Paper Reading Journals #8
Dementia risk analysis using temporal event modeling on a large real-world dataset
Basic Info
| Item | Content |
|---|---|
| Title | Dementia risk analysis using temporal event modeling on a large real-world dataset |
| Journal | Scientific Reports |
| Author | R. Andrew Taylor, Aidan Gilson, Ling Chi, et al. |
| Year Published | 2023 |
| Link | https://doi.org/10.1038/s41598-023-49330-8 |
| Geography of Study | USA |
| Area of Study | Dementia, EHR |
Description
(Including data and methods)
Data: 442,278 patients in New England health system. Method: (1) Temporal Event Analysis - Fisher’s Exact Test in Bonferroni & Bernoulli Trail (2) trajectory clustering (3) Dynamic Time Warping
Main Contribution
(Innovative method or Application) Innovation Type: Application innovation. This paper applied temporal statistical analysis to identify dementia-related risk events that occur prior to the onset of dementia. It combined temporal event modeling and clustering methods to uncover the time-dependent features of dementia progression. This approach may improve the explainability and interpretability of dementia occurrence.
Data Source
Emergency Department record in New England health system with n=442,278
Favorite Thing
This paper provides a potential modeling approach to trace dementia pathways before onset by incorporating temporal features. We may consider adopting this method in our future modeling framework.