Page 24 - Recommendations-for-biostatisticians-in-managing-and-conducting-medical-research-consultations

etc. Thus, we can now clearly see from Figure 2 what sort of clinical parameters will have to
be tested in the TEV model. If the proposed research is a prospective study, then CF will
guide the researcher in preparing a scientifically valid and robust questionnaire for
identifying those variables of interest which appear to be the risk factors contributing to
the development of each clinical parameter.

However, if the proposed research is a retrospective study in which all the variables
have already been pre-specified, then the CF should be in line with the data available in the
dataset. In this case, the researcher will not be able to exercise freedom in determining
which of the variables of interest can also be identified as the risk factors for this disease.
Instead, he/she will still be allowed to make the most out of the available variables found in
the dataset, by identifying certain patterns which appear to emerge from the dataset and
then drawing conclusions from them. This commonly occurs in the analysis of secondary
data such as the patients’ case-notes or the patient registry data.

From Figure 2, it seems apparent that drawing the CF will be very useful for
research proposals that aim to (i) test specific hypotheses with regard to an
evaluation of the association between a dependent variable and an independent
variable, to (ii) identify risk factor(s) for a disease and to (iii) set up a disease model.
The ultimate aim for drawing the CF is to convert the research objective into a
simplified figure which will be easier to understand. In order to draw a valid CF that
can facilitate a biostatistician to be better able to gain a sound understanding of the
background subject matter prior to offering a biostatistical consultation, it is very
important to first establish an effective two-way communication between the research
client and the statistical consultant (i.e. biostatistician).

Once a valid CF has been developed, then it will then be much easier for the research
client to decide which variables are identified as ‘independent variables’, and which
variables are identified as ‘dependent variables’, and what other variables will also be
collected. By gleaning an understanding of the concept of CF and the importance of its
applications, the biostatistician will be equipped with the skill to understand the
significance of the research idea, which shall then be directed to the underlying research
problem and also lead to formulation of study objectives. This is an important skill for a
biostatistician to master in order to enable him/her to use the technique of CF to
summarize the research objective into a simplified figure.