https://users.phhp.ufl.edu/rbauer/EBPP/intention_to_treat_stats.pdf

Statistical considerations in the intent-to-treat principle.

The combination of randomized and historical controls in clinical trials.

The multicenter trial of cryotherapy for retinopathy of prematurity (ROP), whose preliminary results are reported in this issue of theArchives, 1 was designed to resolve the uncertainty about the value of treatment and to evaluate its drawbacks and possible benefits. Preliminary results weigh heavily in favor of treatment, and thus constitute an important contribution to the management of ROP, a disorder with the potential to blind many hundreds of premature infants each year. 2 See also pp 457 and 471. By comparing paired sample data from infants with symmetric disease and independent sample data from infants with asymmetric disease in 172 premature infants at 23 centers, the study confirmed that cryotherapy conferred a clear benefit at three months after treatment. The benefit represents an approximate 50% reduction in unfavorable events defined as the presence of a retinal fold, retinal detachment, or retrolental tissue (21.8% of the treated eyes vs

Multicenter trial of cryotherapy for retinopathy of prematurity.

We assessed the efficacy of adding nifedipine to the conventional treatment of unstable angina in 138 patients in a prospective, double-blind, randomized, placebo-controlled trial. There was no difference between the two groups in the dose of conventional antianginal medication or in age, prior myocardial infarction, ejection fraction, or other risk factors. Failure of medical treatment (defined as sudden death, myocardial infarction, or bypass surgery within four months) occurred In 43 of 70 patients given placebo and in 30 of 68 given nifedipine. KapIan–Meier survival-curve analysis of the number and time dependence of treatment failures demonstrated a benefit of nifedipine over placebo (P = 0.03). The benefit was particularly marked in patients with ST-segment elevation during angina (P = 0.02). Side effects (transient hypotension or diarrhea) required with-drawal of the drug from four patients given nifedipine and from one given placebo. We conclude that the addition of nifedipine to conventi...

Nifedipine in Unstable Angina: A Double-Blind, Randomized Trial

An Overview of Intravesical Therapy for Superficial Bladder Tumors

In 1952, Marc Daniels and Austin Bradford Hill (now Sir Austin) reported that streptomycin combined with para-aminosalicylic acid was more effective than bed rest, or the use of either drug alone, in the therapy of pulmonary tuberculosis in young adults.1 The patients participating in this trial, after they had been accepted on the basis of well defined criteria, were "randomly allocated to one of the treatment groups." So, according to A. L. Cochrane,2 the randomized clinical trial was launched and revealed as a practical means of applying the experimental method to clinical medicine. In spite of its availability for 20 . . .

F. J. Ingelfinger

Papers

The randomized clinical trial.