Adults with primary, spontaneous, non-traumatic intraparenchymal haemorrhage confirmed on CT. The ICH Score was derived in 152 adults at a single US centre (1997–1998) to predict 30-day all-cause mortality.

ICH Score	Observed 30-day mortality in derivation cohort
0	0%
1	13%
2	26%
3	72%
4	97%
5	100%
6	No patients; risk presumed extremely high

External validation in nine cohorts (≈3,800 patients) shows generally good discrimination for mortality (pooled c-statistic around 0.8) across diverse settings.
Calibration is variable; contemporary, maximally treated cohorts often have lower mortality at a given score than the original UCSF cohort.
Guidelines recommend using a severity score such as the ICH Score at presentation but caution against basing treatment limitation or prognostic decisions on a single score because of withdrawal-of-care and self-fulfilling prophecy bias.

Use this score early after presentation to support discussions about prognosis, level of care and neurosurgical input alongside neurological examination, comorbidities, haematoma expansion risk and patient values.

The primary source or current guidance provides the model definition, source population and supporting evidence.

Intracerebral Haemorrhage (ICH) Score