P-001: COMMON DATA MODEL FOR SICKLE CELL DISEASE SURVEILLANCE: CONSIDERATIONS AND IMPLICATIONS

Abstract

Purpose: With no existing population-based national registry, there is a lack of comprehensive understanding of the sickle cell disease (SCD) population living in the United States. The Sickle Cell Data Collection Program (SCDC), funded by the Centers for Disease Control and Prevention (CDC), is addressing this need by developing state-level surveillance programs. SCDC currently operates in 11 states, maintaining identifiable individual-level data within each state. However, there is no data model to guide the harmonization of data collected for SCD public health surveillance. We describe an ongoing process for establishing and maintaining a common data model in the unique setting of public health surveillance and identify key elements for surveillance reporting on SCD. Materials and methods: The core data sources common across SCDC states include newborn screening, vital records, and Medicaid claims. However, many programs have additional data including electronic medical records (EMR), state all-payer data, and clinical cohorts. The SCDC Common Data Model will standardize key data elements by identifying data instruments, specific variables and formats. Key data instruments are shown in Figure 1. We are developing data hierarchy guides to standardize variables across multiple data sources. The creation of this Common Data Model is overseen by a committee from SCDC states with expertise in clinical care, claims data, epidemiology, database construction, and data analysis, with input from SCD champions. Results: The SCDC has identified a core set of results that are feasible for most states to attain, are measured with reasonable accuracy, and have high public health and policy relevance. These Core Surveillance Data are reported annually by SCDC participant states and compiled by CDC. Birth Information: We report numbers of SCD births in 1-year and 5-year increments; by sex, race, ethnicity, and SCD genotype. Newborn screening records are the primary source, with state laboratory confirmatory testing used to identify SCD genotype. Case Numbers: We estimate the number of prevalent cases across the state during a given date range and report by county, sex, and age group (<10, 10-19, 20-29, 30-39, 40-49, 50-59, 60+ years). The primary source for these tables is healthcare claims datasets linked with newborn screening and clinical case reports. Cases are identified as possible, probable, or confirmed SCD using validated definitions. Deaths: Deaths are reported by age (<20, 20-49, >=50 years) and stratified by sex. State vital records are a core source of death information, followed in priority by Medicaid data, and clinical data when available. Healthcare Utilization: Core Surveillance Data reports acute care utilization including number of emergency visits, inpatient hospitalizations, hospital length of stay and visit’s payer type. Results are reported by age group (<10, 10-19, 20-29, 30-39, 40-49, 50-59, 60+ years) and payer type (e.g., Medicaid, self-pay). Conclusion: To standardized national SCD surveillance, we are implementing a common data model that will allow SCDC to move towards a distributed data network. In this report, we describe the process of developing and maintaining a common data model in this unique setting and identify key data elements for public health reporting on SCD.C. TAKEMATO declares a conflict of interest: Consultancy, Expert: Data Safety Committee member: Novartis trial for Eltrombopag for aplastic anemia Research support/Scientific studies: Cure Sickle Cell Initiative: funding for sickle cell cohort identification GBT pharmaceuticals: voxelotor studies in sickle cell disease Daiichi Sankyo: anticoagulation for pediatric thrombosis Novartis: Eltrombopag for immune thrombocytopenia A. SNYDER declares a conflict of interest: Research support/Scientific studies: DD20-2003 Sickle Cell Data Collection support from the Centers for Disease Control and Prevention