(409) Development of common data elements for maternal health research: Results from a Delphi consensus process

Common data elements (CDEs, i.e., standardized questions paired with allowable responses) are critical for providing consistency in data collection and promoting research collaborations. This study aims to determine a minimum reporting set of CDEs by NICHD grantees conducting maternal health (MH) research.


Study Design:

NICHD convened a 64-expert panel including representatives from large research networks, ACOG, SMFM, CDC, HRSA, and NIH. Experts formed 2 working groups supporting development of biomedical and psychosocial CDEs. Two modified Delphi processes, each with 2 voting rounds, were undertaken. In Delphi 1 (03/24-05/24; 84% response rate), experts voted on MH constructs for standardized data collection categorized as: Tier 1 (all MH research), Tier 2 (thematic MH research areas), or neither. Constructs for which >80% of experts affirmed the importance of standardization were retained for categorization to Tier 1 or 2 based on majority vote. The Tier 1 list was restricted to constructs with >75% of experts identifying the construct as Tier 1, forming the minimum set of constructs for all MH research. For measuring these constructs, in Delphi 2 (07/24-08/24; 70% response rate), experts used 5-point Likert scales to rate the feasibility and validity of specific CDEs identified via landscape analysis and literature searches.


Results:

Experts voted on 267 biomedical and 194 psychosocial constructs; 202 biomedical (146 Tier 1; 56 Tier 2) and 144 psychosocial (90 Tier 1; 54 Tier 2) constructs were retained. Restriction to constructs with >75% Tier 1 votes yielded 35 biomedical and 22 psychosocial constructs in the minimum set of constructs for all MH researchers (Table). A total of 89 biomedical and 115 psychosocial CDEs were rated by experts during Delphi 2. Of these, depending on available data source (EHR, survey, other), the minimum reporting set includes 50-60 CDEs/CDE bundles.


Conclusion:

Reporting of CDEs across MH research is expected to improve data quality, study consistency, data interoperability, and our understanding of how to improve MH.