Cardiovascular event risk assessment — Fusion of individual risk assessment tools applied to the Portuguese population

Cardiovascular disease (CVD) is the world's largest killer, responsible for 17.1 million deaths per year. Thus, the improvement of the prognosis of this disease is an important factor to defeat the current statistics. Although there are several risk tools available to assess the risk of occurrence of a cardiovascular event within a given period of time, these tools present some major drawbacks. In particular, each individual tool considers a reduced number of risk factors, does not permit to incorporate additional clinical knowledge and presents difficulties in coping with missing risk factors. In order to overcome the identified weaknesses, a flexible framework is proposed here, based on the fusion of a set of distinct risk assessment tools. The methodology is based on two main hypotheses: i) it is possible to derive a common representation for the individual risk assessment tools, ii) it is possible to combine (fusion) the obtained individual models, in order to achieve the referred goals. Additionally, through the implementation of optimization techniques, an increasing in the global risk prediction performance is also investigated. The validation of the strategy is carried out through the combination of three current risk assessment tools (GRACE, TIMI, PURSUIT) developed to predict the risk of an event in coronary artery disease (CAD) patients. The combination of these tools is validated with two real patients testing datasets: i) Santa Cruz Hospital, Lisbon/Portugal, N=460 ACS-NSTEMI1 patients; ii) Santo André Hospital, Leiria/Portugal, N=99 ACS-NSTEMI patients. Considering the obtained results with the available datasets it is possible to state that the initial goals of this work were achieved. This evidence makes this work a valid contribution for the improvement of the risk assessment applied to cardiovascular diseases.

[1]  Peter Herbison,et al.  Global Registry of Acute Coronary Events (GRACE) hospital discharge risk score accurately predicts long-term mortality post acute coronary syndrome. , 2007, American heart journal.

[2]  Nir Friedman,et al.  Bayesian Network Classifiers , 1997, Machine Learning.

[3]  R. Jackson.  Guidelines on preventing cardiovascular disease in clinical practice , 2000, BMJ : British Medical Journal.

[4]  Kevin B. Korb,et al.  Data mining cardiovascular Bayesian networks , 2005 .

[5]  Sotiris B. Kotsiantis,et al.  Supervised Machine Learning: A Review of Classification Techniques , 2007, Informatica.

[6]  Geoffrey I. Webb,et al.  Discretization for naive-Bayes learning: managing discretization bias and variance , 2008, Machine Learning.

[7]  Alexey Tsymbal,et al.  Ensemble feature selection with the simple Bayesian classification , 2003, Inf. Fusion.

[8]  E. Antman,et al.  The TIMI risk score for unstable angina/non–ST-elevation MI: a method for prognostication and therapeutic decision making , 2001 .

[9]  Geoffrey I. Webb,et al.  A comparative study of Semi-naive Bayes methods in classification learning , 2005 .

[10]  E W Steyerberg,et al.  Predictors of outcome in patients with acute coronary syndromes without persistent ST-segment elevation. Results from an international trial of 9461 patients. The PURSUIT Investigators. , 2000, Circulation.

[11]  G. Specchia,et al.  Management of acute coronary syndromes in patients presenting without persistent ST-segment elevation. , 2002, European heart journal.

[12]  Paulo Carvalho,et al.  Long term cardiovascular risk models' combination , 2011, Comput. Methods Programs Biomed..

[13]  Carlos Aguiar,et al.  TIMI, PURSUIT, and GRACE risk scores: sustained prognostic value and interaction with revascularization in NSTE-ACS. , 2005, European heart journal.

[14]  E. Antman,et al.  The TIMI risk score for unstable angina/non-ST elevation MI: A method for prognostication and therapeutic decision making. , 2000, JAMA.