This is more of a "subjective" and "qualitative" study, since there is often subjective interpretation of the data due to the uncontrolled exposure. However, the data is no less valuable than those from a toxicological study, given that the exposures are more realistic (or representative of real exposures). Perhaps one can use the toxicological study results to determine threshold pollution levels (e.g., "smog alerts" or occupational air quality standards) and the epidemiological study results to show that the chosen thresholds can adequately protect public health in the long run.

Due to the uncontrolled nature of the pollutant exposure, there are a number of uncertainties in the results. Plus, there are competing and synergistic effects of the subjects' environments and behaviors (for example, subjects living in hot or cold climates, smoking, diet, exercise, etc.). Some of these factors are difficult to factor into the results because people have the ability to move from one place to another, so there will be wide variations in responses for different individuals in the study population.

It is unfortunate that there are a number of uncertainties in the interpretation of epidemiological data. It is easy to interpret epidemiological data to suit nearly any special interest. Occasionally, the results of suspect epidemiological studies become the focus of unscrupulous media coverage or special interest group lobbying. (Note that it is just as easy to falsify data in a toxicological/clinical study to produce desired outcomes, but in this case it is the actual data that is fraudulent while in the epidemiological case, it is the interpretation of the data that is the problem.)