About usefulness of GERAADA score

We read with interest the first clinical validation of the new GERAADA score performed by Luehr et al. [1]. However, we would like to contribute with some reflections about their conclusive statement, with which the authors promote a prospective clinical trial to further evaluate the score as a useful tool to allow for improved decision-making.

The GERAADA score has been intended to predict the 30-day mortality in patients undergoing surgery because of acute aortic dissection. The GERAADA score represents not only the first score dedicated for this scope but also the first score for emergency cardiac surgery ever concepted [2]. In fact, all the prior scores used were conceived for elective interventions. Remembering that acute aortic dissection is a life-threating condition with not much alternative to surgery, and that both clinicians and patients suffer under the pressure of time, we think that colleagues in the heart surgery community should question themselves about the usefulness of this powerful tool. While the scores for elective surgeries can be useful to better plan the interventions, selecting more experienced operators for cases with higher expected mortality, this type of strategy is not applicable for emergency interventions, where often the surgeon on call is alone. The implicit—and never explicated—question in the article so far is the following: can an intervention be refused on the basis of a mathematical evaluation? We are convinced that some reflections are indispensable due to the ethical implications.

A first reflection should be made on the fact that every score—however, precise it is—is always a snapshot of the past (i.e. of the patients enrolled between 2006 and 2015) that does not take into account the continuous surgical and medical technical-scientific progress. Second, it should be reflected about the precision of the score, namely its discrimination (which is measured through the area under the curve). The GERAADA score showed a mean area under the curve between 0.673 and 0.712 [1, 2], which in statistics is considered a ‘poor/fair’ prediction, and is far from the value of 0.8, which is considered ‘good’ [3]. To conclude, there will always be variables, which are not considered in the mathematical model and which instead can be decisive for the final decision not to operate (e.g. presence of a malignant metastatic tumour) or to operate in any case even in the case of very high risk (e.g. the firm will of the patient to be operated).

Probably, the best use of the new GERAADA score should not be the patient selection or the decision-making but rather the quality control and the performance comparison between different hospitals for the retrospective evaluation and better resource management purposes.

REFERENCES