MGH301 Epidemiology and Biostatistics with Special Reference to Social Epidemiology

The process of understanding how to include social epidemiological theories into a project can be met with a lot of complexities. Methods which can even provide the most evidenced based on other health interventions is often inappropriate when issues such as mental illness is under investigation (Berkman, Kawachi, & Glymour, 2014).That is to say, it cannot be approached in the absence of social dimensions especially in societies with different social norms.

Therefore, a vigorous and a comprehensive method should be applied to get results that can be used as inferences for the whole population. Koenen (2014), argued that the extent to which other social constructs influence study participants needs to be carefully observed. One can achieve this by trying to link which specific theory may influence how data is being collected and analyzed. Social factors affecting an individual is not independent of the population and the context in which social variables are analyzed should be predetermined by an appropriate methodology (Oakes & Kaufman, 2006).

Berkman et al. (2014) further suggests that one must have a deeper understanding of the social factors as far as putting them into studies are concerned. For example a suitable and a most appropriate study design should be chosen when dealing with mental health issues. In this case, a retrospective case control studies is well appropriate for such a study because it can be a good way of tracking back how predictors or exposures are connected to health outcomes (Koenen, 2014). In addition to that, it also gives an idea of how other social factors could be related and a large sample size is required in order to account for dropouts. However, in order to determine the correct sample size, a pilot study is often necessary to be conducted (Campbell, M., Machin, D., & Walters, S. 2010) especially when dealing with issues like mental health that has limited data in low income countries (Sankoh, Sevalie, & Weston, 2018).

Since the issue of mental illness is often regarded as unmentionable in most settings, targeting school going children might be faced with a lot barriers. With the outcome of our event being binary in nature, logistic regression will be used and other covariates will be adjusted for in the multiple logistic regression.

One of the major problems associated with studies within the paradigm of social epidemiology is its inability to detect for the minimum differences of a health outcome (Berkman et al. 2014) and there could be a reverse causal relationship between mental illness and school dropouts (Oakes & Kaufman, 2006). This is definitely a problem especially when dealing with mental illness because of the densities of its procedures. Another prominent challenge is the inadequate human and financial resources needed to carry out follow ups and in circumstances where the data is to be collected digitally, its reliability and consistency is often question (Koenen, 2014). Moreover, this becomes burdensome in situations where mental health outcomes such as depression, anxiety or stress are to be observed in participants. Berkman et al. (2014) further states that extrapolating studies in other parts of the world could be challenged with local norms and values.

For example many studies on mental health are conducted in the west and cannot be used in low income countries unless it is domesticated into that context which sometimes become difficult. This is because the context in which these people live are totally parallel in nature. Kuh, Ben-Shlomo, Lynch, Hallqvist, and Power (2003), suggests that the analysis of social variables is demanding especially if the data exists as multilevel or repeated data in relation to how other social factors influences our health.

In conclusion, the method to use in any given social epidemiological study should be carefully examined. This is important as it will prevent coming up of with wrong conclusions. Since the field of social epidemiology is relatively a new area, a lot of researches needs to be done in order to address the challenges mentioned above.