The Diabetes Prevention Program (DPP) was a randomized clinical trial designed to prevent type 2 diabetes (T2D) in adults with prediabetes. The Diabetes Prevention Program Outcomes Study (DPPOS) is the 30-year follow-up of that cohort, focusing on T2D, prediabetes, and related complications. Cognitive assessments within DPPOS began in 2009 and were expanded in 2022 to examine cognitive impairment, including Alzheimer’s disease (AD) and AD-related dementias (ADRD), among surviving participants. To align DPPOS with national standards for AD/ADRD research and to enable data sharing with the National Alzheimer’s Coordinating Center (NACC), the team implemented the NACC-UDSv3 standardized data framework in 2022.
The integration aimed to preserve the longitudinal integrity of DPPOS while adding standardized AD/ADRD-specific items, and to streamline cognitive outcomes adjudication through automated reporting. The source reports that implementing NACC-UDSv3 into an established longitudinal cohort not originally designed for dementia research was a primary objective of this effort.
The implementation process began with a comparison between the 16 NACC-UDSv3 data forms and existing DPPOS items. Overlapping and similar items collected in DPPOS were identified and harmonized to avoid duplication while maintaining data fidelity. Missing or non-overlapping elements required by NACC-UDSv3 were added to the DPPOS data collection to produce a dataset that conformed to NACC-UDSv3 specifications.
The harmonization process prioritized mapping each required NACC-UDSv3 item to equivalent DPPOS measures when possible; when direct equivalents did not exist, new items were introduced. The goal was to create a dataset that could be shared with NACC and used within the DPPOS research workflows without disrupting established longitudinal measures.
The harmonized forms were adapted for electronic data capture (EDC) using MIDAS (Multimodal Integrated Data Acquisition System) hosted at George Washington University. MIDAS was configured to collect both the harmonized NACC-UDSv3 items and the existing DPPOS cognitive and clinical measures.
Automated reports were developed within MIDAS to support cognitive adjudication. These reports integrated current and prior neuropsychological test scores to present consolidated information for adjudicators, streamlining review and decision-making. The automated reporting was designed to reduce manual data reconciliation and to provide a consistent presentation of longitudinal cognitive performance for each participant.
In the first wave of the DPPOS-AD/ADRD study using the harmonized dataset implemented in MIDAS, the team completed 1,561 cognitive adjudications successfully. The source reports that this demonstrates the operational feasibility of integrating NACC-UDSv3 into a long-standing cohort study and of using electronic capture and automated reporting to support adjudication workflows.
No additional quantitative performance metrics, error rates, or time-to-adjudication statistics were reported in the source. The successful completion count is the implementation outcome presented.
The DPPOS-AD/ADRD project illustrates that a standardized, national AD/ADRD data framework can be incorporated into an ongoing longitudinal study that was not originally designed for dementia research. Key components of this integration included systematic harmonization of overlapping items, addition of missing standardized items, adaptation of forms for EDC, and development of automated adjudication reports.
The authors suggest that the combination of harmonization, electronic capture, and automated adjudication may serve as a practical framework for other cohorts seeking to adopt NACC-UDSv3 and to align with national AD/ADRD research standards. The source frames these processes as potentially generalizable operational steps for similar long-term studies wishing to enable AD/ADRD-focused analyses and data sharing with NACC.
The source confirms that the study followed relevant ethical guidelines, obtained institutional review board approvals at participating centers, and collected written informed consent from participants prior to initiating study procedures. The trials referenced are registered on ClinicalTrials.gov (Diabetes Prevention Program: NCT00004992; Diabetes Prevention Program Outcomes Study: NCT00038727; Diabetes Prevention Program Outcomes Study AD/ADRD: NCT05704309), as reported in the source.
The datasets used for the implementation were de-identified, and the source states that participant identifiers were not available outside the research group.
Funding for the work reported in the source included multiple U01 grants from the National Institute of Diabetes and Digestive and Kidney Diseases and a U19 award from the National Institute on Aging, as listed in the source. Specific grant numbers and sources are provided in the original article.
Competing interests disclosed in the source include consulting and other relationships for select authors: Jose A. Luchsinger disclosed consultancy for Merck KGaA and Novo-Nordisk, receipt of study drug/placebo from Merck KGaA for an NIH-funded trial, royalty payments for a book, and editorial stipend; Owen Carmichael disclosed NIH and Eli Lilly Corporation grant support. The remaining authors reported no conflicts of interest in the source.
The implementation reported in the source shows that NACC-UDSv3 can be integrated into an established longitudinal cohort like DPPOS through item-level harmonization, addition of missing standardized elements, EDC adaptation via MIDAS, and automated adjudication reporting. The first-wave completion of 1,561 cognitive adjudications demonstrates operational feasibility. The authors propose that this approach — harmonization, electronic capture, and automated adjudication — may provide a practical model for other cohorts aiming to adopt national AD/ADRD data standards and to facilitate data sharing with NACC.
Note: The source is a preprint posted on medRxiv and has not been peer-reviewed. Additional implementation details, validation metrics, or long-term operational outcomes were not reported in the source document.