Skip to main content
. 2020 Mar 2;17:8. doi: 10.1186/s12981-020-00265-4

Table 3.

Association between hair measure and self-reported measures of optimal adherence

Adherence measures Hair TFV concentration Kappa statistics Univariate modela Multivariate modelb
Optimal
123 (45.9%)
Suboptimal
145 (54.1%)
cOR (95% CI) p aOR (95% CI) p
Frequency
 Optimal 230 (85.8%) 109 (88.6%) 121 (83.4%) 0.048 1.54 (0.76, 3.14) 0.229 1.37 (0.66, 2.86) 0.400
 Suboptimal 38 (14.2%) 14 (11.4%) 24 (16.6%)
Percent
 Optimal 261 (97.4%) 122 (98.4%) 140 (96.6%) 0.017 2.16 (0.41, 11.34) 0.362 2.37 (0.44, 12.95) 0.318
 Suboptimal 7 (2.6%) 2 (1.6%) 5 (3.4%)
VAS
 Optimal 206 (76.9%) 102 (82.9%) 104 (71.7%) 0.107* 1.91 (1.06, 3.46) 0.032 1.88 (1.03, 3.45) 0.041
 Suboptimal 62 (23.1%) 21 (17.1%) 41 (28.3%)
CAS
 Optimal 197 (73.5%) 98 (79.7%) 99 (68.3%) 0.109* 1.82 (1.04,3.19) 0.036 1.80 (1.02, 3.18) 0.043
 Suboptimal 71 (26.5%) 25 (20.3%) 46 (31.7%)

* < 0.05. Data are in numbers and percentages [n (%)], TFV is tenofovir. VAS is visual analog scale. CAS is composite adherence score

aLogistic regression; cOR is crude odd ratio; aOR is adjusted odd ratio

bAdjusted for age, sex, ethnicity, marital status, work status, education experience, monthly household income, duration of HIV diagnosis, ART regimen and duration of TDF-based ART