OBJECTIVE: The aim of this study was to determine concurrence of administrative claims data and medical records abstraction data with regard to selected comorbid conditions among hypertensive subjects. METHODS: Study subjects were incident hypertensives identified by: 1) at least one claim with an ICD-9 diagnosis code of hypertension and at least one claim for an anti-hypertensive drug; or 2) at least two claims with an ICD-9 diagnosis code of hypertension, preceded by at least six months of health plan enrollment free of anti-hypertensive prescriptions. Nurse abstractors reviewed medical records for documentation of seven comorbid conditions, including ischemic heart disease, congestive heart failure, chronic renal failure, cerebrovascular disease, complicated hypertension, hyperlipidemia, and diabetes. To determine whether comorbidities noted in medical records data appeared in claims data, we reviewed physician, facility, and pharmacy claims for a twelve-month period after the initial hypertension indicator to identify ICD-9 diagnosis codes, CPT procedure codes, or prescriptions for drugs of interest. RESULTS: 118 subjects were included in the analysis. Claims and medical records data, respectively, identified ischemic heart disease (14/15); congestive heart failure (7/10); chronic renal failure (0/1); cerebrovascular disease (3/1); complicated hypertension (18/3); hyperlipidemia (44/73); and diabetes (25/18). CONCLUSION: The extent to which claims data identified comorbidities compared to medical records data varied greatly by comorbid condition. Some of this variation may be due to limitations of medical records and claims data, including: 1) subjects may have seen additional physicians whose records were not abstracted, but subjects’ comorbidities were in claims data; 2) medical records may not contain a patient’s diagnosis, but pharmacy claims contain therapies; and 3) claims data may not identify chronic conditions if a patient is not receiving therapy.