NCIRD provides subject matter expertise for 19 nationally notifiable conditions for which NCIRD programs rely on the electronic data in NNDSS. In support of the CDC Surveillance Strategy and the CSTE call for data standardization and harmonization (15-EB-01 “Common Data Structure for NNDs”), NCIRD participates in the NNDSS Modernization Initiative (NMI) with mumps and pertussis. In addition, to address provisioning challenges for NNDSS historical data already received at CDC, NCIRD initiated and supports the NNDSS Data Accessibility Project (NDAP).
METHODS:
For mumps, pertussis, and other conditions (e.g., varicella) for which message mapping guides (MMGs) are being developed, NCIRD Surveillance Office compares and aligns variables from all data sources (e.g., worksheets, older MMGs, NETSS, NBS, electronic laboratory reporting, immunization information systems) to identify similar concepts across conditions. Opportunities for standardization and harmonization are described and NCIRD subject matter experts then select appropriate harmonization options. For NDAP, NCIRD staff (e.g., epidemiologists, informaticians, SAS/SQL coders, data managers) are collaborating on data store access, validation, and analysis to provide historical data to CDC programs in HL7, NETSS, and NBS formats.
RESULTS:
As of April 2015, there were 110 total mumps MMG variables and 118 total pertussis MMG variables. Of those, 81 (74% of mumps, 69% of pertussis) were harmonized. Of 26 total mumps vaccine variables and 24 total pertussis vaccine variables, 24 (92% mumps, 100% pertussis) were harmonized. All of the 49 total lab variables on each MMG were harmonized. The previous (2009) varicella MMG had 161 total variables, of which 110 (68%) were not harmonized. NCIRD harmonization for varicella (October 2015) identified 120 total varicella variables; 94 (78%) were harmonized with mumps, pertussis, or other source. Of 22 varicella vaccine variables and 54 varicella lab variables, 22 (100%) and 52 (96%), respectively, were harmonized.
As of December 2015, SAS/SQL code is being written to validate and analyze NDAP mumps, pertussis, and varicella data, although production data is not yet available.
CONCLUSIONS:
Harmonization is possible across clinically and epidemiologically distinct conditions, particularly for laboratory and vaccine related variables, when supported by Center surveillance infrastructure, cooperation of program SMEs, and collaboration with NMI partners. The methods are extensible across infectious and non-infectious program areas but may be challenged by lack of variable repositories as scope broadens. Although NDAP may address access to historical data, benefits of NMI cannot be fully realized until relational data from new harmonized MMGs can be provisioned to CDC programs.