Data Science

AR² – Autism Replication, Validation, and Reproducibility Center

Posted on by Jodi Snare

Page last updated February 4, 2026

Study Design: Retrospective Observational Trial
PCORnet Infrastructure: Common Data Model (CDM), Patient partners or engagement
Principal Investigator: Judy Zhong
Institution: Weill Cornell Medicine
PCORnet® Network Partner: INSIGHT
Funder: NIH
Funding Date: 2025
Study Duration: 2025 – 2028
Participating PCORnet® Clinical Research Networks: ADVANCE, GPC, INSIGHT, PEDSnet, STAR
Therapeutic Area: Intellectual and Developmental Disabilities (IDDs); Data Science
Status: Active, not recruiting

Research Question(s):

The study focuses on making autism research more reliable and widely applicable. We are creating the Autism Replication, Validation, and Reproducibility (AR²) Center to ensure that findings from autism data science projects can be independently verified and reproduced by providing complete, transparent packages, including data, code, and documentation.

By using diverse datasets from across the U.S., we aim to confirm that research results hold true for different populations and settings. This effort will promote best practices, share resources openly, and help translate autism research into real-world clinical care and policy.

Transfer Learning NLP to Improve Adoption of Clinical Text in Multi-Site Studies

Posted on by Jodi Snare

Page last updated April 14, 2026

Study Design: Other, Data Science
PCORnet Infrastructure: Common Data Model (CDM), Patient partners or engagement
Principal Investigator: Yonghui Wu
Institution: University of Florida and University of Florida Health
PCORnet® Network Partner: OneFlorida+
Funder: Patient-Centered Outcomes Research Institute (PCORI); (Project webpage)
Funding Date: 2024
Study Duration: 2024 – 2027
Participating PCORnet® Clinical Research Networks: INSIGHT, OneFlorida+
Therapeutic Area: Data Science
Status: Active, not recruiting

Research Question(s): Can large language models - an advanced form of artificial intelligence (AI) - improve the generalizability of patient information extraction and clinical phenotyping across different healthcare systems?

Semantic Data Quality Standards for Multi-Center Clinical Research Studies and Networks

Posted on by Jodi Snare

Page last updated April 14, 2026

Study Design: Other, Methods to improve study design, methods to support data research networks
PCORnet Infrastructure: Common Data Model (CDM), Single IRB, Patient partners or engagement, Clinical Research Collaboration Agreement
Principal Investigator: L. Charles Bailey
Institution: The Children's Hospital of Philadelphia
PCORnet® Network Partner: PEDSnet
Funder: Patient-Centered Outcomes Research Institute (PCORI); (Project webpage)
Funding Date: 2021
Study Duration: 2021 – 2026
Participating PCORnet® Clinical Research Networks: PEDSnet
Therapeutic Area: Data Science
Condition: Data quality assessment, data quality analysis, data quality reporting, standards development
Status: Active, not recruiting

Research Question(s):

  1. Can we find ways to more accurately describe how suitable data are to answer a specific research question?
  2. What are the tools that can be used across studies to consistently describe whether the data are high quality?

Primary Publication(s):

Razzaghi H, Dickinson K, Wieand K, et al. A multifaceted approach to advancing data quality and fitness standards in multi-institutional networks. J Am Med Inform Assoc. 2025;ocaf181. doi:10.1093/jamia/ocaf181