The HEP score was developed as an expert-derived pre-test probability model for immune HIT in adults receiving unfractionated or low molecular weight heparin in hospital. It is intended for patients with new thrombocytopenia and/or thrombosis where HIT is a realistic differential diagnosis.

Study / setting	Patients	Reference standard	Key HEP findings
Cuker et al. 2010, single-centre derivation/validation	50 adults referred to a HIT reference laboratory	Expert panel adjudication plus PF4/heparin ELISA and serotonin release assay	HEP showed greater interobserver agreement and discrimination than the 4Ts (AUC ~0.91 vs 0.74). A screening cut-off of HEP ≥2 was 100% sensitive and ~60% specific for expert-defined HIT, and modelling suggested about a 40% reduction in direct thrombin inhibitor use compared with usual practice.
Joseph et al. 2015, external validation	51 inpatients with suspected HIT (mixed medical, surgical and ICU)	Confirmed HIT defined by positive ELISA and serotonin release assay with expert adjudication	HEP and 4Ts had similar discrimination (AUC ~0.73–0.74). HEP ≥2 remained 100% sensitive but only ~16% specific for confirmed HIT, so low scores were useful to rule out HIT but many patients with higher scores did not have HIT.
Besch et al. 2022, surgical ICU pilot	119 surgical ICU patients with suspected HIT (6 confirmed)	Independent committee using ELISA, platelet aggregation testing and serotonin release assay	HEP ≥5 had sensitivity 100% and specificity ~92% (AUC ~0.97) and outperformed the 4Ts; no patient with confirmed HIT had HEP <5, suggesting that in this cohort HEP <5 effectively ruled out HIT.
Kram et al. 2022, mechanical circulatory support	374 HIT evaluations in 299 adults on mechanical circulatory support (IABP, VAD, ECMO and related devices)	Composite of PF4/heparin ELISA optical density ≥1.0 or positive serotonin release assay	Both HEP and 4Ts had only modest discrimination (AUC ~0.63–0.68). At recommended cut-offs (for example HEP ≥3), sensitivity was low and a low HEP score did not reliably exclude HIT in patients requiring mechanical circulatory support.

Prospective comparison work suggests that overall diagnostic accuracy of HEP and 4Ts is broadly similar, with HEP potentially performing better among less experienced clinicians and in some ICU cohorts. All studies emphasise that pre-test probability scores should be combined with PF4/heparin antibody testing and, where available, platelet activation assays.

Low HEP scores (typically ≤1) generally correspond to a very low probability of a positive platelet activation assay and can help avoid unnecessary HIT testing in stable ward patients. Scores ≥2 identify patients who warrant further evaluation, while scores ≥5 fall in a high-probability, high-specificity range in published cohorts. Absolute post-test risk still depends on the clinical context and local HIT prevalence.

The score has not been validated in paediatric populations, pregnancy or outpatients, and most local pathways still reference the 4Ts as the primary pre-test tool. The HEP score can be used alongside the 4Ts where available, particularly when intermediate 4Ts results leave uncertainty.

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

HIT Expert Probability (HEP) Score for suspected immune heparin-induced thrombocytopenia