Longitudinal Modeling of CD4 Cell Count Trajectories among HIV Patients Receiving Antiretroviral Therapy in Pretoria, South Africa
To assess immune recovery among HIV patients undergoing antiretroviral therapy (ART), it is essential to monitor the CD4 cell count. Optimizing patient management is critical for understanding CD4 progression and its determinants. A longitudinal dataset comprising 4,317 observations from 841 HIV patients was analyzed using linear mixed-effects models. Random intercept and slope models were fitted to account for within-patient correlation. Nonlinear CD4 trajectories were modeled using quadratic time effects, and covariates including age, sex, tuberculosis status, and adherence were incorporated. CD4 count increased significantly over time, with an initial rapid increase followed by a plateau ($\beta_1 = 97.88$, $p < 0.001$; $\beta_2 = -58.14$, $p < 0.001$). Significant heterogeneity occurred in baseline levels and rates of change. Male patients had lower CD4 counts than females ($\beta = -32.30$, $p < 0.001$). Tuberculosis status had a negative but borderline effect, while age and adherence were not statistically significant. The findings support individualized treatment strategies and continuous monitoring.