In real world, what we mean by exposure assessment can be quite complex. In exposure  assessment for local issues, you will need to actually conduct local measurements in the media where you think an individual may be exposed. Such measurements will need to take into account the following:

Step III: Dose Response Assessment

This will be different for diabetes and bladder cancer. For diabetes, we will try to find out what would be the maximum acceptable dose that we can allow. This is the reason, for non-cancer diseases, we refer to environmental health risk assessment as health safety assessment. Here, the ultimate purpose is to test as to how much of the exposure can be safe for humans even though we know that whatever may be the exposure, there will be some harm. Is it possible to identify a certain level of exposure at which the health effect will not be produced? Whatever that value may be, we will set that value as a maximum allowable amount of exposure. Therefore, when we use dose response studies for non-cancer health outcomes, we would aim for dose response studies at different levels to test whether attributable risk would have exceeded the expected or frequency or incidence of the disease in the non-exposed population. 
What does the dose-response relationship between exposure to inorganic arsenic and diabetes (Type II diabetes) look like? 
Wang et.al. (2013) conducted a meta-analysis on the association between exposure to inorganic arsenic through drinking water and the risk of development of diabetes (Type II DM); while they did not provide a table of data, as part of the meta analysis, they provided a dose-response chart as follows (Figure \ref{688210}):