As victims of the Flint water crisis, we decide to recommend the Cohort study as the most effective study design to get the best information about the potential risk factors that might have caused toxic water. A cohort study is a study that proceeds from exposure to suspected risk (causes) to outcome (effects). In the cohort study, investigators select two quite different groups to compare: a group exposed to particular factors and another group who unexposed to the factors (Reingold, 2016). After following them overtime or looking backward and examining, if the exposed group turns out to have a higher incidence of disease, we are able to determine there should be a strong association between the exposure and the outcome.
The retrospective cohort study design -A study done after the occurrence of both the exposure and outcome – is fittest to investigate rare exposures in a small, well-defined population because “the investigators can make it a point to identify an adequate number of subjects who have an unusual exposure (Boston University, 2016).” By using health records at certain period to distinguish those who exposed to the water in Flint and other area residents who did not involve exposure to the potential toxic water, we can easily check and compare the incidence of disease and then figure out what happened. Through data, such as how many people experienced hair loss or a skin rash in different groups, investigator will calculate odd ratio, deciding whether it is a risk factor or not. Multiple outcomes also can be measured for one exposure.
On the other hand, cohort study design has some potential limitations. First, the cohort study is susceptible to various potential biases, such systematic errors leading false association between exposure and outcome (Jeon, 2016). Selecting study participants who must be free of the exposure of particular sources under investigation is really important because it can create a biased measure of association. There is also observation bias. For example, some people misunderstand the questions and give different responses to the same questions. (Boston University, 2016).
Also, there is data-reliability problem. If records were not designed for the study, the available data may not be a good quality to trust and utilize. Some data that is highly related to exposure or outcome status may be absent or not be recorded until recently. In other words, adjusting for confounding factors is necessary to get a correct answer and conclusion. Lastly, although retrospective cohort study takes relatively less time and cost than prospective cohort study because outcome and exposure have already occurred, it still need more budget and time than case-control study or cross-sectional study. Without adequate funding and guaranteed long investigation time, the cohort study does not be initiated.