OBJECTIVES: Electronic health records (EHR) have been increasingly used for comparative effectiveness research. It is unclear how care-discontinuity, defined as receiving care outside of an EHR system, may affect data completeness and study validity using EHR. We aimed to quantify care-continuity of an EHR system and compare the misclassification of key variables in patients with high vs. low care-continuity. METHODS: Study cohort comprised all patients ≥ 65 in EHR from two large US provider networks linked with Medicare insurance claims data from 2007/1/1 to 2014/12/31. By comparing EHR and claims data, we quantified care-continuity by the Mean Proportion of Encounters Captured (MPEC) by the EHR system. Within levels of care-continuity, we quantified misclassification by Mean Standardized Differences between the proportions of 40 key variables based on EHR alone vs. linked claims-EHR data (MSD_{_40_variables}, <0.1 was used to indicate satisfactory variable classification). We compared patient characteristics in those with high vs. low EHR continuity. RESULTS: Based on 104,403 patients in EHR system 1 and 79,336 in EHR system 2, the mean capture proportions of all records were 24% and 18% in system 1 and 2, respectively. The misclassification of key variables (MSD_{_40_variables}) based on EHR alone was 11.5-19.6 fold greater in those with lowest level of care-continuity (MPEC< 10%) than that in those with highest level of care-continuity (MPEC >= 80%), across EHR systems and years during follow-up. In both systems, capturing at least 60% of the encounters in an EHR was required to have satisfactory variable classification. The patient characteristics in those with high and low care-continuity were found to be comparable. CONCLUSIONS: Care-discontinuity may lead to substantial misclassification in key variables. Researchers may consider restriction to those with high care-continuity to improve study validity when relying exclusively on EHR data.