Page 30 - PATIENT REGISTRY DATA FOR RESEARCH: A Basic Practical Guide

offer a better method for patient identification by indicating his/her patient identification

number and also specifying which eye he/she is having problems (Mohamad-Azaz & Goh,

2018).

After having compiled a list of criteria to identify a case as a duplicate, the researcher

will then need to adopt a systematic approach for removing the duplicates. Using NDR report

as an example to illustrate the situation in which a patient may have been enrolled in more

than one centre on two different dates, the investigator may decide to retain only one copy of

a patient's medical record in the registry, by basing either on time (for example: to keep the

one prepared on the earliest or most recent date) or (ii) by basing on the content of the

medical record (for example: to keep the one containing more information).

For the former, the selection can be made very easily (and probably even quicker) by

using an IT software. However, it may pose a risk of losing vital patient information if the

selected patient's medical record is not the one having the most complete set of information

pertaining to the patient's medical condition. In contrast, selecting the record containing the

most complete set of information will be very time-consuming since it will require a very

careful evaluation of each medical record before omitting them one at a time. After removing

all the duplicates, it is necessary to perform a validation step to ensure there are no more

duplicates to be found in the data set.

An efficient way to avoid duplicate data being entered into the registry is to create a

mechanism in the system in which a duplicate will be detected as soon as the same patient's

identifiers (such as his/her name or identity card number) have been entered into the registry

more than once. By using this approach, the database system will keep a permanent record of

a patient's data whenever his/her data have been entered into the registry database, which will

then alert the investigator as soon as the data from the same patient are being entered into the

registry again.