The main clinical features of Alzheimer’s Disease and Alzheimer’s Disease Related Dementias (AD/ADRD) are progressive impairments of cognitive function and changes in behavior. Early intervention may be important for successful disease modification, but there are currently significant limitations in early prediction of cognitive decline and AD/ADRD using standard research and clinical tools as they are not sensitive enough for diagnosis at presymptomatic or early stages of disease. Potentially more sensitive approaches (e.g., neuroimaging, fluid biomarkers, neuropsychological tasks, digital and passive measures) can be expensive, difficult to interpret, or have unclear performance in some individuals and groups, and may require access to academic medical centers, protected databases, or industry partners to ascertain data. Data sources, analytical algorithms, interpretations, and applications of test results have known (and unknown) biases, methodological limitations, and questionable predictive validity, especially in groups that have historically been underrepresented in or excluded from participation in AD/ADRD research.
The National Institute on Aging (NIA), a component of the National Institutes of Health (NIH), seeks to stimulate the use of data resources with appropriate sample diversity, including data relevant to populations disproportionately impacted by AD/ADRD. For example, in Black American older adults, the amount of the protein amyloid – which has long been considered a biological hallmark for AD – might have a smaller role in determining cognitive impairment than other factors such as co-occurring chronic medical conditions (hypertension, diabetes) and sociodemographic and systemic factors, each of which has been found to contribute to racial and ethnic disparities in dementia diagnoses (Wilkins et al., 2022). This highlights the importance of identifying novel biomarkers, including non-biological predictors (e.g., social determinants of health) in adults from underrepresented racial and ethnic groups (Dark and Walker, 2022). The goal of this challenge competition is to inform novel approaches to early detection that might ultimately lead to more accurate tests, tools, and methodologies for clinical and research purposes.
Advances in artificial intelligence (AI), machine learning (ML), and computing ecosystems increase possibilities of intelligent data collection and analysis, including better algorithms and methods that could be leveraged for the prediction of biological, psychological (cognitive), socio-behavioral, functional, and clinical changes related to AD/ADRD.
To make progress, the Challenge aims to address the need for:
- Data from a wider set of sources and types, including data relevant to low-resourced, underserved communities disproportionately impacted by AD/ADRD to better understand and address biases in existing data sources;
- Open, shareable data stored in trusted repositories to determine “distributional robustness” of predictive algorithms; and
- Algorithms that meet “right to explanation” mandates (i.e., if an AI algorithm impacts people, people have a right to an explanation of how AI conclusions were reached).
To meet these needs, the Challenge focuses on identifying, accessing, and using data from sources such as:
- Research datasets, including NIH-supported research (e.g., population-representative longitudinal studies like the Health and Retirement Study, Alzheimer’s Disease Sequencing Project, Alzheimer’s Disease Neuroimaging Initiative, and open datasets), and non-NIH-supported research,
- Real world data like electronic health records, Centers for Medicare & Medicaid Services claims, or from users themselves (e.g., through direct-to-consumer blood-based biomarkers and online cognitive testing),
- Social media or device use, and/or
- Combined data from different sources.
These are only examples, and data need not be limited to sources generated by NIH-supported research.
The overarching goal of this Challenge is to spur and reward the development of solutions for accurate, innovative, and representative early prediction of AD/ADRD. To achieve this goal, the Challenge will feature three phases that successively build on each other.
Phase I: Find IT! The first phase will focus on identifying, accessing and/or building representative, inclusive, open, shareable datasets that can be used for early prediction of AD/ADRD with an emphasis on addressing biases in existing data sources. The primary track will invite solvers to find, curate, and/or contribute existing data that may be used in the Challenge. An alternate track will invite submissions that describe methods for collecting new data in a way that advances the goals of the Challenge.
Phase II: Build IT! Building on the work from data discovery and Phase I, the second phase will focus on advancing state-of-the-art, ethical, and inclusive algorithms and analytic approaches with an emphasis on explainability of predictions.
Phase III: Put IT All Together! A final phase of the Challenge will bring together the top teams and their work from the previous phase to demonstrate algorithmic approaches on diverse datasets for early prediction of AD/ADRD. This phase will include an innovation event where the top teams share their results and pitch their solutions and discuss findings, learnings, and recommendations.
An emphasis of the Challenge is to build teams bringing diverse perspectives to generating solutions that will benefit populations that have been disproportionately impacted by AD/ADRD. Submissions should describe how the proposed solution might address health disparities and should reference the NIA Health Disparities Research Framework to facilitate identifying and proposing tools, technologies, and products that reflect the life course perspective or theory, as well as relevant levels of analysis among the different domains described in the NIA Health Disparities Research Framework.
Anticipated Dates:
- Phase I
- Challenge launch – September 1, 2023
- Executive summary drafts due – January 17, 2024
- Submissions due – January 31, 2024
- Finalists notified – March 18, 2024
- Finalists provide data access for verification – June 1, 2024
- Winners selected and notified after verification – June 17, 2024 through August 30, 2024
- Winners publicly announced – September 2024
- Phase II
- Challenge launch – October 22, 2024
- Model Submissions due – December 19, 2024
- Report Submissions due – January 22, 2025
- Verified winners notified – March 31, 2025
- Winners publicly announced – April 16, 2025
- Tentative Phase III Dates
- Challenge launch – April 8, 2025
- Submissions due – July 1, 2025
- Winners publicly announced – November 2025
Statutory Authority to Conduct the Challenge:
The NIA is conducting this Challenge pursuant to authorities under Section 2002 of the 21st Century Cures Act, 42 U.S.C. 283q, and the America Creating Opportunities to Meaningfully Promote Excellence in Technology, Education, and Science (COMPETES) Reauthorization Act of 2010, as amended, 15 U.S.C. 3719. NIA, part of the NIH, leads a broad scientific effort to understand the nature of aging and to extend the healthy, active years of life. NIA is the primary federal agency supporting and conducting AD research. NIA is conducting this Challenge as a way to advance AD/ADRD prediction and diagnosis. This challenge is also aligned with the objectives of the 42 U.S.C. 283q, which calls on NIH to support challenges in areas of biomedical science that could: 1) realize significant advancements and 2) improve health outcomes in human diseases and conditions, particularly with respect to human diseases and conditions for which public and private investment in research is disproportionately small relative to Federal Government expenditures on prevention and treatment activities, that are serious and represent a significant disease burden in the United States, or for which there is potential for significant return on investment to the United States.
Awards:- $650,000
Deadline:- 23-01-2025