An Efficient and Doubly Robust Empirical Likelihood Approach for Estimating Equations with Missing Data Academic Article uri icon