Efficacy versus Effectiveness

Efficacy

Efficacy refers to whether an intervention works under ideal conditions for a specific outcome. Regulatory agencies such as the U.S. Food and Drug Administration require that medications demonstrate efficacy prior to their approval for commercialization. These premarketing studies are referred to as drug registration trials and generally [Page 432]aim to show superiority of the proposed agent versus placebo. This superiority is measured using a very specific outcome, such as reduction of symptoms using a rating scale designed and validated for that purpose, or a reduction in a laboratory measure, such as decrease in blood cholesterol levels. These clinical trials can be very large, enrolling multiple hundreds of patients across many study centers in several countries. Attempts are usually made to ensure a homogeneous test population. Intervention choice is randomized and subjects are followed double-blind. These clinical trials also monitor for adverse events, usually relying on spontaneous reporting but also including safety scales when certain tolerability problems are anticipated, such as extrapyramidal symptoms encountered with the use of antipsychotics. Clinical registration trial reports include information on both efficacy and tolerability under these artificial study conditions, but the aim of these reports is not to provide a synthesis for clinical guidance but to prove that the intervention is efficacious. Whether or not the intervention is efficacious and effective in a routine clinical practice is not certain. This is especially problematic when the patients who receive the intervention in clinical practice are unlike the subjects who received the intervention under controlled conditions. A good example of this are the registration trials of intramuscular antipsychotics for the treatment of agitation associated with schizophrenia or bipolar mania. Patients in these trials were required to provide informed consent and may represent a population that is very different from the agitated patient involuntarily brought to an emergency department by the police in terms of level of cooperation, degree of agitation, comorbid medical conditions, and presence of active alcohol or drug use. Perhaps the biggest objection to the use of registration trial data is that the comparator of placebo is not appropriate for the clinician whose main interest is to know how the new intervention compares with the old established one.

...