This paper provides an overview of intent-to-treat analysis versus as-treated analysis. The first step in evaluating a therapy must be an unbiased intent-to-treat analysis. This is the only sure way to protect against potential bias. As-treated analyses, however, can be used in some situations, and methodology to provide objective secondary analyses should be pursued. A two-arm clinical trial of an active treatment versus placebo control will be used to illustrate the two procedures.
FarwellJJr.LeeYJHirtzDGSulzbacherSIEllenbergJHNelsonKB. Phenobarbital for febrile seizures: Effects on intelligence and on seizure recurrence. New Engl J Med.1990;322:364–369.
2.
HirtzDGLeeYJEllenbergJHNelsonKB. Survey on the management of febrile seizures. Am J Dis Child.1986;140:909–914.
3.
NelsonKBEllenbergJH. Predictors of epilepsy in children who have experienced febrile seizures. New Engl J Med.1976;295:1029–1033.
4.
KaplanELMeierP. Nonparametric estimation from incomplete observations. J Am Stat Assoc.1965;52:202–223.
5.
WolfSMCarrADavisDC. The value of phenobarbital in the child who has had a single febrile seizure: A controlled prospective study. Pediatrics.1977;59:378–385.
6.
PilgaardSHansenFJPaerregaardP. Prophylaxis against febrile convulsions with phenobarbital: A three-year prospective investigation. Actz Paediatr Scand.1981;70:67–71.
7.
EfronBFeldmanD. Compliance as an explanatory variable in clinical trials. J Am Stat Assoc.1991;86:9–26.
8.
PeduzziPWittesJDetreKHolfordT. Analysis as-randomized and the problem of non-adherence: An example from the veterans affairs randomized trial of coronary artery bypass surgery. Stat Med.1993;12:1185–1195.
