A large randomized double-blinded study compares a new inhalational steroid agent to fluticasone in patients with moderate persistent asthma. The following results are highlighted in the study: The investigators conclude that the study drug is more effective in this patient population. Which of the following is the major limitation of this conclusion?
A)Clinical significance
B)Lack of power
C)Lead-time bias
D)Surrogate endpoint
E)Type 1 error

Question

A large randomized double-blinded study compares a new inhalational steroid agent to fluticasone in patients with moderate persistent asthma.  The following results are highlighted in the study: The investigators conclude that the study drug is more effective in this patient population.  Which of the following is the major limitation of this conclusion?
A)Clinical significance
B)Lack of power
C)Lead-time bias
D)Surrogate endpoint
E)Type 1 error

Quizplus · Accepted Answer

Explanation: Patients taking the new inhalational steroid agent used rescue doses of albuterol 23.6 times per month, while patients taking fluticasone had 24.1 uses per month. This change was statistically significant (P <0.05), but the difference of 0.5 (= 24.1 − 23.6) uses per month is so small that it lacks clinical significance. Another example would be if a new antihypertensive medication showed a statistically significant 0.5 mm Hg decrease in blood pressure compared to an existing antihypertensive drug. The practical clinical value of this result would be limited. There are no preset guidelines regarding what constitutes clinical significance; much will depend on the specific scenario and disease process. (Choice B) Power is the ability to detect an effect if that effect exists. A larger sample size increases power. This study is reported to be large so lack of power is unlikely. (Choice C) Lead-time bias occurs when several disease interventions are compared and one of the interventions diagnoses the disease earlier than the others. This leads to the erroneous presumption that there is an effect on the outcome (eg, the intervention prolongs survival outcome). There is no indication of lead-time bias in this study. (Choice D) Surrogate endpoints (eg, laboratory test, biomarker, physical finding) are meant to substitute for clinically meaningful endpoints and are typically validated and used in multiple studies. Examples include CD4 cell count (surrogate for HIV complications) and hemoglobin A_1c levels (surrogate for diabetes complications). Rescue inhaler use is commonly used as a surrogate for asthma control; the lack of clinical significance of the results seen is a much more prominent limitation of the study. _1c (Choice E) A type 1 error (α) describes the probability of seeing a difference when there is no difference in reality. In this study, it would mean that the statistically significant difference seen between fluticasone and the new agent is not actually present. However, even if that difference were statistically significant, it would not be clinically meaningful. Educational objective: A statistically significant result may not be clinically significant (not meaningful in practice).

A Large Randomized Double-Blinded Study Compares a New Inhalational Steroid