Clinically relevant difference

For continuous data, the sample size is inversely proportional to the square of the clinically relevant difference. So if the crd is reduced by a factor of two then the sample size is increased by a factor of four, if the crd is increased by a factor of two then the sample size is reduced by a factor of four. In our earlier example the sample size requirement to detect a difference of 8 mmHg was 33 

For binary data this same relationship between the crd, in terms of the absolute difference in success rates, and the sample size is only approximately true. In the example we were looking to detect an improvement in the success rate from 35 per cent to 50 per cent, an absolute difference of 15 per cent and we needed a sample size of 227 patients per group. If we were to halve that difference and look for an improvement from 35 per cent to 42.5 per cent then the sample size requirement would be 885 per group, an increase in the sample size by a factor of 3.9. 

---

The optimal myeloablative preparative regimen is challenging to study because several indications for HCT (e.g., SCID and thalassemia) are rare enough that it is not feasible or is cost-prohibitive to conduct clinical trials that are powered adequately to detect clinically relevant differences. The longterm outcomes of busulfan-cyclophosphamide (BU-CY) and... 

We see that in contrast to the type-1 error, the type-11 error is defined as occurring when accepting the null hypothesis if it is false. The power of a test is defined to be the probability of detecting a true difference and is equal to 1 — probability (type-11 error). The type-11 error and power depend upon the type-1 error, the sample size, the clinically relevant difference (CRD) that we are interested in detecting and the expected variability. Where do these values come from ... 

If the clinically relevant difference is large the sample size will be small. 

To illustrate the use of the formula suppose we are designing a trial to compare treatments for the reduction of blood pressure. We determine that a clinically relevant difference is 5 mmHg and that the between-patient standard deviation 0 is 10 mmHg. At)q)e-1 error is set at 0.05 and the type-11 error at 0.20. Then the required sample size, per group, is... 

In a placebo-controlled hypertension trial, the primary endpoint is the fall in diastolic blood pressure. It is required to detect a clinically relevant difference of 8 mmHg in a 5 per cent level test. Fiistorical data suggests that CT= 10 mmHg. Table 8.4 provides sample sizes for various levels of power and differences around 8 mmHg the sample sizes are per group. 

The sample size calculation should be detailed in the trial publication, indicating the estimated outcomes in each of the treatment groups (and this will define, in particular, the clinically relevant difference to be detected), the type I error, the type II error or power and, for a continuous primary outcome variable in a parallel group trial, the within-group standard deviation of that measure. For time-to-event data details on clinically relevant difference would usually be specified in terms of the either the median event times or the proportions event-free at a certain time point. 

Risk-based decision criteria would then have to relate to clinical relevance different levels of understanding (e.g., correlative, causal, mechanistic) will need to be recognized within this context. This general approach is utilized in some current regulatory policies in the desired state the approach can be extended to other areas. For example in current regulatory policies ... 

When recombinant and urinary versions of follicle-stimulating hormone were compared under double-blind conditions in an in vitro fertilization program in a randomized, multicenter study (12), the former was more potent. There were no clinically relevant differences in safety between the two products and no cases of ovarian hyperstimulation syndrome. 

In this example of a group difference of 3 mmHg, development would probably not continue. However, at what point would a decision to continue likely be made This leads to another question What is the smallest effect size that is clinically meaningful, or clinically relevant This effect size can be called the clinically relevant difference (CRD). Its determination is a clinical one, not a statistical one. This determination may well be strongly influenced by existing empirical evidence (for example, actuarial statistics), but, unlike statistical significance, its determination is not simply formulaic. 

The clinically relevant difference (CRD) that the test is required to detect. This is the treatment effect size, i.e., the difference between the mean drug treatment group response and the mean placebo treatment group response, that the sponsor deems clinically relevant. 

Determining the clinically relevant difference to look for in the study is relatively straightforward. Another way to conceptualize the clinically relevant difference is as the smallest effect size that is clinically meaningful. This can be based on clinical input. For example, a decrease in SBP of 10 mmHg may be thought by the sponsor to be clinically relevant in this context. 

Drug response is likely to be the result of a complex function of the influence of many genes interacting with environmental and behavioral factors. Whether PK-PD variability translates into clinically relevant differences in drug response depends on further issues including compliance, the availability of alternate drugs and doctor/patient perception of side-effects [10]. 

The criterion of bioequivalence applies if there is a similarity in bioavailability (statistically proven) that is unlikely to result in clinically relevant differences in efficacy and/or safety. 

Liquid release. Dose adjustment may be needed. Benefit is less fluctuation of levels. Retard is given as twice daily dosing (rather than three times day for standard). Chewable tablets reach peak plasma concentration more slowly than syrup (6 hours vs. 2 hours) Suitable for long-term use. No clinically relevant difference between oral dose forms. No dose adjustments between oral formulations... 

EM), and ultrarapid (UM) metabolizer genotypes. The figure shows that the amount of dose adaptation varies among substrates, and that clinically relevant differences in dosages exist for some antidepressants, whereas others are only slightly influenced by the CYP2D6 polymorphism. 

An understanding of the pharmacogenetics of pharmacodynamics is probably less advanced than that of pharmacokinetics, but inherent variability in pharmacodynamics may be greater than in pharmacokinetics. In turn, whether pharmacokinetic-pharmacodynamic variability translates into clinically relevant differences in drug response depends on further clinical and operational issues, such as compliance, and doctor/patient... 

There are a number of values of the treatment effect (delta or A) that could lead to rejection of the null hypothesis of no difference between the two means. For purposes of estimating a sample size the power of the study (that is, the probability that the null hypothesis of no difference is rejected given that the alternate hypothesis is true) is calculated for a specific value of A. in the case of a superiority trial, this specific value represents the minimally clinically relevant difference between groups that, if found to be plausible on the basis of the sample data through construction of a confidence interval, would be viewed as evidence of a definitive and clinically important treatment effect. 

PpLACEBO value that is considered the minimally clinically relevant difference (CRD). 

The definition of the minimally clinically relevant difference of interest involves clinical, medical, and regulatory experience and judgments. The appropriate sample size formula depends on the test of interest and should take into account the need for multiple comparisons (either among treatments or with respect to multiple examinations of the data). The project statistician provides critical guidance in this area. 

This issue is difficult to address in generality because the answer is dependent on correlation between measures and also on clinically relevant differences for different measures and all sorts of correlation structures and patterns for differences can be envisaged. However, some qualitative feel for the effect can be gained by making two radically simplifying assumptions first that the values are conditionally independent given some latent variable, and second that some common effect size (ratio of clinically relevant difference to standard deviation) is targeted. (It will also, of course, be assumed that the outcomes are Normally distributed.)... 

It is desired to run a placebo-controlled parallel group trial in asthma. The target variable is forced expiratory volume in one second (FEVi). The clinically relevant difference is presumed to be 200 ml and the standard deviation 450 ml. A two-sided significance level of 0.05 (or 5%) is to be used and the power should be 0.8 (or 80%). What should the sample size be. ... 

The third complication is that there is usually no agreed standard for a clinically relevant difference. In practice some compromise is usually reached between true clinical requirements and practical sample size requirements. (See below for a more detailed discussion of this point.)... 

It may be a requirement that the results be robust to a number of alternative analyses. The problem that this raises is frequently ignored. However, where this requirement applies, unless the sample size is increased to take account of it, the power will be reduced. (If power, in this context, is taken to be the probability that all required tests will be significant if the clinically relevant difference applies.) This issue is discussed in section 13.2.12 below. 

Figure 13.1 Power as a function of clinically relevant difference for a two-paraUel-group trial in asthma. The outcome variable is FEVi, the standard deviation is assumed to be 450 ml, and n is the number of patients per group. If the clinically relevant difference is 200 ml, 80 patients per group are needed for 80% power.

In fact, (13.4) would imply that we knew, before conducting the trial, that the treatment effect is either zero or at least equal to the clinically relevant difference. But where we are unsure whether a drug works or not, it would be ludicrous to maintain that it cannot have an effect which, while greater than nothing, is less than the clinically relevant difference. 

---