Raking is a standard method in survey sampling that makes use of auxiliary information on the population to improve upon the simple Horvitz-Thompson estimator. Lumley et al (Int Stat Review 2011) presented an example of how raking can be used in the setting of error-in-variables to improve upon the simple regression calibration estimator (Prentice, Biometrika 1982) in an application involving a time-to-event outcome, where this error-correction method is expected to have some bias. We further explore raking as a method to improve upon the regression calibration estimator for a more general underlying error model where there may be correlated errors in both the outcome and covariates. This setting is common in medical applications that depend on electronic health records data, for which the data were not collected to support research questions under study. We will discuss the choice of the auxiliary variable and features of the underlying estimation problem that affect the degree of improvement the raking estimator will have over the simpler, biased regression calibration approach. Results from a simulation study and applications to real data will be presented.