Sometimes clinical trials collect survival data, which have some variables measured longitudinally. This type of data is mostly analyzed using Cox proportional models with time dependent covariates. The longitudinal variables are treated as time dependent covariates. When there is association between a longitudinal variable and the time to event, estimates produced from separate models may be biased. The study uses Cox proportional models with time dependent covariates for survival data and linear mixed effects regression models for the longitudinal data. For the joint analysis, the joint modeling between repeated measurement and time to an event is used. The method is applied to data from a randomized clinical trial for the malnourished HIV positive patients who were on ART at Queen Elizabeth Central Hospital. One group received corn soya blend (CSB) and other group received ready to use therapeutic food (RUTF). Results from joint modeling showed that there is significant association between body mass index (BMI) and time to death of a patient, p < 0.001. Both joint model and Cox proportional model with time dependent covariates showed that the type of food did not have significant effect on the time to death of patients. Hemoglobin levels, sex of patient and use cotrimoxazole were significantly associated with time to death of malnourished HIV positive patients. It was also observed that some variables which were not significant in the separate models became significant in the joint model. This shows the importance of using joint models. Joint modeling of longitudinal and survival data gives unbiased estimates.