What question did this study set out to answer?

To prospectively validate five published multivariate prognostic models for acute stroke and compare their accuracy against two simple clinical variables.

May 1, 1992Journal of Neurology Neurosurgery & PsychiatryOpen Access

Predicting the outcome of acute stroke: prospective evaluation of five multivariate models and comparison with simple methods.

View Full Paper

Ask AI

Bookmark

Authors

JGJohn GladmanNottingham University Hospitals NHS Trust DHDavid HarwoodUniversity of Maryland, College Park DBDavid BarerUniversity of Edinburgh

Discussion

Member takes

Implication

Prospective study finds simple clinical signs match complex multivariate models for predicting stroke outcomes, indicating limited clinical utility for complex algorithms.

Key Points

To prospectively validate five published multivariate prognostic models for acute stroke and compare their accuracy against two simple clinical variables.
Prospective validation cohort of 102 consecutive stroke patients admitted to a district general hospital.
Evaluated five multivariate models (Belfast, Guy's Hospital, Uppsala, Edinburgh, Bristol) alongside two basic clinical measures: conscious level on admission and urinary continence at four weeks.
Guy's Hospital and Uppsala acute models predicted death with 75% accuracy compared to 65% for admission conscious level alone, whereas the Belfast model achieved 50% accuracy.
Urinary continence at four weeks predicted long-term recovery with accuracy similar to the multivariate Edinburgh model, while the Bristol model performed poorly.
Multivariate models showed substantially lower accuracy in prospective testing than in their initial development, with only the Uppsala model offering modest utility for trial stratification.

Cite This Study

Gladman et al. (1992) studied this question.

synapsesocial.com/papers/6a7caecea64782d1c66ee7d0 https://doi.org/10.1136/jnnp.55.5.347

View Full Paper

Ask AI

Bookmark

Implication

Prospective study finds simple clinical signs match complex multivariate models for predicting stroke outcomes, indicating limited clinical utility for complex algorithms.

Predicting the outcome of acute stroke: prospective evaluation of five multivariate models and comparison with simple methods.

Discussion

Member takes

Implication

Key Points

Cite This Study

Predicting the outcome of acute stroke: prospective evaluation of five multivariate models and comparison with simple methods.

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider