Page 36 - PATIENT REGISTRY DATA FOR RESEARCH: A Basic Practical Guide

recode the existing variables which are commonly found in patient registry databases, just

like those found in any other databases.

For example, a new variable called body mass index (BMI) can be derived from

height and weight calculations, and then recoded into a set of varying categories according to

a pre-specified classification system. While standardized classification systems are readily

available for some variables such as BMI and stages of chronic kidney disease; many other

variables do not have such standardized classification systems, and therefore it will be

necessary to classify them by using other means.

For instance, age can be categorized into either a '5-year 'or '10-year' age groups.

However, since there is usually a widely-accepted way of categorizing a variable, the

researcher may review the existing literature to search for the most common way of

categorizing a variable; which will then enable him/her to make a comparison of these

variables between different studies. For example, the age group can be categorized according

to an international standard which is widely used (WHO, 2007; WHO, 2008; Bujang et al.,

2012). Once these new variables in a registry data set have been computed and/or recoded,

these data should again be checked for accuracy in a validation step to ensure both the

computation and the recoding of these data are correct.