Effect sizes from included studies and meta-analysis model summaries
| Full samples | HIV-negative subgroups | HIV-positive subgroups | ||||
|---|---|---|---|---|---|---|
| OR (95% CI) | Weight (%) | OR (95% CI) | Weight (%) | OR (95% CI) | Weight (%) | |
| Unprotected anal intercourse (UAI) | ||||||
| Chiasson et al38 | 0.98 (0.78 to 1.23) | 11.4 | 1.08 (0.82 to 1.42) | 11.5 | 2.33 (1.18 to 4.60) | 9.7 |
| Grov et al14 | 1.52 (1.27 to 1.83) | 12.0 | 1.51 (1.19 to 1.92) | 12.8 | 2.50 (1.64 to 3.81) | 13.1 |
| Menza et al41 | 1.39 (0.99 to 1.96) | 9.7 | 1.43 (0.97 to 2.11) | 7.5 | 0.84 (0.38 to 1.84) | 8.5 |
| MINTS II12 42 43 | 0.97 (0.73 to 1.27) | 10.7 | 0.92 (0.76 to 1.11) | 15.3 | 1.56 (0.86 to 2.84) | 10.7 |
| Mustanski49 | 0.40 (0.17 to 0.94) | 4.0 | ||||
| NHBS I39 | 1.36 (0.93 to 1.99) | 9.1 | 1.10 (0.84 to 1.45) | 11.4 | 1.21 (0.68 to 2.16) | 11.0 |
| NHBS II40 | 0.98 (0.88 to 1.09) | 12.9 | 0.98 (0.88 to 1.09) | 19.7 | ||
| SILAS IV | 1.23 (1.14 to 1.33) | 13.1 | 1.23 (1.13 to 1.34) | 20.6 | 1.39 (1.12 to 1.73) | 15.5 |
| Smith et al46 | 1.25 (0.75 to 2.08) | 7.3 | 1.00 (0.61 to 1.64) | 12.1 | ||
| Tieu et al47 | 2.12 (0.80 to 5.60) | 3.3 | 0.73 (0.22 to 2.42) | 1.1 | 4.02 (1.80 to 8.98) | 8.3 |
| Wilson et al48 | 3.44 (1.92 to 6.15) | 6.4 | 3.44 (1.92 to 6.15) | 11.0 | ||
| Model summary | 1.24 (1.01 to 1.52) I2=88%; Q10=46.1 (p<0.0001)* | 1.13 (0.99 to 1.29) I2=70%; Q7=22.2 (p=0.0024) | 1.75 (1.26 to 2.43) I2=73%; Q8=25.7 (p=0.0012) | |||
| UAI in group sex | ||||||
| Chiasson et al38 | 0.52 (0.30 to 0.92) | 29.8 | 0.44 (0.21 to 0.92) | 28.9 | 0.86 (0.12 to 5.94) | 8.4 |
| Grov et al14 | 1.69 (1.16 to 2.48) | 33.6 | 1.81 (1.08 to 3.05) | 33.3 | 3.38 (1.42 to 8.04) | 29.7 |
| SILAS IV | 1.51 (1.28 to 1.78) | 36.6 | 1.48 (1.23 to 1.78) | 37.8 | 1.56 (1.09 to 2.24) | 61.9 |
| Model summary | 1.14 (0.57 to 2.28) I2=91%; Q2=13.3 (p=0.0013) | 1.11 (0.50 to 2.48) I2=89%; Q2=10.6 (p=0.0051) | 1.87 (1.03 to 3.38) I2=37%; Q2=3.1 (p=0.2) | |||
| Serosorting | ||||||
| Chiasson et al38 | 1.52 (1.09 to 2.10) | 15.6 | 1.41 (1.00 to 1.99) | 15.2 | 2.83 (1.10 to 7.26) | 10.7 |
| Grov et al14 | 3.41 (2.40 to 4.85) | 15.2 | 2.91 (1.92 to 4.43) | 13.8 | 4.74 (2.43-9.25) | 15.8 |
| Menza et al41 | 1.37 (1.12 to 1.67) | 17.6 | 1.43 (1.14 to 1.78) | 17.6 | 1.37 (0.86 to 2.18) | 21.0 |
| NHBS I39 | 1.80 (1.40 to 2.32) | 16.0 | 1.84 (1.42 to 2.39) | 17.0 | 0.92 (0.23 to 3.75) | 18.1 |
| MINTS II12 42 43 | 0.99 (0.73 to 1.34) | 16.8 | 0.90 (0.70 to 1.16) | 16.9 | 1.60 (0.90 to 2.84) | 6.0 |
| SILAS IV | 1.47 (1.38 to 1.56) | 18.8 | 1.40 (1.31 to 1.49) | 19.5 | 2.16 (1.79 to 2.62) | 28.4 |
| Model summary | 1.61 (1.18 to 2.19) I2=92%; Q5=31.4 (p<0.0001) | 1.51 (1.14 to 2.01) I2=90%; Q5=28.0 (p<0.0001) | 2.06 (1.40 to 3.02) I2=64%; Q5=11.7 (p=0.0397) | |||
| Potential serodiscordant UAI (SDUAI) | ||||||
| Chiasson et al38 | 0.89 (0.67 to 1.19) | 14.9 | 0.78 (0.52 to 1.17) | 13.9 | 1.53 (0.72 to 3.28) | 11.5 |
| Grov et al14 | 1.01 (0.81 to 1.26) | 17.3 | 0.95 (0.71 to 1.26) | 17.6 | 1.03 (0.66 to 1.60) | 14.3 |
| Menza et al41 | 0.89 (0.70 to 1.13) | 16.8 | 0.99 (0.72 to 1.37) | 16.4 | 0.63 (0.40 to 1.02) | 14.0 |
| MINTS II12 42 43 | 0.90 (0.55 to 1.46) | 9.6 | 1.31 (0.73 to 2.37) | 9.4 | 0.38 (0.14 to 1.02) | 9.6 |
| NHBS I39 | 1.34 (0.93 to 1.94) | 12.5 | 1.50 (1.10 to 2.05) | 16.8 | 0.90 (0.50 to 1.62) | 13.0 |
| SILAS IV | 0.76 (0.70 to 0.83) | 21.0 | 0.72 (0.65 to 0.81) | 22.8 | 0.86 (0.73 to 1.00) | 16.0 |
| Tieu et al47 | 2.37 (0.94 to 5.98) | 3.8 | 0.73 (0.22 to 2.42) | 3.2 | 4.02 (1.80 to 8.98) | 11.1 |
| Wilson et al48 | 2.90 (1.20 to 7.01) | 4.1 | 2.90 (1.20 to 7.01) | 10.4 | ||
| Model summary | 1.01 (0.83 to 1.23) I2=73%; Q7=27.2 (p=0.0003) | 0.97 (0.77 to 1.22) I2=71%; Q6=24.4 (p=0.0004) | 1.13 (0.70 to 1.82) I2=86%; Q7=27.8 (p=0.0002) | |||
| SDUAI with strategic positioning | ||||||
| Chiasson et al38 | 1.23 (0.33 to 4.61) | 8.6 | ||||
| Grov et al14 | 2.73 (0.32 to 23.37) | 3.3 | ||||
| MINTS II12 42 43 | 0.42 (0.04 to 4.68) | 2.6 | ||||
| SILAS IV | 2.31 (1.56 to 3.41) | 85.6 | ||||
| Model summary | 2.10 (1.43 to 3.10) I2=1%; Q3=2.6 (p=0.5) | |||||
*Q10 denotes the value of the test statistic for k–1=10° of freedom; p denotes the type II error for rejecting the null hypothesis that the pooled estimate is without heterogeneity.
NHBS, National HIV Behavioral Surveillance; SDUAI, serodiscordant UAI.