Development and validation of a prediction model of length of ICU stay for benchmarking: a retrospective cohort study

Hideki Endo, Satoru Hashimoto · 2026

Read Development and validation of a prediction model of length of ICU stay for benchmarking: a retrospective cohort study by Hideki Endo, Satoru Hashimoto (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/pmc-13473596.txt.

Description

Hideki Endo and Satoru Hashimoto built in 2026 a prediction model for intensive-care stay length, designed for fair benchmarking across Japanese hospitals. Efficient intensive-care use depends on knowing how long patients should stay, yet existing models perform unevenly and may not fit Japan. The cohort drew on the Japanese Intensive Care Patient Database, covering adults admitted to 87 units between April 2022 and March 2023, with stay in days as the outcome.

Additive models using severity score, disease category, emergency surgery, and admission source were compared, and a gamma model with a log link performed best, explaining 41.5 percent of deviance in cross-validation. Applied to hospitals, the model yielded standard stay ratios and observed-minus-expected stays, shown in funnel plots. Correcting for extra variation cut the share flagged beyond 95 percent control limits from 69 percent to 9.2 percent. The authors present the model as a basis for comparing intensive-care use.

Catalog record

CREATORHideki Endo, Satoru Hashimoto
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13473596

Browse the full library · About the Tower of Babel