VALIDATION OF DATA QUALITY IN THE MALAYSIAN PATIENT REGISTRY INFORMATION SYSTEM FOR TYPE I PAEDIATRIC DIABETES CASES

Abstract

INTRODUCTION
Type 1 Diabetes Mellitus (T1DM) is a significant chronic condition in children, yet Malaysia lacks comprehensive population-based data on its prevalence and clinical features. To address this, the Patient Registry Information System – Non-Communicable Disease (PRIS-NCD) was developed within the Malaysian Health Data Warehouse (MyHDW) to enable longitudinal data collection. This study aimed to evaluate the validity of PRIS-NCD data following pilot implementation in a national paediatric referral centre.

METHODOLOGY
One hundred twenty paediatric patients with T1DM under follow-up in Hospital Putrajaya were retrospectively notified into the PRIS-NCD registry. We assessed data validity by comparing 47 variables between registry entries and abstracted electronic medical records (EMR). The analysis focused on exact agreement rates and missing data percentages to determine concordance and completeness.

RESULT
Of 120 cases, 115 were included in the analysis. The mean exact agreement between the registry and EMR data was 95.4% at diagnosis and 94.7% at follow-up. Most variables showed agreement rates exceeding 90%, except for BMI at diagnosis (86.8%), insulin test at diagnosis (88.6%), and microalbuminuria at follow-up (68.3%). Missing data were generally low in both datasets, with registry data showing slightly fewer missing values compared to EMR data despite being a secondary source.

CONCLUSION
The PRIS-NCD registry demonstrated high validity in recording T1DM patient data, supporting its use for epidemiological insights and healthcare planning. Rigorous data entry procedures likely aided high agreement rates. However, a broader evaluation of completeness, comparability, and timeliness will require nationwide implementation and real-time data capture. We recommend continued improvements in data entry practices and user education to sustain and enhance data quality.