A Nationwide Japanese Medical Claims Foundation Model: Balancing Model Scaling and Task-Specific Computational Efficiency
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917433003999232 |
|---|---|
| author | Aratake, Nanae Tosaki, Taisei Okamoto, Yuji Uchino, Eiichiro Nakamura, Masaki Matsui, Nobutomo Hatakama, Akiko Okuno, Yasushi |
| author_facet | Aratake, Nanae Tosaki, Taisei Okamoto, Yuji Uchino, Eiichiro Nakamura, Masaki Matsui, Nobutomo Hatakama, Akiko Okuno, Yasushi |
| contents | Clinical risk prediction using longitudinal medical data supports individualized care. Self-supervised foundation models have emerged as a promising approach for leveraging large-scale unlabeled healthcare records. In natural language processing, scaling laws suggest that larger models achieve predictably lower pretraining losses, supporting the foundation model paradigm. However, for structured medical data, characterized by a limited vocabulary and sparse observations, whether increasing model size consistently improves downstream predictions is unclear, as most studies evaluate only a single model scale. In this study, we evaluated the relationship between model scale and downstream task performance for structured medical foundation models. Using a random sample (2.3 million patients, 32 hospitals) from a nationwide 519-hospital Japanese claims database, we pretrained encoder-only Transformers at five scales (2.2M-101M parameters) for disease incidence and medication prediction. Downstream performance saturated at task-dependent thresholds: disease prediction benefited from larger models (32M-101M), whereas medication prediction saturated at 11M, reducing pretraining time by 178 h. Across all tasks, the best-performing model consistently outperformed a Light Gradient Boosting Machine baseline in the area under the precision-recall curve. These findings indicate that, unlike the monotonically decreasing pretraining loss, the optimal model size varied depending on task characteristics. This task-dependent saturation provides practical guidance for balancing predictive performance and computational cost in structured medical foundation models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_22348 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | A Nationwide Japanese Medical Claims Foundation Model: Balancing Model Scaling and Task-Specific Computational Efficiency Aratake, Nanae Tosaki, Taisei Okamoto, Yuji Uchino, Eiichiro Nakamura, Masaki Matsui, Nobutomo Hatakama, Akiko Okuno, Yasushi Machine Learning Clinical risk prediction using longitudinal medical data supports individualized care. Self-supervised foundation models have emerged as a promising approach for leveraging large-scale unlabeled healthcare records. In natural language processing, scaling laws suggest that larger models achieve predictably lower pretraining losses, supporting the foundation model paradigm. However, for structured medical data, characterized by a limited vocabulary and sparse observations, whether increasing model size consistently improves downstream predictions is unclear, as most studies evaluate only a single model scale. In this study, we evaluated the relationship between model scale and downstream task performance for structured medical foundation models. Using a random sample (2.3 million patients, 32 hospitals) from a nationwide 519-hospital Japanese claims database, we pretrained encoder-only Transformers at five scales (2.2M-101M parameters) for disease incidence and medication prediction. Downstream performance saturated at task-dependent thresholds: disease prediction benefited from larger models (32M-101M), whereas medication prediction saturated at 11M, reducing pretraining time by 178 h. Across all tasks, the best-performing model consistently outperformed a Light Gradient Boosting Machine baseline in the area under the precision-recall curve. These findings indicate that, unlike the monotonically decreasing pretraining loss, the optimal model size varied depending on task characteristics. This task-dependent saturation provides practical guidance for balancing predictive performance and computational cost in structured medical foundation models. |
| title | A Nationwide Japanese Medical Claims Foundation Model: Balancing Model Scaling and Task-Specific Computational Efficiency |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2604.22348 |