A 62-year-old man with diabetes, hypertension, and hyperlipidemia comes to the emergency department of an academic medical center with chest pain, nausea, vomiting, and diaphoresis. An electrocardiogram demonstrates ST elevation in the anterior leads, and cardiac enzymes are markedly elevated. Investigators at the center are designing a randomized control trial to test the hypothesis that drug B will decrease the mortality associated with acute ST-elevation myocardial infarction compared to standard of care. To ensure that investigators will not miss a difference between drug B and standard of care (if a difference truly exists), which of the following would they want to maximize?

Question

A 62-year-old man with diabetes, hypertension, and hyperlipidemia comes to the emergency department of an academic medical center with chest pain, nausea, vomiting, and diaphoresis.  An electrocardiogram demonstrates ST elevation in the anterior leads, and cardiac enzymes are markedly elevated.  Investigators at the center are designing a randomized control trial to test the hypothesis that drug B will decrease the mortality associated with acute ST-elevation myocardial infarction compared to standard of care.  To ensure that investigators will not miss a difference between drug B and standard of care (if a difference truly exists), which of the following would they want to maximize?

Quizplus · Accepted Answer

Statistical power, (1 - β), represents a study's ability to detect a difference when one exists. It is the probability of rejecting the null hypothesis when it is truly false - ie, the probability of finding a true relationship. Power depends on sample size and the difference in outcome between the groups being tested. In this study, the researchers want to detect a difference between drug B and standard of care if one exists; they want to maximize power. β is the probability of committing a type II error (Choices B and D). Type II error occurs when researchers fail to reject the null hypothesis when it is truly false. This causes investigators to miss true relationships. An example of a type II error would be a study finding that aspirin does not affect platelet function when, in fact, it does. Therefore, if β is set at 0.2, the power will be (1 - β) = 80%; there will be an 80% chance of rejecting the null hypothesis when it is truly false. (Choices A and C) Type I error occurs when researchers reject the null hypothesis when the null hypothesis is really true. That is, a study finds a statistically significant difference between 2 groups when one does not truly exist. An example would be if a study concluded that candy improves heart failure mortality when, in fact, it does not. α is the maximum probability of making a type I error that a researcher is willing to accept. Generally, α is compared to the p-value, the probability of observing a given result (or more extreme) due to chance alone assuming the null hypothesis is true (eg, if there is no real difference between the groups). The value of α is typically set at 0.05, meaning that researchers are willing to accept up to a 5% chance of making a type I error. In such a scenario, if p <0.05, the result is said to be statistically significant. Educational objective: Power (1 − β) is the probability of rejecting a null hypothesis when it is truly false. It is typically set at 80% and depends on sample size and difference between outcomes.

A 62-Year-Old Man with Diabetes, Hypertension, and Hyperlipidemia Comes to the Emergency