Page 47 - PATIENT REGISTRY DATA FOR RESEARCH: A Basic Practical Guide

Chapter 4 - Approach for Statistical Analysis

It is important to always bear in mind that prior to subjecting the registry data to the

proposed statistical analysis, researchers must be clear about the intended study objectives

and the significance of all the study variables in relation to the objectives of the study. All

statistical analyses should be conducted based on the research question and the study

objectives. These analyses should be conducted only after data cleaning has been performed

and the data are properly organized for the subsequent statistical analyses. Data cleaning for a

patient registry database can often be a very difficult task because of the sheer volume and

the wide variety of data that can possibly be collected by a registry.

Therefore, the function of data preparation of patient registry data (which includes

data collection, data preparation and data analysis) is usually performed by experienced

statisticians or their equivalents. Say, for example, a researcher or data analyst is in charge of

conducting the statistical analysis of patient registry data; it is then likely for him/her to be

mentored and supervised by a senior statistician or a statistical consultant. In addition, the

researcher or data analyst should also be well-versed in the handling of patient registry data,

and be equipped with adequate prior knowledge in the design of these registries (including

their clinical parameters, definitions of all the variables), the design of its case report form,

the procedures for data collection and data preparation for statistical analysis of all the patient

registry data, all of which have already been discussed previously.