Table 3:
Diagnostic accuracy of each test for ulnar and radial arteries
| Parameter | Examined using the modified Allen test | Examined using the heart rate–monitoring application | p value† |
|---|---|---|---|
| Ulnar artery assessment, n = 736 | |||
| No. (%) of positive test results* | 126 (17.1) | 49 (6.7) | < 0.001 |
| Overall diagnostic accuracy, % (95% CI) | 84.0 (79.1 to 87.9) | 92.1 (88.7 to 94.6) | 0.002 |
| Specificity, % (95% CI) | 84.6 (79.6 to 88.5) | 93.3 (90.2 to 95.5) | < 0.001 |
| Sensitivity, % (95% CI) | 40.0 (14.3 to 72.7) | 37.5 (12.5 to 71.5) | 0.9 |
| Positive likelihood ratio (95% CI) | 2.6 (−0.3 to 5.4) | 5.6 (0.1 to 11.1) | |
| Negative likelihood ratio (95% CI) | 0.7 (0.2 to 1.2) | 0.7 (0.3 to 1.0) | |
| Radial artery assessment, n = 736 | |||
| No. (%) of positive test results* | 102 (13.9) | 30 (4.1) | < 0.001 |
| Overall diagnostic accuracy, % (95% CI) | 84.0 (76.6 to 87.6) | 95.9 (93.4 to 97.5) | < 0.001 |
| Specificity, % (95% CI) | 85.6 (81.1 to 89.1) | 96.9 (94.6 to 98.3) | < 0.001 |
| Sensitivity, % (95% CI) | 46.7 (25.4 to 69.2) | 50.0 (23.1 to 76.9) | 0.9 |
| Positive likelihood ratio (95% CI) | 3.2 (1.3 to 5.2) | 16.4 (1.5 to 31.2) | |
| Negative likelihood ratio (95% CI) | 0.6 (0.3 to 0.9) | 0.5 (0.2 to 0.9) | |
| All arteries assessed, n = 1472 | |||
| No. (%) of positive test results* | 228 (15.5) | 79 (5.4) | < 0.001 |
| Overall diagnostic accuracy, % (95% CI) | 84.0 (79.9 to 87.4) | 94.0 (91.9 to 95.6) | < 0.001 |
| Specificity, % (95% CI) | 85.1 (80.9 to 88.5) | 95.1 (93.2 to 96.6) | < 0.001 |
| Sensitivity, % (95% CI) | 45.0 (26.4 to 65.2) | 43.8 (23.9 to 65.9) | 0.9 |
| Positive likelihood ratio (95% CI) | 3.0 (1.5 to 4.6) | 9.0 (3.2 to 17.8) | |
| Negative likelihood ratio (95% CI) | 0.6 (0.4 to 0.9) | 0.6 (0.3 to 0.8) | |
Note: CI = confidence interval.
*
A positive test result indicates that there is inadequate collateral blood flow; this includes all true-positive and false-positive results.
†
Values of p were generated using a generalized estimating equation model for the randomization arm parameter (i.e., modified Allen test versus heart rate–monitoring application) with binomial distribution using a logit link where the outcome was a correct test result.