CaRi-Heart for predicting cardiac risk in suspected coronary artery disease: early value assessment

Prognostic ability

How well CaRi‑Heart predicts cardiac risk in a UK population, especially in the following groups:

The AI‑Risk algorithm predicted cardiac death and MACE in the whole cohort and for people without obstructive CAD. AI‑Risk appeared to overestimate risk in people with obstructive CAD. AUC results indicate that adding AI‑Risk to QRISK3 and CAD‑RADS 2.0 improved its prognostic ability for cardiac mortality and MACE.

The confidential results on the prognostic performance of the AI-Risk algorithm in people from different demographic and socioeconomic backgrounds are available from Chan et al. (2024b).

Clinical decision making

How CaRi‑Heart results affect clinical decision making compared with UK standard clinical practice

In a real-world study, using CaRi‑Heart results led to changes in clinical management in 45% of people, which included:

Clinical outcomes

How clinical outcomes might change for people with suspected CAD who have had CaRi‑Heart testing and appropriate treatment

There is no direct evidence on how this could result in changes to clinical outcomes. Linked-evidence modelling has been used to establish the effect size of statins on different cardiovascular outcomes in different risk groups.

Costs

The costs to the NHS of using CaRi‑Heart

A model was developed to compare AI‑Risk in addition to standard care versus standard care over a person's lifetime. Using linked-evidence modelling, a simulation of 5,000 people after implementing CaRi-Heart estimated a reduction in events of:

ICERs were calculated using 3 different costs for AI-Risk: £300, £500 and £700 per scan. ICERs were £1,371, £2,307 and £3,244, respectively.

The budget impact over 5 years was calculated assuming £700 per scan and gradual uptake up to 20% by year 5. This came to £30,169,955 for year 5.