From the graph above and the table that we constructed based on the chart, we see that the risk estimates are mostly between 1.05 and 1.10 with and for two regions, Waipapa and Wairakei drains, we would have RRs of about 1.7 and 2.8.
If the health outcome is a cancer or a tumour, the situation would become more complex.
For a cancer, say bladder cancer, we cannot talk of a threshold dose as cancer starts at molecular level and no one can say at what molecular level a cancer process does not or does begin. At the end of a dose-response assessment, we do two different things depending on whether we are dealing with non-cancer or cancerous health effects. When we would conduct Environmental Health Risk Assessment for Diabetes, we would identify a level of the dose at which the excess risk of the disease among those who are exposed is nil or minimum. When we would conduct Environmental Health Risk Assessment for cancerous health risks, we would consider the dose-response curve and for the exposure level that is average or the exposure level that we have determined from the exposure assessment stage we would identify what would be the excessive risk of cancer above and beyond the baseline risk.
For cancerous health effects, we can have information on the basis of animal studies or we can have information on the basis of epidemiological studies.
Step IV: Risk Characterisation
In the step of risk characterisation, we start with two pieces of information: dose-response analysis and the exposure assessment. We know for instance, for the different concentrations of the the arsenic in drinking water, there are different relative risks of the health outcomes. These are read off from the results of epidemiological studies conducted to assess the dose-response effects. Thus we have two pieces of information from the environmental health risk assessment process itself: first, we have information on Relative Risk (or Odds Ratio). If we know the baseline risk (that is, disease rates in the non-exposed population), then we can estimate the absolute risk or excess risk or attributable risk due to the exposure as follows:
Absolute Risk = Relative Risk * Incidence Rate in the Non-exposed
We will estimate the absolute risk from the relative risk estimate and the incidence rate among the non-exposed and then we will use this information to characterise the risk.
Thus, for diabetes, we have one set of risk characterisation, and for cancer we have another set of risk characterisation. If we were to use the data above where we knew from earlier studies the extent to which people in the Waikato region were exposed to inorganic arsenic in their drinking water or the sediments (assuming all of the inorganic arsenic were to leach into their drinking water), and if we used the risk estimates we obtained from the meta analysis in Wang's study on the dose-response association between exposure to inorganic arsenic in drinking water and the risk of Type II diabetes, we would need another information in order to assess the extent to which this exposure would translate to a realistic risk for the people living in the Waikato region. This information would be the incidence rate of the development of Type II DM. We have used a prevalence estimate data from a 2000 paper by Moore and Lunt (2000) to assess the excess risk of Type II DM if people in the region were to have similar prevalence of Type II DM they found in their earlier study for all of NZ \cite{Moore2000S65}. We could modify the previous table and would come up with a new set of estimates as follows (Table \ref{577947}):