In this cohort study of 60,507 hospital admissions, we found that although the SPM marginally outperformed existing prediction scores in balanced accuracy for classification of sepsis, it suffers from poor timeliness, limiting its clinical application for sepsis diagnosis and treatment.

As with all questions of testing performance, the balance between missed true cases and overtreatment of false positives must be weighed.

In the case of sepsis, prioritization of timely treatment is paramount, given the potentially severe consequences when the diagnosis is missed or delayed.

Irrespective of the clinical utility of the SPM as a prediction model, sepsis remains an area that is underresearched and underdeveloped from the perspective of prediction modeling.

The current tools continue to leave enormous gaps in our ability to fully determine which patients need urgent treatment for sepsis and reduce the high burden of associated negative health outcomes.

Read More