The last few decades have seen a cascade of research in methods for estimation and inference of causal effects, including complex methodology for longitudinal and survival outcomes. Over the same time period, joint modeling of longitudinal and survival data has arisen as a popular method for addressing issues like measurement error in longitudinal variables for survival analysis. This paper synthesizes approaches from each field to develop a novel joint modeling method which estimates causal parameters while adjusting for both confounding and measurement error by using inverse-probability weights. This approach can be applied in a variety of contexts, but is particularly useful for HIV clinical trials, where researchers are now interested in estimands defined by usage of pre-exposure prophylaxis (which is not randomized and subject to measurement error). The proposed estimator is shown to be consistent and asymptotically normal under certain assumptions and is evaluated in several simulation studies. The method is then applied to data from the HTVN 704 / HPTN 085 clinical trial to assess the joint causal effect of a randomized antibody treatment and longitudinal usage of (non-randomized) pre-exposure prophylaxis on HIV-1 acquisition.