Significant Difference: P Doesn't Always Equal .05

Here's another gem where the name of the lead author, Georgetown University's John Marshall MD, will be familiar to many in this important therapeutic area: colorectal cancer. The article in Journal of Oncology Practice does a nice job of telling the story of a typical 3-step PI-CME project. While only 27 people (mostly physicians) participated, outcomes improved significantly, especially with respect to patient safety and supportive care. 

Here's something unusual: The small sample sizes required investigators to choose a non-standard significance value of 10% (P < .10) to detect important trends.