(C-116) Evaluating the Readiness of Saudi Oncology Real-World Data for Standardization and Quality Enhancement

Background: Data standardization and quality are essential in generating reliable real-world evidence (RWE), which support regulatory and clinical decision-making, particularly in cancer drug treatments.However, studies evaluating the standardization and quality of oncology real-world data (RWD), both internationally and in Saudi Arabia, remain uncommon.

Objectives: This study represents the first phase of an initiative to assess the readiness of Saudi multicenter oncology real-world data (RWD) for standardization. Additionally, data quality assessment was conducted with a focus on completeness and accuracy—two essential quality dimensions that determine the reliability of fitness-for-purpose.

Methods: Data were sourced from a centralized database, the Saudi Real-World Evidence Network (SRWEN), pertained to electronic health records (EHRs) from five Saudi tertiary healthcare centers. Patients (≥18 years old) diagnosed with common (breast, thyroid, colorectal) and rare (gastric, hepatocellular and renal) forms of cancer between 2016 and 2023, were included. Readiness for standardization was evaluated by assessing the level of alignment of the existing extracted variables vs. available elements outlined in the framework of Minimal Common Oncology Data Elements (mCODE) initiative (the latter was designed to establish and maintain standardized, computable oncology data formats). Percentage completeness (i.e., the extent to which the variables in the SRWEN contained complete data entries), and accuracy of the available variables was assessed and compared across cancer types and centers. All outcomes were analyzed descriptively using RStudio (Version 1.2.5033).

Results: A total of 20,671 patients were identified (15,683 [67.0%] with common cancer). Data alignment with the mCODE framework varied across centers. Core elements (e.g., patient demographics, cancer diagnosis dates, and comorbidities), were generally available. However, data on cancer clinical status, location, staging, and medications were inconsistently reported across cancer types and centers. Completeness was higher in common vs. rare cancer. Demographic information and diagnosis dates were generally complete. Certain variables, such as cancer clinical status, exhibited extremely low completeness (below 1%). Overall data accuracy was high, with rates exceeding 80% across evaluated domains.

Conclusions: The findings reveal considerable variability in oncology data format across centers within the SRWEN. While data accuracy is generally high, critical gaps in completeness exist, particularly concerning clinical status and outcomes. Addressing these issues is essential for enhancing RWD's utility in supporting research and regulatory decisions in cancer care.