Wednesday, April 11, 2012
The Quality Reports section of Pediatrics was developed to share real-world experience in conducting quality-improvement projects. There are important lessons from such projects that go beyond statistical inference testing. The stories related to rapid tests of change – what was tried and seemed to work or not work – is a cornerstone of Quality Reports. As I have said, quality-improvement is not simply about the p-value. However, statistical testing, when used appropriately, can illuminate the understanding of quality-improvement projects. Smith et al. (doi: 10.1542/peds.2011-2412) present their work regarding the treatment of community-acquired pneumonia and better antimicrobial stewardship. There are many lessons about how to roll out and implement guidelines. However, I was struck by their approach to data analysis which helped to clarify trends over time. This was a clever approach to a complex modeling problem. Such an analysis is not always necessary and lack of this type of formal statistical analysis should not hold back those who want to share their experience in quality improvement from submitting their work. For ideas and inspiration about presentation, look to the SQUIRE guidelines.
Posted by Dr. Lewis R. First at 12:01 AM