Clinically important difference

In comparing the results from trials 1 and 2 it is clear that the p-value does not tell the whole story. In terms of p-values they are indistinguishable but the first trial has demonstrated a clinically important difference while trial 2 has detected something that is clinically irrelevant. 

It is all too common to see a conclusion that treatments are the same (or similar) simply on the back of a large p-value this is not necessarily the correct conclusion. Presentation of the 95 per cent confidence interval will provide a statement about the possible magnitude of the treatment difference. This can be inspected and only then can a conclusion of similarity be made if this interval is seen to exclude clinically important differences. We will return to a more formal approach to this in Chapter 12 where we will discuss equivalence and non-inferiority. 

Competing perspectives create tension, for example pharmacoeconomics versus clinical importance. Differences in perspective may be irreconcilable because they relate to a perceived encroachment turf wars can erupt between clinical, marketing and pharmacoeconomics departments within the same company, in spite of all three professing the same goal, that is to successfully market a worthwhile drug in a proper fashion. 

A major advantage of quantitative data is that the study size required to determine whether or not a clinically important difference exists between two treatments will usually be much less than that needed with qualitative data. The calculation of study size with ordinal data is more problematic, but somewhat more data will usually be needed than with quantitative data but often considerably fewer than with qualitative data. 

If a study has been designed to have a power of 95% of detecting a specified clinically important difference (CID), the predicted 95% CI for the difference between the treatments that will be obtained when the study is completed is the observed difference between the treatments 0-544 x CID. In the example described in Section 7.4.3.3.1, the CID was 30% (pi being 30% and p2 being 60%), so the predicted 95% CI is the observed difference between the treatments 0.544 X 30% (= 16.3%). Suppose, for example, that the study data gave a 20% difference between the two treatments, the predicted 95% CI for the true difference would be 20% 16-3%, or 3-7% to 36-3%. 

The colloids, in particular albumin, are expensive solutions. Therefore, it is difficult to justify the additional cost of colloidal products unless the benefit-to-risk ratio is substantially greater than that associated with inexpensive crystalloid solutions. This does not appear to be the case based on randomized, controlled studies and meta-analyses comparing colloid and crystalloid solutions for acute circulatory insufficiency. Because other colloids, such as hetastarch, almost always have been compared with albumin and not with crystalloid solutions in published clinical studies (with no clinically important differences being found), there is no reason to suspect that these other colloids have any unique advantages as volume expanders. Adverse effects associated with colloids appear to be uncommon and generally are extensions of their pharmacologic activity (Table 24—4), but this is also true of crystalloids. The benefit-to-risk ratio appears to be similar for colloids and crystalloids thus, based on cost, crystalloids are preferred for initial treatment of circulatory insufficiency. 

There is little evidence to support clinically important differences with regard to the frequency of ulcers and upper GI complications among most available nonaspirin, nonselective NSAIDs (see Table 33-3) when used in equipotent anti-inflammatory dosages. However, the nonacetylated salicylates (e.g., salsalate) and newer NSAIDs (e.g., etodolac, nabumetone, and meloxicam) may be associated with a decreased incidence of GI toxicity. NSAIDs that selectively inhibit cyclooxygenase-2 (COX-2) decrease the incidence of gastroduodenal ulcers and related GI complications when compared to the nonselective NSAIDs. The use of buffered or enteric-coated aspirin confers no added protection from ulcer or GI complications. ... 

The geriatric population is arbitrarily defined as comprising subjects 65 years or older. The older the population likely to use the drug, the more important it is to include the older age range, 75 years and older. For drugs used to treat diseases not unique to, but present in, the elderly, a minimum of 100 subjects usually would allow detection of clinically important differences between the elderly and younger subjects with respect to efficacy as well as adverse reactions. 

In 1994, the benefit of fibrinolytic therapy in the treatment of the elderly was brought into question by the Fibrinolytic Therapy Trialists (FTT) Collaborative Group (3). This meta-analysis included nine fibrinolytic trials incorporating over 58,600 patients. Of these patients, 17,000 were 65-74 years of age and 5754 were >75 years of age. This analysis revealed that the benefit of fibrinolytic therapy diminished with advancing age relative risk reduction (RRR) of death associated with fibrinolytic therapy was 22% in patients <55 years of age, 19% in patients 55-64 years of age, 16% in patients 65-74 years of age and only 4% in patients >75 years of age. However, while the relative effectiveness of fibrinolytic therapy diminished with advancing age, the 4% RRR translated into a 1% ARR, or NNT = 100, still a clinically important difference and similar to the 1.2% ARR in patients <55 years of age. Therefore, even small relative reductions in risk can be important when applied to groups with increased risk. 

Beaton DE, Boers M, Wells GA (2002) Many faces of the minimal clinically important difference (MCID) A literature review and directions for future research. Curr Opin Rheumatol 14 109-114... 

Jaeschkle R, Singer J, Guyatt GH (1989) Measurements of health status Ascertaining the minimally clinically important difference. Control Clin Trials 10 407 15 Jones PW (2002) Interpreting thresholds for a clinically significant change in health status in asthma and COPD, Eur Respir J 19 398 04... 

Frid, A., and Linde, B., 1993, Clinically important differences in insulin absorption from abdomen in IDDM, D/flhetes Res. Clin. Pract. 21 137-141. 

A 12-month open study in primary care showed no clinically important differences in adverse events between flexible and fixed dosing of budesonide - - formoterol. Maintenance and reliever therapy was associated with a significantly lower daily dose of budesonide and direct cost savings, with at least equivalent efficacy [63 ] there were similar findings in other studies [64. ... 

Liver In a study of the potential hepato-toxic or liver tumor-inducing effects of long-term danazol prophylaxis in 92 patients with hereditary angioedema, half of whom took danazol [89 ]. There were no clinically important differences between the liver function parameters in years 0 and 5 in the two groups. Abdominal ultrasound did not detect neoplastic or other potentially treatment-related alterations in the liver parenchyma. 

---