Epidemiological studies should normally be the most interesting considering human health because they take into account individuals in their environment. Establishing a relationship between an environmental factor and an illness is however a delicate process because a factor can cause disorder in one person and not in another. Moreover, it is difficult to isolate a particular factor in the multitude of factors that constitute our living environment (chemical, physical factors…) and our individual characteristics (genetic and socio-economic factors…).
Therefore, epidemiological studies need to study a large number of people. Based on data, researchers get an overview of the relationship between the studied factor and a disease. Despite its interest in human health, it is important to note that epidemiology gives correlations, rarely causal relationships. If an association is found between a factor and a disease, it does not mean that the factor caused the disease because to be causal, several criteria must be verified (see information page on epidemiological studies).
Other study methods are needed to improve the understanding of the pathophysiological mechanisms and to improve the credibility of epidemiological studies.
Well-conducted in vitro studies can reveal mechanisms of action at the cellular or molecular level that can explain pathophysiological effects. But the results of in vitro studies do not necessarily mean that an effect will be observed in vivo.
It should also be remembered that if the in vitro methods have a high sensitivity (few or no false negatives, i.e. negative results that do not reflect reality: they are false because they should not be negative), they have a low specificity (many false positives, i.e. positive results that should not be). It means that a negative result is probably really negative, but a positive result must be confirmed by other in vitro or in vivo studies.
Results obtained by one method must be confirmed
by other study methods.
From Table 1, it is clear that no method is neither perfect nor infallible. Each of them requires a very rigorous working protocol. Despite their efforts, researchers can hardly take into account all parameters.
Take the example of epidemiological studies: it is needed to not only consider a sufficient number of individuals, which is not always possible in the study of « rare » diseases, but also to identify confounders, to adapt the working protocol… See further informations in Bias of epidemiological studies.
In laboratory studies, it is needed to properly control parameters to which cells or animals are submitted and to work with specific animal or cell models, in accordance with the purpose of the study. Further information are available in in vitro studies and in vivo studies.
According to the difficulty of having perfect experimental conditions and the fact that all parameters cannot be controlled, results of one single study are rather meaningless.
Results of a single study are not sufficient to validate a theory. It is compulsory to replicate the study and to compare with results of other laboratories.