Review Logic for Phenotypic Review

Source of data as measure of quality

Manual data entry provides the highest quality of phenotyping data but, due to limited resources and time, this is not possible for all sites. There are three levels of phenotyping data that exist within the Epi25 phenotyping dataset and these are graded based on the source of information.

GRADE I: Sub-phenotype provided to the Broad Institute with the sample, no additional clinical information exists in Epi25 phenotyping database.
GRADE II: Phenotyping information exists in the Epi25 database and was imported from an external clinical database.
GRADE III: Phenotyping information exists in the Epi25 database and was provided manually by completion of REDCap phenotyping forms.

Review of inclusion/exclusion criteria (GRADE II & GRADE III)

Three sub-phenotypic groups are included in Epi25 and each group has a specific set of inclusion/exclusion criteria. The data quality review process involves checking the Epi25 phenotyping dataset to confirm that the cases included in the study meet these criteria.

Specific fields within each sub-phenotype are assessed to verify each inclusion/exclusion criteria. Depending on the data entered in that field, the case will either be passed, require manual review or, if there is missing data in that field, require follow up with the site to request the information be provided. Note that to pass overall, a case must pass each specific field that is assessed.

Responses that require manual review are first checked by the phenotyping coordinator and then by a clinician if further input is required. The manual review process requires reviewing all the available information about a case and determining if there is enough evidence to support their assignment to a specific sub-phenotype. If the data is concerning or further clarification is needed, the phenotyping coordinator will reach out to the site PI.