Figure 7.

Multivariable logistic regression modelling of the total number of viral infections amongst all participants. The model shows that female sex, HIV+ status, and older age are independently associated with having more chronic/latent viral infections (after additional adjustment for ethnicity, tobacco smoking status, and region of birth). This model was selected by minimizing Akaike’s Information Criterion (AIC) and maximizing the coefficients of determination (r²). The center points depict the β value and lines depict 95% confidence internals. Red intervals designate statistical significance. Negative β values indicate that the specified variable has the fewer viral infections compared to the reference group, whereas positive β values indicate the presence of more viral infections. The coefficient of determination is shown at the top. n = 6 excluded due to missing data (region of birth); n = 2 excluded due to South American region of birth (too few events).