PMean: Slapping the word pilot on a failed study

Steve Simon

2019-03-03

Someone was asking on the MedStats listserv about a study that had gone off the rails. They had recruited only about a third of the patients that they had wanted. Things were going pretty well in the first arm of the study

Anyway

  1. Analyze the study as originally planned
  1. Call this a pilot study and provide descriptive analyses only.
  2. Recognize that the data is so fatally flawed that any analysis of the data would be inappropriate.

This is what I suggested.

If you stop a study early because of reasons that are related to the p-value that you are seeing

If you stop a study for other reasons

Stopping because of a very high dropout rate could be a problem

If you are a stickler for accuracy

So go ahead and analyze the data.

But please

There are too many people who can’t afford to run a good study and know that it is way under-powered