SOME STUDY DESIGN SPESIFIC STATISTICS

This is a short bonus chapter, in which we first discuss two of the most simple and commonly used statistical methods in clinical research:

1. Relative Risk

2. Odds Ratio.

Then, in the last section of the chapter, we present to you a little comparison between Incidence and Prevalence in statistical terms.

The scope of this study design within clinical research is to outline one or more Risk Factors for a certain outcome event (e.g. disease). It follows the exposure forward in time leading up to the outcome. The results of cohort studies are usually expressed as Observed Frequencies of exposure and outcome. The table template below outlines this concept.

Fitting your dataset to this table allows you to estimate the Risk of developing the disease of interest within the population from which your sample was drawn:

Estimated Risk= Number of Participants Developing the Disease During the Study Period Cohort's Total Number=a + bn

However, more important to the aims of cohort studies, you may want to estimate the risk of developing the disease for the exposed and the unexposed participants separately.

Fortunately, this can be done using the same principle of Estimated Risk:

Estimated Riskexp= aa + c

Estimated Riskunexp=bb + d

The Relative Risk (RR) is the statistic that describes the association of the exposure and the disease. It can be assessed by comparing the estimated risk of disease for each of the two groups (exposed and unexposed): RR= Estimated RiskexpEstimated Riskunexp=a / (a + c)b / (b + d)

An RR = 1 means the exposure has no effect on the disease outcome as both the exposed and the unexposed participants developed it with the same proportion.

An RR > 1 means the exposure may be a Risk Factor of the disease outcome; and an RR < 1 means the exposure may be a Protective Factor against the disease outcome.

When assessing RRs, always look for the CI of the calculation and the strength of the association (is the RR double, triple, more).

Like the Cohort Studies, this study design is also concerned about Risk Factors. However, this design goes back from the outcome and outlines all possible exposures that may have led to it.

This fact is what prevents us from calculating an absolute risk in case-control studies. Instead, we calculate the Odds Ratio (OR) from the observed frequencies:

The Odds Ratio is a ratio of the odds of being a case in the exposed to those of being a case in the unexposed groups.

The odds are the quotient of probabilities of the cases over those of the free of the disease in the group (i.e Oddsexp= aa + c ca + c=ac ; Oddsunexp=bb + ddb + d=bd ).

Therefore, Odds Ratio can be given by: OR=a / cb / d=a db c

OR indicates the odds associating a certain exposure to a disease outcome. An OR = 1 means the odds are the same for exposure and non-exposure. An OR > 1 means the odds of disease development in the presence of the exposure is higher; and an OR < 1 may mean that the exposure is protective against the disease.