A Bayesian Approach to Improving Spatial Estimates After Accounting for Misclassification Bias in Surveillance Data for COVID-19 in Philadelphia, PA