Table 1

 Some pragmatic and intrinsic limitations of randomised controlled clinical trials in rheumatoid arthritis

Pragmatic limitations
    Relatively short observation period in most clinical trials in chronic diseases
    Inclusion and exclusion criteria in clinical trials often restrict eligibility to a small minority of patients who will receive the intervention
    Inflexible dosage schedules and concomitant drug therapies may limit the apparent efficacy of certain drugs such as methotrexate
    Surrogate markers used in clinical trials are reversible measures of inflammatory activity, which may be suboptimal indicators of long term irreversible damage and poor outcomes—for example, tender joints may be poor predictors of joint deformity and work disability
    Statistically significant results are not necessarily clinically important and vice versa
    Rare adverse events cannot be captured in clinical trials of fewer than 10 000 subjects
Intrinsic limitations
    The design of a clinical trial may greatly influence the results, despite inclusion of a control group
    Clinical trial reports generally ignore individual variation in responses to treatments
    Interpretation of results and adverse events is not standardised and may introduce bias into the reporting of the results of any clinical trial
    The format of a clinical trial may distort the “placebo effect” in either direction by informing patients that they are to receive one of several regimens rather than the “best” therapy