2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Sunday, December 13, 2026
06:00 p.m. - 08:00 p.m.
MCCNO - Hall F
Access, Representation, Population Health, Technology and Procedures, Serologic Tests, Blood Banking, Treatment Considerations, Supportive Care, Biological Therapies, Transfusion, Gene Therapy, Non-Biological Therapies, Opioids, Workforce, Artificial Intelligence (AI), Machine Learning, Clinical Practice (Health Services and Quality), Clinical Guidelines, Health Economics, Patient Safety, Practice Models, Quality Improvement, Quality of Care, Systems-Based Hematology, Telemedicine, Diseases, Hemoglobinopathies, Sickle Cell Disease, Metabolic Disorders, Iron Overload, Research, Clinical Research, Health Outcomes Research, Patient-Reported Outcomes, Real-World Evidence, Registries, Study Population, Human, Adult, Pediatric
PAPER

4360 Development of an operational electronic health record-based registry to identify patients with sickle cell disease across an integrated health system

Madhav Vissa, MD, Erik Stewart2, Karen Grant3, Beth Anne Martin4, Marsha Treadwell, PhD5
1 University of California, San Francisco, Benioff Children's Hospitals, Oakland, California, United States
2 University of California San Francisco, San Francisco, California, United States
3 University of California, San Francisco, Benioff Children's Hospitals, Oakland, California, United States
4 University of California San Francisco, San Francisco, California, United States
5 University of California, San Francisco, Benioff Children's Hospitals, Oakland, California, United States

Background 
Population-based surveillance from the California Sickle Cell Data Collection (SCDC) Program estimates that more than 1,000 individuals with sickle cell disease (SCD) reside in the San Francisco Bay Area, yet only approximately a third receive care through the region's largest comprehensive SCD program at UCSF Benioff Children’s Hospital (BCH) Oakland. Clinic-based registries may underestimate patients with SCD interacting with a health system, whereas diagnosis-based electronic health record (EHR) registries may overestimate them because of misclassification of sickle cell trait. We sought to develop a reproducible EHR-based registry capable of identifying patients with SCD across UCSF to support clinical operations, quality improvement, and equitable access to comprehensive SCD care. 

Methods 
A broad Epic parent registry was first generated using all SCD- and sickle cell trait-related diagnosis concepts. Patients with isolated SCT encounter diagnoses were excluded after validation against physician-adjudicated cohorts, while patients with both SCT and non-trait SCD diagnoses were retained because miscoding of confirmed SCD as SCT was common.  

A curated reference cohort was assembled by integrating existing physician-adjudicated datasets, including the ASH Research Collaborative Data Hub, prior institutional chart review, and targeted clinician adjudication, minimizing redundant manual review. Patients were then hierarchically stratified by receipt of specialty SCD care, frequency of non-trait encounter diagnoses, and other structured evidence of SCD (including referral, invoice, and hospital account diagnoses). Observed confirmation rates among manually reviewed patients within each stratum were used to estimate the projected composition of the remaining patients in that stratum and guide iterative refinement of the operational registry. Encounters dating as far back as 2012 (including conversion encounters from a historical instance of Epic at BCH Oakland prior to merging with UCSF) were included.  

Results 
The parent registry identified 3,534 patients with sickle cell-related diagnoses. Exclusion of patients who only carried SCT diagnoses markedly reduced over ascertainment while preserving patients with confirmed SCD who also carried incidental SCT diagnoses. 

Manual review demonstrated that the likelihood of confirmed SCD increased progressively with increasing structured EHR evidence. Patients evaluated in dedicated SCD clinics had the highest confirmation rate (318/351 [90.6%] manually reviewed; stratum n=1000), followed by patients with ≥2 non-trait encounter diagnoses (115/139 [82.7%] manually reviewed; stratum n=182), 1 non-trait encounter diagnosis (67/115[58.3%] manually reviewed; stratum n=174), and no non-trait encounter diagnoses (30/278 [10.8%] manually reviewed; stratum n=651 total). Excluding this final group would omit a relatively small number of patients with confirmed SCD while substantially reducing false-positive ascertainment, a tradeoff mitigated by the dynamic nature of the registry. Patients subsequently receiving appropriately documented SCD encounter diagnoses are automatically incorporated. Patients with ≥1 non-trait encounter diagnosis demonstrated a high projected probability of true SCD while maintaining broad case ascertainment, so this threshold was selected for the operational Epic registry. The final Epic registry identified 1,100 patients, including 1,040 who were also identified by the curated reference cohort. 

Conclusions 
Development of a high-fidelity operational SCD registry required extensive iterative refinement, yet the resulting implementation relies on straightforward EHR logic that can be operationalized and maintained. This approach leveraged existing physician-adjudicated cohorts, minimizing redundant manual review while enabling a continuously updated operational registry. Registry refinement intentionally prioritized sensitivity over specificity because false-positive patients could subsequently be removed through targeted review, whereas missed patients would not be identified for outreach or quality initiatives. Rather than relying solely on specialty clinic populations, this registry enables health systems to identify patients with SCD who interact with the health system outside dedicated SCD programs, creating opportunities for outreach, evaluation of engagement in comprehensive SCD care, and population-level quality improvement. 

Madhav Vissa, MD: Novo Nordisk - Honoraria; CSL Behring - Research Funding; Novo Nordisk - Consultancy (Includes expert testimony); Agios - Consultancy (Includes expert testimony)