Comparison of Methods for Incorporating Related Data When Developing Clinical Prediction Models: A Simulation Study
Haya Elayan, Matthew Sperrin, Glen P Martin, Frieder Braunschweig, Jonas Faxén, Joakim Alfredsson, David A Jenkins · 2026
Read Comparison of Methods for Incorporating Related Data When Developing Clinical Prediction Models: A Simulation Study by Haya Elayan, Matthew Sperrin, Glen P Martin, Frieder Braunschweig, Jonas Faxén, Joakim Alfredsson, David A Jenkins (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13520826.txt.
Description
Haya Elayan and colleagues simulate in 2026 when extra data helps or harms clinical risk models. Extra sets may come from other times, lands or care settings. Rules say models should use target group data. Spare sets still tempt model builders.
Fake target and helper groups varied in likeness and size. Four build plans were compared on fit, sorting skill and stability. Borrowing data helped in most runs. Recalibration often beat target-only builds. One recalibration variant proved shaky. A weighting plan stayed steady across most scenes. A Swedish heart registry case shows the lessons at work in real care.
Catalog record
| CREATOR | Haya Elayan, Matthew Sperrin, Glen P Martin, Frieder Braunschweig, Jonas Faxén, Joakim Alfredsson, David A Jenkins |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |