Are acute coronary syndromes risk models too complex? reply

We thank Drs Gale and Manda for their interest in our study.1 We believe it is important to explore why validated risk scores are often not applied in the ‘real world’, and concur that their perceived complexity may constitute the greatest barrier to more widespread use. However, our findings should not be construed as promoting one risk score over another–rather, our study highlights the important and inevitable tradeoffs between complexity and accuracy. We agree that age and haemodynamic variables are the most powerful prognosticators. Although risk scores incorporating only these variables are purported to be ‘simpler’,2 in reality, their application still requires the use of a calculator and a nomogram for conversion into an estimated risk of adverse events. Thus, it remains unclear whether these ‘simpler’ risk scores are necessarily more user-friendly and less time-consuming, compared with the more ‘sophisticated’ ones. For example, the GRACE risk score calculator, which consists of readily available clinical information, is easy to use, and can be readily downloaded onto a PDA or accessible on the website.3 A major strength of the GRACE risk score is its applicability across the full spectrum of acute coronary syndromes. Because reperfusion therapy should be promptly administered to all patients with ST-elevation myocardial infarction in the absence of contraindications (although the optimal type of reperfusion therapy may depend on clinical presentation and local availability), accurate risk stratification is more relevant in the initial management of non-ST-elevation acute coronary syndrome, which represents a more heterogeneous condition with a variable prognosis. We chose all-cause mortality as our primary study outcome because it was the most robust endpoint. Furthermore, surveillance for myocardial (re-)infarction and the decision to proceed with ‘urgent’ revascularization, especially in the short-term, were probably influenced by physicians' risk assessment. Finally, randomized controlled trials have shown that an early invasive strategy improves long-term outcome.4 Therefore, risk stratification tools that can identify patients with worse long-term outcome are most useful in guiding treatment decisions. Of note, the TIMI risk score demonstrates better discrimination for mortality than the composite endpoint, even in the original derivation cohort.5 Thus, our conclusions appear to be robust and not critically dependent on the chosen endpoint. With respect to the correlations among the risk scores and physicians' assessment, we agree that the highly significant P-values were expected. However, the important point is that there were only weak to moderate correlations―a substantial proportion of patients would be classified into different risk categories, according to these three risk scores and physicians' assessment. This may account for the treatment-risk paradox observed.6 The most important implication of our study is that systematic application of any validated risk score in routine clinical practice will likely improve risk stratification, and consequently, management decisions and patient care. We believe that it is worth ‘taking the trouble’ to apply these risk scores, which can effectively supplement clinical judgment.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.