“…However, for many of these risk-stratification tests, high positive and negative predictive values have not been consistently obtained. A multifactorial model 40 may be scientifically most appropriate to determine the risk of cardiac arrest, but it is unclear whether such a model can be widely applied clinically. Thus, selection of patients for trials examining the efficacy of the ICD in the primary prevention of sudden death has typically been based on ejection fraction and, at most, 1 or 2 other risk-stratification tests.…”