Methodology for Small Clinical Trials

Abstract

Small clinical trials are trials in which the number of patients does not enable the objective of the study to be appropriately met with the usual methodological rules. This situation is common in the case of rare diseases, in paediatrics, in certain cancer pathologies or when the number of patients exposed to the treatment needs to be limited. The principal methodological problems are initially identified, and the classical methods (controlled, randomised, double-blind trial using parallel groups, crossover trial, factorial design, trial performed with several measures repeated over time, add-on design, randomised withdrawal design or early-escape design) and more uncommon methods (sequential approaches, meta-analyses, the 'N of 1' method and other methods that facilitate decision making or modelling) are then discussed. Subsequently, recommendations are made to ensure that the results obtained are not a matter of chance, and to increase the level of proof. Keywords: small trials, statistical methods, level of proof Small clinical trials are trials conducted and analysed using patient numbers lower than is required to meet the aims of the study according to the more common methodological rules. This paper therefore does not concern the standard phase I and II trials where the number of patients included in the study, although quantitatively small, is methodologically adapted to the aims of the study. It concerns all trials that aim to establish the proof of whether any given effect, be it of efficacy or tolerance, is present. PHARMACOLOGIE CLINIQUEThis situation is often the case with rare diseases, in paediatrics or in cancer treatment, when the external limitations of patient availability do not enable the recruitment of patient numbers that meet the usual assessment criteria. It is also the case when the number of patients exposed to the investigational treatment methods needs to be limited, either because of constraints regarding the particular protection of individuals (children, pregnant women), or because of the toxicity of the product (cancer treatment), or even its lack of efficacy (placebo, weak doses). The methodological issue in 'proof of concept' studies, the results of which lead to the decision on whether or not to continue developing a molecule, although based on a different context, can be related to this field.The main issue surrounding these small trials is the level of proof provided by the results of these studies, which reflects the relevance of using non-standard methods.Conducting a small trial through necessity reveals various levels of difficulties, some of which are not entirely specific to small trials, but are increased because of the reduced patient numbers. These include the following:1. The risk of not reaching a conclusion, even though there is a real difference between the therapeutic regimens being compared, because of the great variability in random fluctuations. Only a quantitatively important effect is likely to be given prominence.2. The caution necessary when extr...

Using a meta-narrative literature review and focus groups with key stakeholders to identify perceived challenges and solutions for generating robust evidence on the effectiveness of treatments for rare diseases

“…Strategies that have been proposed for enhancing statistical efficiency at the design phase for clinical evaluative studies of rare disease treatments include factorial trials and adaptive designs. Factorial trials are designed to test multiple treatments simultaneously using the same study population, thus reducing the overall number of participants needed [ 2 , 33 , 39 , 40 , 46 , 49 , 53 , 57 ]. For example, in a 2 × 2 factorial design participants are randomized to either treatment A or control group A, and then randomized again to treatment B or control group B, which effectively reduces the sample size needed to test these two treatments by 50% because the same participants are being randomized [ 40 ].…”

A novel placebo-controlled clinical study design without ethical concerns – The free choice paradigm

“…This free-choice paradigm (FCP) may be regarded as a modification of the ''adaptive response design'' [25], the ''early-escape design'' [26] and other adaptive strategies [27], but its essential difference is that patients have the full selection (step-up, stepdown) with every drug intake. It may offer an alternative approach to common drug test procedures, though its statistics have still to be established.…”

Applying Evidence Standards to Rehabilitation Research

“…Such studies can in fact provide much stronger evidence than simple pre-post research designs or other studies lacking a comparison group, provided certain criteria are met. 7,8,23,24 With interrupted time series, key criteria include evidence (not assumption) of baseline and continuing stability of the trend line and rapid effect onset. 7,8,25 (Crossover and ABA designs are particularly strong and powerful for smaller samples if combined with randomization).…”

“…7,8,25 (Crossover and ABA designs are particularly strong and powerful for smaller samples if combined with randomization). 9,24 Individual baseline and n of 1 studies are not incorporated into standard evidence grading or are assigned to low levels (grade 3 or 4), regardless of design or context. But is this always so?…”