Table 2.
Results of the mixed-effect Bayesian inference-based integrated nested Laplace model showing the association between race/ethnicity, year and HIV clinical outcomes before and after socio-ecological factors were adjusted
| Individual-level variable adjusted | Both individual and socio-ecological factors adjusted | |||
|---|---|---|---|---|
| Odds ratio of immune reconstitutiona | Odds ratio of viral suppression | Odds ratio of immune reconstitutionb | Odds ratio of viral suppression b | |
| Race/ethnicity | ||||
| Black | Reference | Reference | Reference | Reference |
| Hispanic | 1.70 (1.54-1.87) | 2.18 (2.07-2.31) | 1.66 (1.50-1.83) | 2.16 (2.04-2.29) |
| White | 2.34 (2.14 -2.56) | 1.95 (1.85-2.05) | 2.28 (2.08-2.51) | 1.88 (1.78-1.98) |
| Other/unknown | 0.92 (0.77-1.26) | 1.37 (1.20-1.57) | 0.98 (0.77-1.25) | 1.36 (1.19-1.55) |
| Year | ||||
| 2012 | Reference | Reference | Reference | Reference |
| 2013 | 1.27 (1.19-1.34) | 1.14 (1.09-1.20) | 1.28 (1.18-1.39) | 1.19 (1.12-1.26) |
| 2014 | 1.78 (1.69-1.88) | 1.93 (1.84-1.20) | 1.74 (1.59-1.90) | 1.97 (1.84-2.11) |
| 2015 | 2.31 (2.18-2.44) | 2.52 (2.40-2.64) | 2.33 (2.14-2.54) | 2.61 (2.43-2.79) |
| 2016 | 2.53 (2.40-2.68) | 3.09 (2.95-3.24) | 2.50 (2.29-2.74) | 3.23 (3.00-3.48) |
| 2017 | 3.35 (3.17-3.55) | 3.64 (3.47-3.81) | 3.52 (3.19-3.88) | 3.93 (3.63-4.25) |
| 2018 | 4.20 (3.97-4.45) | 4.07 (3.88-4.27) | 4.51 (4.05-5.03) | 4.65 (4.26-5.08) |
| 2019 | 5.91 (5.59-6.26) | 6.39 (6.07-6.72) | 6.61 (5.87-7.43) | 7.34 (6.67-8.08) |
Findings were derived from Bayesian inference-based integrated nested Laplace approximation. Results are presented as odds ratios and 95% credible intervals
aModel included race/ethnicity, measurement year, age at diagnosis, years since HIV diagnosis, and transmission risk category
bModels adjusted for race/ethnicity, measurement year, age at diagnosis, years since HIV diagnosis, and transmission risk category, in addition to the following county-level socio-ecological factors: social economic factors rank, county averaged years of potential life lost, county averaged mentally unhealthy days, low birthweight percentage, percentage of adults reporting binge or heavy drinking, uninsured percentage, ratio of population to primary care physicians, high school graduation rate, and crime rate