Roy Adams
Johns Hopkins University, MD·Computer Science·Early-career
Electronic health records (EHRs), the digital charts hospitals keep for every patient, could reveal a great deal about Alzheimer's disease and related dementias (ADRDs). They could show how the disease varies between people and how it is treated in practice.
- Fields
Computer Science
Medicine & Clinical Research
Mathematics & Statistics
- Topics
Alzheimer's disease
dementia
electronic health records
- Grants
- 1 known grant, through Jul 2027
What they work on
- Develop machine learning methods that combine several imperfect record-based clues for dementia while modeling the error and bias in each, to label whether a patient has Alzheimer's disease or a related dementia.
- Build a continuous electronic health record (EHR) measure of dementia severity that accounts for common error sources such as fragmented care and selection bias, meaning the uneven way patients enter a health system.
- Apply the phenotyping methods to a large Johns Hopkins EHR database and validate them against memory center data, an ongoing longitudinal dementia study, and chart-review severity ratings.
More about the research
But these records are messy. Care is often split across many providers, and the patients who end up in a given health system are not a random sample. Diagnoses are also missed or recorded inconsistently. Errors like these can bias studies and clinical tools, especially for patients who face barriers to getting care.
Why it matters
Studies and clinical tools built on health records can give misleading answers about dementia when records are incomplete or biased. More reliable ways to identify dementia and measure its severity from routine records could support more accurate research on how Alzheimer's disease differs between patients and how it is treated. They could also reduce bias against patients who face barriers to healthcare access.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Tianxi Cai
Harvard University, MA·Mathematics & Statistics
Develop semi-supervised risk prediction methods that combine noisy electronic health record surrogates, such as billing codes, with a small set of chart-reviewed outcome labels, even when predictors are numerous.
Alexis Battle
Johns Hopkins University, MD·Computer Science
Mentor early-stage investigators in study design, data pipelines and quantitative methods for multimodal, longitudinal data in aging and Alzheimer's disease research.
Narges Razavian
New York University, NY·Computer Science
Run a pilot embedded pragmatic clinical trial in primary care to test the acceptability, usability and adoption of an existing AI/machine learning tool and decision support for detecting cognitive impairment.
The school
Johns Hopkins University: 100 labs on PI Pages, $4.1B a year on research, 622 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.