Page 52 - PATIENT REGISTRY DATA FOR RESEARCH: A Basic Practical Guide

including the approach to be adopted for statistical analysis of registry data (Please see

Table 5.1).

4.2 Presentation of the Final Results

The overall presentation of the final results should be formatted in such a way that

will directly address the research objectives which were set earlier. Therefore, all tables and

figures obtained from statistical analysis should be presented clearly in order to effectively

convey its key message to its wide target audience, including all the stakeholders. There are

various ways of presenting findings. The proper way for presenting the results shall depend

on the research objective, scope of the proposed research study and type of target audience.

Most importantly, it is strongly recommended for both the statistician and researcher

to have a clear idea of both the ideal and expected results, before commencing the statistical

analysis. By doing so, the statistician can develop a versatile data analysis plan which will

ideally be built on the research protocol. Having a clear plan of action for data analysis will

also guard against data-driven results, which is important for research integrity and quality.

Hence, the drafting of the dummy tables should be done collaboratively by both statistician

and researcher, which is important for ensuring that the proposed data analysis will be

feasible, specific and focused.

Since it can be a difficult task for a novice or inexperienced researcher to draft these

dummy tables, it will be recommended for him/her to refer to other dummy tables found in

previously published research articles, especially those from peer-reviewed journals, as a

guide for them to follow when they are drafting a new dummy table. To illustrate this point,

this e-book/book lists down a total of 33 published scientific articles found in peer-reviewed

journals and were written by various authors, all of which had reported on research studies

conducted using patient registry data (Please see Appendix 1).