Skip to main content
. 2023 Feb 8;39(2):57–67. doi: 10.1089/aid.2022.0054

Table 4.

Multivariable Analysis of Factors Associated with Recent and Long-Term HIV Infection in Multinomial Logistic Regression, Kenya 2012 and 2018 Combined (n = 38,001)

Characteristic Recent infection compared to HIV-uninfected
Long-term infection compared to HIV-uninfected
aOR 95% CI aOR 95% CI
Year
 2012 3.2 1.2–8.5 0.92 0.79–1.1
 2018 (Ref.)        
Sex
 Male (Ref.)        
 Female 2.4 0.77–7.5 3.1 2.6–3.7
Province        
 Nyanza/Western 1.9 0.76–4.8 2.3 1.9–2.6
 Other (Ref.)        
Age group (years)
 15–24 3.1 0.76–13 0.17 0.13–0.20
 25–34 4.2 1.4–13 0.64 0.56–0.74
 35–64 (Ref.)        
Education
 No education/Primary (Ref.)        
 ≥Primary 2.8 0.88–9.1 1.5 1.3–1.7
Lifetime number of sex partners
 0–1 (Ref.)        
 2–3 5.2 1.6–17 1.8 1.5–2.1
 ≥4 8.6 2.8–26 2.9 2.4–3.6
 Unknown/Missing 13 2.4–71 1.3 0.91–2.0
Circumcision status (males only)
 Circumcised (Ref.)        
 Not circumcised 2.4 0.49–11 4.3 3.4–5.4
Ever tested for HIV
 Yes (Ref.)        
 No 4.1 1.5–11 0.33 0.26–0.42
Genital ulcer/sore in last 12 months
 Yes 0.91 0.25–3.3 1.8 1.4–2.3
 No (Ref.)        
Used condom at last sex in last 12 months
 Yes (Ref.)        
 No 2.1 0.48–9.7 0.15 0.13–0.18
 Not sexually active 0.56 0.10–3.1 0.44 0.37–0.53
 Unknown/Missing 1.9 0.36–10 0.38 0.28–0.52

Recent and long-term HIV infection according to full RITA based on LAg-avidity assay, antiretroviral treatment (self-report and biomarker) and viral load ≥1,000 c/mL. Items in bold are significant (CI excludes 1.0). Odds ratios rounded to two significant digits.