Question about the Health Impact Results

Dear BenMAP users,

I have a few question regarding to the health impact results. I knew the BenMAP program is meant to assess health impacts due to the change in air pollution level. If I understand correctly, the “Point Estimates” is the mean estimates of the health effect change due to the change in air pollution level. But is there any way that I can get the not only the changes in health effects, but the total health effects due to raw air pollution data.

In Dr, Punger and Dr. West’s (2013) paper Section 3.3, they used BenMAP’s Krewski (2009) health impact functions and provided the results of estimated total mortality due to exposure to modeled PM2.5 at 12 km instead of the changes in the mortality due to baseline and control case. Is it the output from BenMAP? If yes, which variable/variable provide this information?

Any suggestion is appreciated,


Sorry for multiple posts.

In Dr. Fann (2012)'s paper Table I, it also shows the PM2.5-related health impacts due to 2006 modeled air quality, it mentioned the health impact is relative to nonanthropogenic background, does it mean the base PM2.5 values are 0 over each grid? And then the change in health impact result actually represent the total premature morality (for example) due to the modeled air quality?


Hello Xiangyu,

I have the same concern in this issue. My case does not consider to the change from basecase and scenario case, I would like to assess the health impacts from the current situation of air quality (specifically from PM2.5). So far, I have an idea that for the change of air quality I will use change from the threshold level to the current level, but I’m not sure whether it right or not. Do you have the solution for this problem?

Ha Chi

Hi Ha Chi,

I think using the threshold level is a possible solution. If I understand correctly, Punger and West (2013)'s study used a threshold of 5.8 as the control case, and modeled PM2.5 as base case to assess the health impacts from current situation of air quliaty. I have sent them an email about this question, but haven’t got any reply yet.

In Dr. Fann’s (2012) paper, he used" nonanthropocentric background" modeled by air quality model as control case. I also read a paper regarding to Black Carbon-mortality relationship (Li, et al. 2016). Dr. Li replied me that they used zero as control case to assess health impact due to BC.

Hope this helps. It will be great if other BenMAP users can correct me if I was wrong and provide more suggestions on this topic!


1 Like

Hi all,

Thanks for posting your questions to the BenMAP discussion forum.

The analytic steps for assessing the total health impacts from air pollution in BenMAP are nearly identical to those for assessing the health impacts from incremental changes in air quality. These steps differ for monitor and model data.

If you are using monitor data, you can perform a 99% rollback using the monitor rollback function. BenMAP is currently unable to process a 100% rollback, but 99% is a close approximation. This provides an estimate of the total health burden associated with air pollution. Because some amount of air pollution may be naturally occurring (i.e., nonanthropogenic), I recommend that you specify a background concentration for this natural air pollution fraction, which I discuss below in the discussion of model data. The 99% rollback will then rollback each monitor value to roughly the background concentration, so that your air quality delta represents the total anthropogenic burden of air pollution. The screenshot below displays this portion of the analysis in BenMAP.

If you are using model data, you still start your analysis with one .csv file that contains your baseline air quality grid (e.g., observed concentrations in 2015). As an example, please see the screenshot below:

This file will serve as your baseline air quality data. You can now create a second .csv file to use as your control data file. Change all data in the “Values” column to whatever concentration you would like to use as a background (non-anthropogenic) level. The lowest observed concentration in the Krewski (2009) study is 5.8 µg/m3, so this is one commonly used option, as you noted in the Punger and West (2013) study. Alternatively, you may select a background of 0 µg/m3 if you wish to estimate the impact of any and all ambient particulate matter (anthropogenic or otherwise). See the screenshot below for an example.

Using your baseline (observed or modeled concentrations) and control (background concentrations) air quality surfaces, BenMAP will provide estimate of the total health impacts attributable to your pollutant of interest.

In either the monitor or modeled case, the result is a comparison of current conditions to an ideal scenario that controls virtually all anthropogenically caused air pollution; as a result the air quality delta and the resulting health effects associated with that delta represent an estimate of the total health burden associated with current levels of your selected air pollutant in your study area.


Josh Bankert


Hi Josh,

Thanks very much for your reply. It’s very helpful!


1 Like

Hi Josh! Nice explanation about rollback function. I would like to do one more question about roll back function. How to analyze annual standarts with this tool.
As you can see at figure below (mark with red arrow) the software analises 24 hour mean values or my input data with annual mean values?
Do you recommend any background value?