Model 1* | Model 2† | Model 3‡ | ||||
Unadjusted ICC (95% CI) | P value§ | Adjusted ICC (95% CI) | P value§ | Adjusted ICC (95% CI) | P value§ | |
Sampling technique for oropharyngeal swab | – | – | – | – | 8.2×10−7 (8.8×10−38 to 1) | 0.497 |
Frequency of swabbing one tonsil only | 2.3×10−5 (1.1×10−24 to 1) | 0.496 | 2.9×10−6 (1.5×10−26 to 1) | 0.495 | – | – |
Frequency of swabbing both tonsils only | 1.2×10−4 (1.2×10−81 to 1) | 0.496 | 5.4×10−6 (1.8×10−19 to 1) | 0.494 | – | – |
Frequency of swabbing oropharynx only | 6.1×10−6 (1.1×10−23 to 1) | 0.495 | 5.4×10−7 (2.1×10−39 to 1) | 0.497 | – | – |
Frequency of swabbing soft palate only | 1.6×10−4 (3.4×10−29 to 1) | 0.495 | 2.3×10−6 (1.1×10−39 to 1) | 0.496 | – | – |
Frequency of swabbing both tonsils together with oropharynx | 1.8×10−4 (1.8×10−26 to 1) | 0.496 | 1.2×10−6 (1.1×10−39 to 1) | 0.497 | – | – |
Frequency of gag reflex | 7.3×10−4 (6.7×10−17 to 1) | 0.474 | 3.7×10−6 (3.9×10−37 to 1) | 0.496 | – | – |
Sampling technique for anorectal swab | – | – | – | – | 1.7×10−7 (4.3×10−40 to 1) | 1.000 |
Distance of swab inserted from the anal verge (cm) | 1.2×10−4 (5.8×10−20 to 1) | 0.498 | 3.9×10−3 (5.5×10−6 to 7.4×10−1) | 0.373 | – | – |
Frequency of patient self-collected swabs | 4.5×10−3 (3.0×10−5 to 4.1×10−1) | 0.334 | 5.4×10−3 (3.5×10−5 to 4.6×10−1) | 0.333 | – | – |
Frequency of clinician-collected swabs without proctoscope | 3.8×10−3 (1.2×10−5 to 5.6×10−1) | 0.356 | 4.7×10−3 (1.8×10−5 to 5.6×10−1) | 0.349 | – | – |
Frequency of asking patients to relax the anal sphincter | 3.6×10−3 (1.2×10−5 to 5.3×10−1) | 0.355 | 4.7×10−3 (2.1×10−5 to 5.2×10−1) | 0.345 | – | – |
Frequency of rotating the swab within the anal canal | 3.0×10−3 (1.8×10−6 to 8.3×10−1) | 0.390 | 4.2×10−4 (2.0×10−30 to 1) | 0.487 | – | – |
Frequency of moistening swab before insertion | 1.2×10−6 (4.7×10−21 to 1) | 0.494 | 5.5×10−6 (2.7×10−26 to 1) | 0.496 | – | – |
*Model 1: unadjusted ICCs were calculated for each sampling technique.
†Model 2: adjusted ICCs were calculated for each sampling technique alone after adjustment for patients’ characteristics.
‡Model 3: adjusted ICCs were calculated for sampling technique after adjustment for patients’ characteristics and all clinicians’ sampling technique.
§Testing the significance of the ICC.
ICC, intracluster correlation coefficients.