brief history of EuroSCORE

The EuroSCORE Project was begun by Samer Nashef and François Roques working together between Cambridge and Martinique in the French West Indies. We gathered data on some 20,000 patients from 128 hospitals in eight European countries. We collected information on 97 risk factors in all the patients. We therefore knew the risk profile of the patients, and we knew which had survived. That information was analysed by biostatistician Philippe Michel and, from that analysis, we devised a simple additive risk model to predict the risk of death for cardiac surgery, published in 1999 ^(1,2).

A few years later, computer technology was becoming widely available and there was no longer a need for the simple additive model, as everyone had access to computers which could calculate the risk more precisely using the full logistic equation which, if you are interested, is:

predicted mortality = $$ {e^{(β_0 +Σβ_iX_i)} \over 1+e^{(β_0 +Σβ_iX_i)}}$$

…where β₀ is the constant of the logistic regression equation and X_i=1 if a categorical risk factor is present and 0 if it is absent. For age, X_i=1 if patient age <60 and increases by one point per year thereafter. The logistic EuroSCORE was published in 2003 ⁽³⁾.

The success of EuroSCORE surpassed all expectations. The model has been used globally, in every continent and in most countries with cardiac surgery. The original paper alone has been cited in more than 4,000 scientific publications, and the term EuroSCORE has entered the medical vocabulary. As a result of better measurement of risk-adjusted clinical outcomes, cardiac surgical results rapidly improved with risk-adjusted mortality falling by over half, with the greatest improvement being seen in countries and hospitals with the most robust monitoring. This meant that within a few years, the model became out of date in that it significantly overpredicted mortality, and it was time to refresh the model. Another data collection took place gathering information on over 22,000 patients from all over the world. The project was led by Samer Nashef and managed by Christopher Smith. The result was a new risk model, EuroSCORE II, developed with the help of biostatistician Professor Linda Sharples and published in 2012 ⁽⁴⁾.

Anticipating that EuroSCORE II may also become outdated, work on EuroSCORE 3 has begun, and we invite hospitals with an interest in participating in this project to register that interest here.