within these estimates, three active modes are considered: cycling, E-cycling and walking, without accounting for recreational activity.Note that the estimates for walking encompass distances covered while transitioning to or from another mode of transport, especially public transportation (e.g train or bus).For each active travel mode, the scenarios express the yearly national demand in terms of Giga passenger-kilometers (Gpkm) at five-year intervals over the period 2015-2050.For this study, passenger-kilometers values were linearly interpolated for each year within the intervals.

Age-specific allocation of active transport demands
For each scenario and mode of transportation, the passenger-kilometers were distributed among age groups according to the mode-specific age distributions observed in the 2019 French mobility survey (Ministry of Ecological Transition, n.d.).We performed our main analysis under the assumption that the age-specific distribution of transport demand do not evolve over time.In addition, we allocated the total number of kilometers cycled over time between E-bikes and mechanical bicycles differently based on age groups and scenarios.This allocation was done in order to achieve two objectives: i) replicate the observed 6.7-year age difference between E-bikers and cyclists as reported in a recent multi-country study (Castro et al 2019) and ii) capture a linear evolution regarding the proportional contribution of E-bikes to the overall cycling distance over time, as assumed in each scenario for 2050.
Further details regarding these calculations are provided in the Supplementary Materials.

Demographic projections
Population projections, including mortality and life expectancy, for different years and age groups, were accessed from publicly available data provided by the French national institute of statistics and economic studies (INSEE, n.d.).These same data were also used by ADEME in the design of all their transition scenarios.

Health impact assessment
We relied on international guidelines to model and report the health effects associated with active transport of the Transitions2050 scenarios for metropolitan France over the 2021-2050 period (Hess et al 2020).Our health modeling method relies on principles similar to the HEAT tool developed by WHO (Götschi et al 2020), but differs in two major aspects.First, we only explicitly considered physical activity for the pathway linking active transport and health while HEAT also assesses the impacts related to air pollution exposure and road injuries.Second, we additionally accounted for the age-distribution of current levels of active transport in France and for specifications of e-bike uses regarding energy expenditure and age distribution.
For each age group, yearly distances travelled were converted into exposure times for walking, cycling or E-cycling using the following average speeds, respectively: 4.8, 14.9 and 18.1 km•h −1 (Götschi et al 2020, Egiguren et al 2021).First, we compared the resulting total duration of active travel to international recommendations for weekly moderate physical activity, averaged over the entire adult population.
The health impacts (number of deaths averted) generated by physical activity associated with active transport were calculated, using a linear dose-response function between duration spent in active transport and the reduction in all-cause mortality, weighted by the specific mortality rate and size of the population: With: We used a relative risk (RR REF ) of 0.89 for all-cause mortality when engaging weekly in 168 min of walking, with maximum risk reduction of 30% and a relative risk of 0.90 for 100 min of cycling, with a maximum risk reduction of 45%.Health impacts are therefore capped above 460 min of walking and 447 min of cycling (Kelly et al 2014).The reference duration corresponds to the duration of active travel that yields the risk reduction stated by the RR REF .We constrained the benefits associated with physical activity generated by active transport to the population aged between 20 and 89 years.This restriction was made based on the fact that the relative risk estimates for cycling and walking were extrapolated from studies encompassing individuals within the age range of 20-93 years (Kelly et al 2014) and the data used for age distribution of cycling distances were collected from individuals aged between 15 and 89 years (Ministry of Ecological Transition, n.d.).The assumed ratio of energy expenditure between moderate assistance E-bikes and bikes was 0.89, as estimated by a recent meta-analysis (McVicar et al 2022).The impacts were evaluated for each age group, taking into account the respective population size and mortality rate.The total count of deaths prevented by each scenario was obtained, together with the number of years of life lost (YLL) prevented, by subtracting from the estimated number of deaths prevented in this scenario the estimated number of deaths prevented in the BAU scenario.YLL is a metric of premature mortality, summing up deaths occurring at each age and multiplying this with the number of remaining years to live up to a selected age limit (here, the projected life expectancy).Finally, based on the estimate of the number of deaths prevented, life tables were generated in order to compute the life expectancy in each scenario.

Monetizing the health impacts
The value of a statistical life year (VSLY) is a macro-economic concept that allows to assess prevention policies based on the monetarized value that civil society confers to a certain risk reduction (Kniesner and Viscusi 2019).We further estimated the intangible costs associated with the health impact of each scenario, using the VSLY that is recommended in France for the socioeconomic assessment of public investments (Emile 2013).In France, the VSLY is projected to be €151 000 in 2025 and this figure is anticipated to increase to €202 000 by 2050.
The VSLY was multiplied by the number of YLL prevented to calculate the monetary impacts of mortality risk reduction.All monetarized values were expressed in € 2022 .
All the estimates were presented alongside uncertainty intervals (UIs), which were determined using the lower and upper bounds of the 95% confidence interval (CI) of the RR used.

Sensitivity analyzes
We performed five distinct sensitivity analyzes.First, we explored a more favorable RR associated with cycling (0.81 instead of 0.90 per 100 min of cycling), as estimated by a recent meta-analysis (Zhao et al 2021).Second, we explored alternative walking and cycling speeds, respectively 3.6 and 13 km•h −1 (The French Agency for Ecological Transition, n.d.b).Third, we investigated an alternative ratio of 0.78 between E-bike and cycling energy expenditures (Berntsen et al 2017).Fourth, for scenarios S1 and S2, we assumed that the high increase of cycling overall would affect the demography of cyclist toward a flatter age distribution, as those observed in countries with a high cycling culture, such as Denmark.We therefore evolved gradually between 2021 to 2035 the distribution of age-specific contribution of kilometers cycled from the levels observed in France (Ministry of Ecological Transition, n.d.) to those reported in Denmark (Center for Transport Analytics, Transport, n.d.) (see supplementary materials).Lastly, we constrained the benefits of active transport to the age range of 20-75 years, as opposed to the wider scope of 20-89 years in the main analysis, in order to make our results comparable to an equivalent use of the HEAT tool (Götschi et al 2020).

Projection of active transport
Time changes in transport demand from 2021 to 2050 are presented in supplementary materials (figure S2) for each mode of active transport and scenario.
Figure 1 compares the estimated average weekly durations of physical activity generated by active transport obtained under each scenarios by 2035 and 2050 with both the 2015 French situation and the minimal recommendations set by the WHO, that is, 150 min of moderate-intensity aerobic physical activity per week.In 2015, the active transport levels in the French population allowed for a weekly average of 80 min of moderate-intensity aerobic physical activity (The French Agency for Ecological Transition, n.d.b).In the S2 scenario, in 2035, the average levels of physical activity through active transport alone are expected to attain the minimal WHO recommendations.The S1 scenario nearly achieves these levels by 2035 and surpasses them by 2050.In contrast, the S3 scenario falls short of reaching the minimal recommended levels of physical activity, both in the medium term (2035) and the long term (2050).Both the BAU scenario and the S4 scenario project minimal increases in active transport compared to 2015 and fail to reach WHO's minimal recommendations, even by 2050.These results are detailed by age groups in supplementary materials (figure S3).

Quantitative health impact assessment
The time-varying health effects expected in each scenario are presented in figure 2 as numbers of prevented deaths and YLL compared to the BAU.The disparities in health impacts among scenarios become more pronounced as time progresses.From 2030 to 2040, the benefits of S1 and S2 scenarios experience a large increase, while the benefits of S3 scenario show a more linear progression.On the other hand, S4 scenario has negative impacts that intensify over time and remain noticeable for all age groups above 50 years old.Deaths prevented by scenarios S1-S3 are highest in the 80-89 age group and YLL prevented increase for younger ages, plateau for the 40-59 and the 60-79 years old age groups and are limited amongst 80-89 years old.
For all age groups, the benefits of S2 surpass all other scenarios while the benefits of S1 outweigh those of S3.The YLL metric makes these benefits apparent for younger age groups (from 20 years old on).The health impacts of S4 remain small, though negative.
Figure 3 distinguishes the health benefits of the different scenarios by type of active transport.Starting from 2030 onwards, scenarios S1 and S2 demonstrate a rapid increase in the health benefits associated with E-cycling, while its progression in S3 is more linear.The health benefits in the first scenario mainly come from cycling, with more or less equal contributions of classical and E-cycling throughout the period from 2021 to 2050.In scenarios S2 and S3, E-cycling emerges as the mode of active transport contributing to the majority of health benefits, while walking and cycling provide similar benefits each year.On the other hand, scenario S4, which is comparable to the BAU scenario in terms of cycling mobility, exhibits lower levels of walking.Consequently, the health detriments observed in scenario S4 can be entirely attributed to the decrease in pedestrian mobility.
Figure 4 presents the annual health and economic benefits of each scenario at medium (2035) and long (2050) term.Here again, scenario S2 shows a larger impact over all other scenarios, S1 exhibits a superiority over scenario S3 and scenario S4 has detrimental health impacts as compared to the BAU.The number of deaths prevented in 2050 by active transport reaches 25 402 [13 687-34 661] with the trajectories projected for scenario S2, 18 345 [10 083-24 833] with scenario S1 and 6,250 [3,469] with scenario S3.Scenario S4 would induce an additional 2,658 106] deaths.

Sensitivity analyzes
Table 2 summarizes the results of our sensitivity analyzes.Overall, the parameter choice which affects the most our estimates is the RR value for cycling.Assuming a more favorable RR leads to notable increases in the YLL estimated, around +50-+70%.Modifying the speeds of walking and cycling has less but substantial effects of about +20% in the YLL prevented.By lowering the speeds, the duration of exposure to active transport is extended, leading to higher health impacts.The third sensitivity analysis, focusing on the ratio of energy expenditure provided by an E-bike compared to a mechanical one slightly reduces the health impact, by less than 10% of YLL prevented.In scenarios characterized by high levels of cycling (S1 and S2), using the gradual shape of the age distribution of cycling kilometers of France in 2021 to Denmark in 2035 (and on to 2050) leads to a minor decrease in the number of prevented YLL.Considering Denmark's relative age contribution for cycling kilometers leads to less deaths prevented in older age groups (Supplementary materials, figure S6) and so has limited impacts on YLL prevented.Restricting the health effects to individuals below the age of 75 reduces the YLL prevented ascribed to active transport by about 20%.
For each analysis and any choices of the parameters, the ranking of scenarios according to their health impacts remained unchanged, with S2 as the most beneficial scenario, followed by S1 and then S3, while S4 yielded negative health impacts.

Main results
In this study, we highlight the widely varying potential health benefits of transport transition pathways towards carbon neutrality, through their impact on active transport.Future pathways promoting modal shift to active modes of transportation facilitate the attainment of the minimal recommendation on physical activity set by the WHO solely through transport activities.Conversely, policies wagering on future technological developments to attain carbon neutrality are likely to intensify the lack of physical activity and its harmful health impacts in the general population.According to the extent of modal shift and compared to the BAU, health impacts at the 2035 horizon may range from 19 000 deaths prevented to 2000 additional deaths, representing respectively €34 billion annual savings and €3 billion additional mortality costs.Qualitative results and the ranking of Transitions2050 scenarios remained consistent through several sensitivity analyzes.

Literature comparison
For all scenarios, the majority of health benefits may be explained by an achievable increase in cycling mobility, particularly through the use of E-bikes.The most optimistic scenario reaches 20 kilometers cycled and 7 kilometers walked per person weekly, representing roughly 20 min per day of active transport.Such increase in active transport seem considerable as compared to the levels reported before 2020.However, they remain lower to those reported in other European countries: In Netherlands, active transport represent between 24 and 28 min daily (Fishman et al 2015) and more than 50% of trips <2 kilometers and 30% of trips between 2 to 5 kilometers are undertaken using bicycles (Goel et al 2021).This highlights the margin of progression for active transport in France, where only 5% of the active population uses a bike for trips covering less than 5 kilometers, while 60% opt for a car (Chantal Brutel, Jeanne Pages, 2021).The anticipated levels of active transport in S1 and S2 scenarios therefore appear to be quite achievable.
The method we used for quantifying health and economic impacts follows the recommended steps and parameters outlined by the WHO and environmental economists.These approaches are widely described and employed in various scientific studies (Emile 2013, Götschi et al 2020, Smith et al 2021).A health impact assessment study on the transport sector was previously conducted in France using a similar method (Barban et al 2022).The negaWatt scenario explored in this study projected active transport levels similar to those observed in the S1 scenario, and the estimated health impacts were similar to those we obtain.Studies comparing multiple scenarios consistently demonstrate significant differences in the health impacts of active transport: Climate policy strategies that prioritize health-promoting pathways and are built upon the behaviors of the population yield more significant health benefits (Hamilton et al 2021, Milner et al 2023).
Preventing, as we predict in scenario S2, 19 000 deaths annually by 2035 solely through deliberate decisions in the development of the transportation sector holds significant importance for public health prevention as it corresponds to 2.8% of all-cause mortality projected for this year (INSEE, n.d.).Indeed, in comparison, in France, it is estimated that exposure to PM2.5 is responsible for 40 000 deaths annually (Santé publique France 2021) while heat-attributable deaths reached up to 7,000 in 2022 (Santé publique France 2023).In the transport sector, all road-related fatalities in France range between 3000 and 4000 deaths annually (excepted in 2020 and 2021) (ONISR (Observatoire national interministériel de la sécurité routière) 2023).From a benefit-cost perspective, based on a UK case-study, the economic gain of prevented mortality generated by an additional 267 regular cyclists per kilometers of a new separated cycleway breaks even on the costs of construction (Candio and Frew 2023).
The adoption of electric modes of transportation is widely recognized as a strategy for mitigating emissions within the transportation sector.However, it is important to acknowledge that electric vehicles, along with the high battery metals demand (European Federation for Transport and Environment 2023), can also contribute to pollutant emissions through the dispersion of road dust particles and the degradation of tires and brakes wear (Bourliva et al 2017).In the transport sector, modal shift and decarbonation of the energy supply can yield larger health co-benefits (Milner et al 2023).The promotion of E-bikes is a cost-effective approach to reduce transport associated carbon emissions while improving health and mobility (Jenkins et al 2022).The allocation of more areas to bicycle infrastructure can also optimize the utilization of public space (Gössling et al 2016).
Other outcomes have to be taken into account when outlining health effects of modal shift such as air pollution, road injuries or noise.As the health risk estimate relies on a prospective cohort, these factors are implicitly incorporated into the relative risks.Regardless of the specific assumptions, the health benefits of active transport consistently outweigh the potential adverse effects stemming from exposure to air pollution and traffic-related incidents (Mueller et al 2018).Among various modes of transportation, car users exhibit highest levels of exposure to air pollutants such as PM2.5, black carbon and CO (de Nazelle et al 2017).The association with road injuries depends greatly on the assumptions made on the fatality and injury rates of cyclists (Quam et al 2017) as well as the quality of transport infrastructures (WHO 2022).Furthermore, each incremental shift from cars to bicycles corresponds to a noteworthy decrease in pollutant emissions (Bernard et al 2021).For instance, it has been estimated that an individual who replaces one car trip per day with a bicycle trip over 200 d can reduce their CO 2 emissions by 0.5 tons per year (Brand et al 2021).Noise pollution would also benefit from modal shift, which would yield positive impacts on health-related quality of life, mediated by annoyance and sleep disturbance (Héritier et al 2014).It is also expected that a general increase of physical activity reduces morbidity related to cardiovascular risks, cancer, osteoporosis, depression and low back pain (Miller et al 2016).Further impacts related to quality of life can also benefit from active transport such as work performance and productivity (Ding et al 2016).

Limitation
Using the transport demand as a proxy for physical activity allows for a quantification of health impacts associated with specific active modes of transport but does not encompass several variations in the population.For example, given the expectation that scenarios S3 and S4 would limit active transport primarily to those who are already physically active, while scenarios S1 and S2 would promote active transport for a broader population, including individuals who are sedentary, a more targeted distribution of transport demand could be achieved.This distribution would involve a heterogeneous allocation of distances within age groups.By implementing such an allocation strategy, the differences in health impacts between high cycling (S1, S2) and low cycling (S3, S4) scenarios could be amplified.Another limitation linked to this proxy is that the physical activities performed out of transport context are not taken into account, despite their anticipated distinct evolution across scenarios.However, recent evidence suggests that, among adults, active transport adds to, rather than replaces, other physical activities (Wanjau et al 2023).
Using a relative risk from a meta-analysis provides a robust association between active transport and mortality reduction but does not allow to specify distinct relative benefits based on age, gender, or socio-occupational category.A recent meta-analysis showed no significant differences in the association of cycling with all-cause mortality for gender groups and among individuals aged above or below 50 years (Zhao et al 2021).Additional cohort studies are needed to assess the potential heterogeneity of the association between more specific age groups and socio-occupational categories.
In a context of modal shifts, conventional cycling could be substituted by e-cycling (Sun et al 2020); however, we did not account for that potential rebound effect as the evidence is not consensual in the literature (de Haas et al 2021).

Uncertainty analysis
Our analysis does not explicitly take into account the potential adverse effects of cycling and walking, such as crashes and exposure to air pollution.However, since our assessment relies on a dose-response function linking physical activity generated by active transport and all-cause mortality, these harmful effects are implicitly considered.In addition, it is very likely that our analysis involves an overestimation of the adverse effects of traffic-related pollution and crashes, especially in scenarios identified as the most favorable (S1 and S2).Indeed, in S1 and S2, the kilometers travelled by cars are assumed to decrease dramatically and substantially more than in TEND, S3 and S4 (table S1).In Europe, differences in overall crash rates between cities are driven largely by crashes that involved motor-vehicles (Branion-Calles et al 2020).This reduction of car traffic in S1 and S2 would also lead to a decrease in the exposure of pedestrians and cyclists to traffic-related air pollution.At last, these scenarios involve a substantial increase in bike volumes (figure S2) while evidence suggests that the individual risk of fatal crashes during walking and cycling decreases as the volumes of active transport increase (Jacobsen et al 2015, WHO, (World Health Organisation).Regional Office for Europe 2022).
Our results are calculated from a projected scenario, the BAU.This inherent uncertainty may introduce variability in the absolute results of each scenario, yet it does not alter the relative hierarchy of their respective impacts.
Additionally, the RR is associated with a broad CI that we propagated as an UI in our calculations.This leads to a factor 3 between the lowest and upper bound of the UI in predicted impacts, with overlapping UIs between some of the scenarios assessed (figure 2).
All in all, it is important to acknowledge that a health impact assessment study only provides theoretical estimates of health and economic impacts, and that its quantitative results should not be over-interpreted.However, the magnitude of effects and the comparison of scenarios allow for a better interpretation of the co-benefits and potential implications of specific public policies.
Several levers are recognized to support important behavioral shifts to active transport.Changes in the built environment such as increased accessibility, favorable land use mix and creation of adapted infrastructures are known to induce active transport (Kärmeniemi et al 2018).When infrastructures are available, increased relative oil price (due to taxes or extraction costs) can also support a modal shift from car to active transport (Chevance et al 2023).

Conclusion
Different carbon neutral trajectories imply distinct health benefits, this holds particularly true within the transport sector.The implementation of public policies promoting modal shift effectively improves population health while reducing the use of fossil fuels and pollutant emissions.In France, directing transportation towards active modes would yield significant health co-benefits.Transition trajectories primarily relying on technological interventions offer minimal benefits and may even exacerbate the lack of physical activity in the population.In a context of reaching towards carbon neutrality, these results may help guide decision-makers in the choice of optimal transition strategies.

Figure 1 .Figure 2 .
Figure 1.Duration (averaged for the 20-89 years old) of physical activity generated by active transport in 2035 and 2050 compared to the French levels in 2015 (all estimated by ADEME Transition scenarios) and the WHO guidelines for moderate physical activity.*WHO recommendations apply for total physical activity, while the explored scenarios only quantified physical activity generated by active transport

Figure 3 .
Figure 3. Years of life lost (YLL) prevented as a function of time and type of active transport, under (a) scenario S1, (b) scenario S2, (c) scenario S3 and (d) scenario S4.
Years of life lost; CI = Confidence interval; RR = Relative risk; MET = Metabolic equivalent of task.a The distribution of age-specific contribution of cycling kilometers evolve gradually between 2021 and 2035 from the levels observed in

Table 1 .
Societal choices, transport sector and active mobility in ADEME's prospective scenarios.
(Kelly et al 2014)th prevented.RR REF : Relative risk of all-cause mortality associated with the reference duration of active travel(Kelly et al 2014).Specific duration a,s : Projected in the scenarios and distributed across age groups.
(Kelly et al 2014): 168 min for walking and 100 min for cycling as estimated in a meta-analysis of 18 cohorts(Kelly et al 2014).This correspond to an energy expenditure of 11.25 MET.hour.week−1 .MR a : Specific mortality rate of an age category a. POP a : Population of the age category a.

Table 2 .
Health impacts of active transport for ADEME's scenarios.Main and sensitivity analysis.