Structural data from Electronic Health Records as complementary information to imaging data for disease prediction. We incorporate novel weighting layer into the Graph Convolutional Networks, which weights every element of structural data by exploring its relation to the underlying disease. We demonstrate the superiority of our developed technique in terms of computational speed and obtained encouraging results where our method outperforms the state-of-the-art methods when applied to two publicly available datasets ABIDE and Chest X-ray in terms of relative performance for the accuracy of prediction by 5.31 % and 8.15 % and for the area under the ROC curve by 4.96 % and 10.36 % respectively. Additionally, the model is lightweight, fast and easily trainable.