01 Set objectives

Agree with key stakeholders what question you are seeking to answer through a health impact assessment and why. For example, you may want to know the number of deaths and illnesses associated with air pollution in a specific location or the government may want to estimate the benefits of implementing a specific policy. At this stage, you may want to consider more sophisticated exposure analysis, using more detailed geographic or demographic data. You should still consider whether to first conduct a screening assessment to provide a basic understanding of who is potentially affected, the impact, if any, on health as well as the potential scale of any impacts before proceeding to a full health impact assessment. It may be useful to refer back to the Emission Inventory or Decision Support guidance to inform objectives or questions for analysis. 

02 Make a plan for health impact assessment, including exposure, socioeconomic and economic analysis if desired

Based on the objectives set in Step 1, your jurisdiction's air quality and health staff should meet jointly to plan for a health impact assessment. Working collaboratively, you should develop a plan to gather appropriate data that is representative for your jurisdiction, to dedicate staff who will learn the appropriate tools and conduct the assessments, and to ensure you have the budgetary resources and buy-in from management to support the assessment process and utilize the results. For a basic health impact assessment, you will follow the steps in Stage 2. The steps that follow provide guidance on conducting more detailed exposure analysis. 

03 Exposure characterization I: Collect AQ monitoring data

With more advanced monitoring data and modeling capacity you may be able to conduct more detailed exposure analysis (e.g., population weighted exposure instead of city-wide average exposure). Where possible, use data from official sources such as a government-run monitoring network. Where there are limits to data available, data produced by other trusted sources such as researchers, global satellite data or modelled estimates. Networks of low-cost sensors may also be useful to provide finer scale data. 

04 Exposure characterization II: Use of air quality monitoring data to validate exposure assessments

You will need information about the amount of air pollution that people are exposed to and, if analyzing the impacts of a policy change, the estimate of the change in air pollution that is expected to result from that policy change. The format of the data required will depend on the tool you are using. It is possible to use data for a range of spatial (i.e., geographic region) and temporal (i.e., time period) scales. However, the spatial and temporal scales should be the same across all data sets (e.g., air pollution, population, and health statistics). 

05 Exposure characterization III: Collection of detailed population data

You will need information about the population that is exposed to air pollution. This could come from a national census, surveys, university studies or other data. The format of the data required will depend on the tool you are using and some tools have population information built in. It is possible to use data for a range of spatial (i.e., geographic region) and temporal (i.e., time period) scales. However, the spatial and temporal scales should be the same across all data sets (e.g., air pollution, population, and health statistics).

At this stage, you should consider using data that is stratified by demographic information (i.e., age, race, sex) which will allow you to quantify impacts for different subgroups.

06  Exposure characterization IV: Collection of detailed health data

At this stage, you may be able to use more detailed information on health impacts, including additional health outcomes (e.g., morbidity in addition to mortality). As above, the data used should match the spatial and temporal scale of the other data used in the analysis. Data could come from official government sources or scientific studies.

07 Run the analysis (including sensitivity analysis)

After formatting the data to input in your preferred tool, you will run your analysis for your main question. You should also run a series of sensitivity analysis to understand how the results change when you adjust key variables or use different C-R functions. If using more detailed health or demographic data you will likely want to run more sensitivity analyses to test your conclusions. It is important to conduct rigorous QA/QC, which could include two analysts running the same analysis to ensure the results are replicable.

08 Implement a communications plan

It is important to carefully consider how you will communicate the results of your analysis and translate technical information into something that can be understood by policymakers and members of the public. First, consider the original question the analysis was conducted to answer and whether the results are useful for that purpose. You will also need to understand any uncertainties in the analysis or critical assumptions made and to be able to explain how changes in those factors could affect the results. If you are using more detailed demographic or health data, there is likely to be public and political interest in any disparities that are revealed by the analysis. The guidance on Public Engagement and Communication also provides helpful advice on how to approach this Step.