what do you think......
Message posted by mike on January 25, 2000 at 12:00 AM (ET)

In a retrospective study of 81 patients, I found an association between the preoperative hemoglobin and the postoperative INR response. Patients are given a blood thinner after surgery and the level of blood thinning is measured in units called the INR. Ability to preoperatively determine sensitive patients to blood thinners would be beneficial.
Logistic regression shows that preoperative hemoglobin has a significant effect on postoperative INR. categorical analysis of preop hemoglobin and postoperative response shows Fisher Exact = 0.04,sensitivity =28%,specificity = 30%, FP=70%
FN=70%, NPV=81%,PPV=4%
Do you agree that these findings show statistical significance BUT not clinical significance?