Clinical Trials as Nomological Machines: Implications for Evidence-Based Medicine