Risk score predicts critical care needs in patients with intracerebral hemorrhage

Now we just need a protocol that will reduce this risk. WHOM will be doing that? Where is the stroke leader that will solve that problem? Describing a problem with no solution is useless. But since we have fucking failures of stroke associations

nothing will be done.

Risk score predicts critical care needs in patients with intracerebral hemorrhage

A novel risk prediction score identified which patients with intracerebral hemorrhage are at low risk for critical care with a high degree of specificity, according to findings from a retrospective cohort study published in Neurology.

Roland Faigle, MD, PhD, assistant professor of neurology at John Hopkins University School of Medicine, and colleagues developed the Intensive Care Triaging in Spontaneous Intracerebral Hemorrhage (INTRINSIC) score to measure predictors of critical care.

“In clinical practice, the INTRINSIC score may be most useful when predicting the absence of critical care needs in order to identify which patients may be triaged to a non-ICU setting. Each score cut-point has a different sensitivity-specificity trade-off,” Faigle and colleagues wrote. “In order to prioritize patient safety, it is desirable to avoid falsely classifying patients as not needing critical care when in fact they do. In addition, high specificity may be particularly desirable when contemplating potential patient transfer (or no transfer) to tertiary centers.”

The score involves a 0-to-9-point system where a systolic blood pressure (SBP) of 160-190 mm Hg earns 1 point, SBP greater than 190 mm Hg earns 3 points, a Glasgow Coma Scale (GCS) of 8 to 13 earns 1 point, a GCS less than 8 earns 3 points, intracerebral hemorrhage (ICH) volume 16 to 40 cm³ earns 1 point, ICH volume greater than 40 cm³ earns 2 points and the presence of intraventricular hemorrhage (IVH) earns 1 point.

Researchers applied the point system to 451 patients with ICH (mean age, 62 years [range, 54-77 years]; 54.1% men). They separated patients into development and validation cohorts, with the risk score applied to the validation group.

Of the 451 patients with ICH, 80.3% received critical care interventions. The study results demonstrated that GCS, SBP, ICH volume and IVH independently predicted critical care need in the development cohort. The most common critical care services involved IV medication infusions for uncontrolled hypertension (67%), mechanical ventilation (47.5%), hyperosmolar therapy for cerebral edema (47.5%) and external ventricular drain placement (22.8%).

The INSTRINSIC risk score applied to individuals who did not require critical care during their hospital stay identified patients with 95.8% specificity. Among patients scoring 0 with no ICU care during their ED stay, 94.4% did not need critical care later. Moreover, 83.3% of patients with a score of less than 2 and no ICU care during their time in the ED did not need critical care.

“We therefore propose a cut-point that predicts absence of critical care needs with high specificity (low false-positives), such as a score of less than 2, which predicted the absence of critical care with 88.5% specificity in the external validation cohort,” Faigle and colleagues wrote. “With increasing resource constraints, such as the height of the COVID-19 pandemic when open ICU beds are a rarity, a higher score cut-point such as less than 3 could be considered.”

In a related editorial, Matthew B. Maas, MD, MS, associate professor of neurology (stroke and neurocritical care) and anesthesiology at Northwestern University’s Feinberg School of Medicine, discussed the urgent need for better service allocation due to the increased demand for critical care during the COVID-19 pandemic. Using Faigle and colleagues’ INTRINSIC score for ICU risk triage as an example, Maas argued that support tools for ICU admissions can be employed to overcome cognitive biases.

“Like other prognostic scores, the INTRINSIC Score is best used alongside other clinical considerations, since there will always be factors that are relevant on an individual level, but aren’t common enough to emerge as statistical predictors in a model,” Maas wrote.