wolfpacker wrote:LastHope wrote:wolfpacker wrote:Note that the 100% number at 5 implants was only for 2 patients. Not exactly statistically significant.
The P value is 0.01. Isn't statistical significance satisfied if < 0.05?
"The risk of specific device infection was strongly correlated and increased based on number of prior IPPs: 1st (6.8%; 3/44), 2nd (18.2%; 4/22), 3rd (33.3%; 4/12), 4th (50%; 4/8), and 5th (100%; 2/2) (R2 = 0.90, p = 0.01). Similarly, overall rates of infection positively correlated with number of prior IPP-related surgeries performed (R2 = 0.97, p < 0.01)."
https://www.nature.com/articles/s41443-018-0026-6
Their overall conclusion that "The risk of specific device infection was strongly correlated and increased based on number of prior IPPs" has p=0.01, which definitely makes sense. But this dataset doesn't really make it clear that there's a hard 100% limit at 5 revisions.
Well said.