ENHANCING ONCOLOGY REAL WORLD DATA QUALITY THROUGH CLINICALLY RELEVANT EDIT CHECKS IN A LARGE US COMMUNITY ONCOLOGY DATABASE
Author(s)
Debra Rembert, MSN, RN, Christine Spillson, RN, BS, BA, MS, Zhaohui Su, PhD, Christina Lewis, RN, Bryanne Tedesco, RN, Jessica Uranga, RN, Michelle O'Brien, RN, Wendy Haydon, MSN, RN, Susan Stepp, RN, Sandra DiLullo, RN, Jessica Paulus, ScD, Janet Espirito, PharmD, Sarah Spark, RN.
Ontada, Boston, MA, USA.
Ontada, Boston, MA, USA.
OBJECTIVES: Ensuring the accuracy of real-world data (RWD) is essential for safeguarding research and regulatory decision-making. Traditional standard edit checks (TSEC) ensure structural accuracy through logic validation (e.g., date sequencing) and identifying missingness, but often overlook clinical plausibility, creating gaps that may compromise data quality. This analysis evaluated the effectiveness of clinically relevant edit checks (CRECs) in oncology versus TSECs on data accuracy.
METHODS: CRECs were defined as the application of additional plausibility checks on clinical variables identified as requiring enhanced examination. Oncology examples include biomarkers (e.g. correct classification of positive vs negative mutations, cytogenetics), diagnosis, and treatment variables. Three retrospective real world research studies with chart abstracted data conducted from 2024-2025 were used to assess data accuracy measured by number of edit checks produced and queries resulting in data modifications. A representative sample of TSECs were applied across all studies. During the quality control process, data quality specialists utilized clinical expertise to implement CRECs specific to tumor types. Descriptive statistics were used to summarize query counts and correction rate for both TSECs and CRECs.
RESULTS: A total of 1100 charts were included in studies of chronic myeloid leukemia (n=500), prostate cancer (n=300), and cholangiocarcinoma (n=300). TSECs were applied to a subset of 272 charts (25%) across 35 variables. This generated 502 edit checks, resulting in 65 queries. With TSECs, corrections impacted 13% of variables reviewed. When CRECs were applied to the same 272 charts, 616 edit checks were generated, resulting in 249 queries. With CRECs, corrections impacted 40% of variables reviewed.
CONCLUSIONS: Implementation of CRECs identified over three times more corrections than TSECs alone. This approach further strengthens the reliability of RWD. Future work should expand CRECs across tumor types and establish new oncology standard edit checks to enable scalable, quality RWD for research and regulatory decision-making.
METHODS: CRECs were defined as the application of additional plausibility checks on clinical variables identified as requiring enhanced examination. Oncology examples include biomarkers (e.g. correct classification of positive vs negative mutations, cytogenetics), diagnosis, and treatment variables. Three retrospective real world research studies with chart abstracted data conducted from 2024-2025 were used to assess data accuracy measured by number of edit checks produced and queries resulting in data modifications. A representative sample of TSECs were applied across all studies. During the quality control process, data quality specialists utilized clinical expertise to implement CRECs specific to tumor types. Descriptive statistics were used to summarize query counts and correction rate for both TSECs and CRECs.
RESULTS: A total of 1100 charts were included in studies of chronic myeloid leukemia (n=500), prostate cancer (n=300), and cholangiocarcinoma (n=300). TSECs were applied to a subset of 272 charts (25%) across 35 variables. This generated 502 edit checks, resulting in 65 queries. With TSECs, corrections impacted 13% of variables reviewed. When CRECs were applied to the same 272 charts, 616 edit checks were generated, resulting in 249 queries. With CRECs, corrections impacted 40% of variables reviewed.
CONCLUSIONS: Implementation of CRECs identified over three times more corrections than TSECs alone. This approach further strengthens the reliability of RWD. Future work should expand CRECs across tumor types and establish new oncology standard edit checks to enable scalable, quality RWD for research and regulatory decision-making.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
RWD158
Topic
Real World Data & Information Systems
Topic Subcategory
Data Protection, Integrity, & Quality Assurance, Health & Insurance Records Systems
Disease
SDC: Oncology