Allocating Treatments to Patients in Clinical Trials

Every medical man commits that act of treachery, Mr. Blake, in the course of his practice... Every doctor in large practice finds himself every now and then, obliged to deceive his patients. 

The purpose of a clinical trial is to improve our knowledge of treatments in order that future patients may benefit. The patients actually recruited on a clinical trial have not presented for this reason, however. They expect their physician to treat them with the best treatment available, although they may, of course, be prepared to make some sacrifice in the interest of others. (In the case of chronic diseases there is also the possibility that they themselves may be numbered among the future beneficiaries.) On the other hand, to make wise choices as to treatment, one has to know something about the alternatives and this means that to get precise information about the value of a given treatment, one will also have to study other treatments which may turn out to be inferior. Furthermore, because patients vary over time and from centre to centre and because physicians and facilities vary also, in order to deal with these sorts of bias the ideal solution will be to study patients under various alternatives concurrently. There is thus a dilemma facing the physician in a clinical trial how does the physician as scientist satisfy the physician as healer and vice versa  

An important element in such conflicts between healer and scientist is the degree of knowledge possessed by the physician. This will have increased by the end of the trial and, if the results are studied during the course of the trial, will increase to some degree before. It is thus possible that a choice which was not believed to be against patients interests at the beginning of the trial will become unacceptable by or before the planned end. Patients are also far from homogenous and this creates further concerns both 

Statistical Issues in Drug Development, 2nd Edition. Stephen Senn 2007 John Wiley Sons, Ltd. ISBN 978-0-470-01877-4 

The most common methods of allocation in clinical trials involve some form of randomization. Simple randomization, for example, would require that every possible split of a given set of patients into two groups was equally likely. Although for even a moderately sized trial, a great imbalance in numbers assigned to experimental and control group would be unlikely using this approach it is, nevertheless, more usual to randomize subject to the constraint that numbers shall be equal. In fact, trialists often go further than this and use the method of randomized blocks. 

---

Senn SJ (1995) A personal view of some controversies in allocating treatment to patients in clinical trials. Statistics in Medicine 14 2661-2674. [See Comment in Statistics in Medicine 15(1) 113-116 (1996)]. 

Chapter 6 Allocating treatments to patients in clinical trials (Berger, 2005)... 

A difiSculty frequently arises in clinical trials because not all patients allocated to a treatment will complete the designated period of treatment or follow up. Non-completion may be a function of the particular treatment allocated, in that the reason for discontinuation of therapy may be inadequate efficacy or an adverse effect. If non-completing patients are omitted from fhe treatment comparison, bias may be introduced. Simply, if a treatment s failures are omitted from further consideration, the treatment will appear to be better than it really is. Thus, the primary efficacy analysis for the comparison of treatments must include all patients allocated to each treatment, whatever their ultimate fate. This is known as intention-to-treat (ITT) analysis. The implementation of an ITT analysis is not always easy, but usually either an endpoint analysis (in which the final assessment, whenever it was made, is used for fhe treatment comparison) or one of the strategies based on ranking the patients (from the best to the worst in the study) proposed by Gould. ... 

Double dummy technique. A means of achieving blinding (masking) in clinical trials where two active treatments are being compared and when (as is nearly always the case) they cannot be matched. A placebo to each is then employed and patients are allocated to receive either A and placebo to B or B and placebo to A. 

Randomization. A means by which an element of chance is allowed to enter the allocation of patients to treatment. In clinical trials (and for that matter other experiments) several restrictions are usually imposed on the randomization, for example the requirement that equal numbers shall be assigned to each treatment in each centre. The allocation of patients to treatment according to chance as opposed to other methods which use either whim or prejudice. 

There are two major types of weaknesses associated with meta-analysis. Bias and heterogeneity. When planning a clinical trial every effort is made to minimise the impact of bias on the results. For example, clinicians and patients are kept unaware of the treatment being used for each individual patient, so-called double-blind studies and patients are randomly allocated to a treatment. However, in meta-analysis there... 

Numerous clinical trials have also attested to the anxiolytic efficacy of hydroxyzine. Controlled trials have confirmed its efficacy and safety at a fixed dose of 50 mg in GAD (64). In a double-blind, parallel-group, multicenter study in France and Great Britain, a total of 244 patients with GAD were allocated randomly to treatment with hydroxyzine (12.5 mg t.i.d.), buspirone 5 mg morning and midday and 10 mg in the evening, or placebo. The results showed both hydroxyzine and buspirone to be more efficacious than placebo, indicating that hydroxyzine can be a useful treatment for GAD (65). 

Crossover trial designs (Figure 31.2) are less often used in clinical practice and are implemented typically only in Phase 1 or 2 of development. Examples of these trials, such as bioequivalence and drug-drug interaction, commonly appear in the literature. In fact, numerous journal articles and textbooks are devoted to the formulation and analysis of these designs. In these trials, patients are allocated randomly to study arms (or sequences). Each patient within the sequence receives a treatment followed by one or more treatments these treatments can be replicated. The treatments can consist of single or multiple doses. 

Randomization is a word which is used to describe the allocation at random (using aleatory devices, such as, for example, dice, cards, lots or random number tables or generators) of treatments to the units in an experiment. In a clinical trial these units are most typically the patients, although occasionally a centre might be treated as a unit, and in cross-over trials, in which patients may receive two or more treatments, a treatment episode is the basic experimental unit. 

Note, that these proportions are true in the long run. If a clinical trial is stopped after a short while and all future patients are then allocated to the treatment which has proved best so far, with no option to reverse this policy, then the randomized trial... 

---