Skip to main content

Data Analysis Services

Data analysis services are provided by the CHEAR Data Repository, Analysis, and Science Center (Data Center), which supports CHEAR by developing and maintaining an extensive exposure data repository, designing and carrying out appropriate data analyses, and developing data science resources used to create combinable datasets for public use.

The Data Center has established data submission and data sharing policies to maximize both the value and security of study data. See Data Submission Agreement and Data Sharing Plans for more details.

The following sections explain the Data Center’s services in more detail.

1. Data Repository

The Data Center repository holds all data generated by the CHEAR Lab Hubs along with pre-existing epidemiologic data provided by CHEAR study principal investigators (PIs).

These data include, but are not limited to:

  • The environmental exposure biomarkers and biologic response data generated by the CHEAR Lab Hubs
  • Previously collected epidemiologic data such as key covariates (e.g., sex, race/ethnicity, age, education)
  • Previously measured biomarkers (including clinical biomarkers)
  • Genetic data
  • Health outcome information

In addition to managing the repository, the Data Center provides the following services:

2. Data Analysis Plan Development

The Data Center works with investigators to develop a data analysis plan that includes:

  • A description of study design and sample size
  • A description of datasets to be analyzed, including:
    • Datasets provided by the investigator
    • Data generated by the Lab Hubs
  • A statement of analysis objectives
  • The type of dependent variable (e.g., continuous scale, binary, time-to-event with censoring)
  • Identification of appropriate covariates/confounders
  • Examination of the Data Dictionary and Codebook to:
    • Understand the variables and their characteristics
    • Link the PI’s epidemiologic data to biomarker results produced by the CHEAR Lab Hubs
  • Consideration of complex correlations among environmental chemicals and indications of whether to focus on select chemicals or evaluate mixture effects
  • A description of how planned analyses will deal with missing data
  • Evaluation of a modeling strategy for:
    • Fitting the available data (e.g., nonlinear or linear regression, logistic regression)
    • Evaluating goodness of fit of the data with the estimated model
    • Selection of statistics for testing for association (e.g., likelihood ratio test, Wald test)

3. Statistical Analysis

The Data Center conducts statistical analyses according to the data analysis plan developed in collaboration with the PI. Services include:

  • Performing or assisting in the proposed statistical analyses based on the modeling strategy chosen in the data analysis plan
  • Designing and populating tables and figures

4. CHEAR Ontology Services

The Data Center is responsible for creating and maintaining the CHEAR Ontology—a common vocabulary for use in the CHEAR program. The Ontology is evolving with the program and will connect to best-in-class existing vocabularies, thus facilitating the integration of data from multiple studies. The Data Center assists PIs in applying the Ontology to their studies. Services include:

  • Facilitating the mapping of variables from data dictionaries into terms consistent with the CHEAR Ontology
  • Incorporating the study's data into the CHEAR Ontology to support collaborative research across the CHEAR consortium, including pooled analyses from cohort studies participating in CHEAR
  • Developing methods and services for comparing similar variables from different data dictionaries, starting with very basic mappings of equivalent terms and moving into more sophisticated analyses of relationships among variables
  • Providing tools and services to manage the CHEAR Ontology evolution

5. CHEAR Knowledge Graph Services

A knowledge graph is a network that connects a wide range of information types relevant to a specific domain, such as children’s environmental health. The CHEAR Knowledge Graph will:

  • Make it possible to query all types of information defined in the CHEAR Ontology
  • Enable users to browse and query the CHEAR Data Center Repository utilizing the vocabulary from the CHEAR Ontology
Page last updated: 
July 19, 2017