OBJECTIVES: The purpose of this paper is to propose a simple and efficient approach for analyzing multivariate longitudinal data (MLD), such as longitudinal health-related quality of life (HRQoL) assessments in observational studies. This will be accomplished using the combination of two popular statistical methods for causal inference and multivariate data, namely the inverse probability-weighted (IPW) estimator and principal component analysis (PCA). METHODS: Multivariate outcomes at each time point will be converted to the first principal component score (FPCS) for each subject. Then all FPCS will be composited into a numerical observation using the area under a curve (AUC). The IPW estimator is used to compare the difference between the two treatments in terms of the estimated AUCs. Finally, the proposed method will be applied to a simulated dataset to determine if there is significant difference between two treatments. RESULTS: The statistical results show that the 95% bootstrap percentile confidence interval (BPCI) is (-3.06, -3.59). Since the BPCI does not contain zero, we claim that the treated group (M=1) is significantly different from the control group (M=0) at a 5% level in overall longitudinal multivariate health outcomes.  CONCLUSIONS: In this paper, we propose a simple and efficient approach to overcome the difficulty of analyzing MLD in practice. We demonstrate how to use our proposed method with a simulated dataset. Our simulated data set allowed us to demonstrate how our proposed method may be particularly useful for analyzing longitudinal HRQoL assessments in medical studies.