Home › Research › How common it is
Evaluating Clinical Foundation Models for Early Alzheimer's Disease and Related Dementia Prediction from Longitudinal EHRs
Farzana, S.; Arian, A.; Rundek, T.; Desvarieux, M.; Ahsan, H.. medRxiv (preprint), 2026-09-03
Topic: Amyloid, tau and blood-test diagnosis
Early identification of Alzheimers disease and related dementias (ADRD) remains challenging despite its importance for timely intervention, management of modifiable risk factors, and care planning. We developed and evaluated ADRD onset prediction models using longitudinal electronic health records (EHRs) from the All of Us Research Program at clinically meaningful lead times of 6, 12, 24, and 36 months before diagnosis, benchmarking interpretable count-based representations against four publicly available pretrained clinical foundation models (CLMBR-T, GPT-style, LLaMA-style, and Mamba) across multiple ADRD phenotype definitions. Count-based models consistently achieved the highest discrimination and calibration across all cohorts and prediction horizons. Predictive performance declined with increasing lead time for all approaches; however, the performance gap between count-based and pretrained representations progressively narrowed, with foundation models achieving comparable AUROC of 0.719 (compared to the AUROC of 0.738 of count-based model) at the 36-month horizon while providing higher sensitivity and F1 scores under a fixed operating threshold. External validation with zero-shot evaluation on UChicago EHRs exhibited limited generalizability for count-based and pretrained clinical foundation model based representations. These findings demonstrate that transparent count-based EHR representations remain the strongest overall approach for ADRD onset prediction, while pretrained clinical foundation models provide complementary advantages for long-term risk identification and establish a benchmark for evaluating transferable clinical representations in temporal ADRD risk prediction.
Cite this study APA, Vancouver, BibTeX · download .bib or .ris
APA 7th
Farzana, S., Arian, A., Rundek, T., Desvarieux, M., & Ahsan, H. (2026). Evaluating Clinical Foundation Models for Early Alzheimer's Disease and Related Dementia Prediction from Longitudinal EHRs. medRxiv (preprint). https://doi.org/10.64898/2026.09.01.26361933
Vancouver
Farzana S, Arian A, Rundek T, Desvarieux M, Ahsan H. Evaluating Clinical Foundation Models for Early Alzheimer's Disease and Related Dementia Prediction from Longitudinal EHRs. medRxiv (preprint). 2026. doi:10.64898/2026.09.01.26361933.
BibTeX
@article{farzana2026-8ac1b7a2,
author = {Farzana, S. and Arian, A. and Rundek, T. and Desvarieux, M. and Ahsan, H.},
title = {{Evaluating Clinical Foundation Models for Early Alzheimer's Disease and Related Dementia Prediction from Longitudinal EHRs}},
journal = {medRxiv (preprint)},
year = {2026},
doi = {10.64898/2026.09.01.26361933},
url = {https://doi.org/10.64898/2026.09.01.26361933}
}
Built from the record alzhub holds for this paper, not looked up. Volume, issue and page numbers are not held in the record. Check the citation against the original before you use it.
Key findings
- Predictive performance declined with increasing lead time for all approaches; however, the performance gap between count-based and pretrained representations progressively narrowed, with foundation models achieving comparable AUROC of 0.719 (compared to the AUROC of 0.738 of count-based model) at the 36-month horizon while providing higher sensitivity and F1 scores under a fixed operating threshold.
- These findings demonstrate that transparent count-based EHR representations remain the strongest overall approach for ADRD onset prediction, while pretrained clinical foundation models provide complementary advantages for long-term risk identification and establish a benchmark for evaluating transferable clinical representations in temporal ADRD risk prediction.
Why it matters
This study compared models for predicting Alzheimer's disease and related dementias from electronic health records and found that simple count-based models performed best overall, with performance declining at longer prediction horizons. It is about risk prediction methods and does not test any treatment, cure or prevention.
New to this subject? Start with our plain-English guide: Alzheimer's in numbers →
Study details
- Study designlongitudinal study
- Evidence strengthModerate
How this study scored 16/100
Preprint — not yet peer-reviewed. It scored 34 on the same rubric, but non-peer-reviewed sources are capped at 16, so it shows as 16. Scored zero for sample size, effect reported, studied in people and randomised / blinded because the record does not state it. We have no data on impact for its field and year, so that was left out: it earned 32 of the 94 points we could judge it on, which is 34 out of 100.
| Component | Points | Why |
|---|---|---|
| Journal tier | 2 / 10 | Tier 4 — not a peer-reviewed journal (medRxiv (preprint)) |
| Journal's citation record | 0 / 4 | Not a journal article, so the journal's record does not count |
| Study design | 7 / 12 | A longitudinal study following the same people over time |
| Strength of evidence | 3 / 5 | Rated “moderate” in our summary |
| Peer-review validation | 4 / 10 | Preprint — not yet peer-reviewed. Hosted on a preprint server. |
| How recent | 5 / 5 | 2026 — within 1 year(s) of the newest study here |
| Sample size | 0 / 4 | Participant count not stated — scores zero (not stated in the record — scored 0) |
| Effect reported | 0 / 3 | No effect size reported — scores zero (not stated in the record — scored 0) |
| Studied in people | 0 / 3 | Not stated whether this was studied in people — scores zero (not stated in the record — scored 0) |
| Randomised / blinded | 0 / 2 | No randomisation or blinding stated — scores zero (not stated in the record — scored 0) |
| Times cited | 0 / 7 | Cited 0 times (OpenAlex, 2026-10-10) |
| Citations per year | 0 / 5 | 0.0 citations a year over 2 year(s) since 2026 |
| Impact for its field and year | — / 6 | No field-weighted citation impact (OpenAlex has none yet) — left out of this score |
| Cited by studies here | 0 / 6 | No other study in this library cites it |
| Still being cited | 0 / 4 | Cited 0 times in 2024–2026 (2026 still in progress; OpenAlex) |
| Senior author's track record | 4 / 4 | Last-listed author Habibul Ahsan: h-index 102 (OpenAlex, whole career) |
| Institution's track record | 2 / 3 | Columbia University: institution h-index 1,824 (OpenAlex) |
| Multi-centre collaboration | 2 / 3 | OpenAlex lists 4 institution(s) in 1 country |
| Indexed in PubMed | 1 / 2 | Not in PubMed, but has a DOI |
| Free to read | 2 / 2 | Free to read |
| Points | 32 / 94 | Scored on 19 of 20 parts. We have no data for the rest, so they are left out: 32 of the 94 points we could judge it on is 34 out of 100. |
| Total | 34 / 100 | Capped to 16 because this is not a peer-reviewed article. |
What each of these 20 components measures, and what this score is not →
Read the original
Full text: free to read — OpenAlex lists this copy as free to read: the link above (checked 6 Feb 1020).
Studies this one cites
Referenced by this study and also in the alzhub library, so you can read them here.
Related studies in How common it is
Also tagged “classification”
835 studies in the library carry the topic tag “classification”; these are this study's neighbours in a fixed, alphabetical-by-hash tour of them, not the closest in content. All 835 tagged studies.