Healthcare resource utilization (HCRU) may be evaluated through patient interviews in multicenter clinical trials but the accuracy of self-reported data should be assessed. OBJECTIVE: This study evaluated the consistency between electronic record data and patient self-reported HCRU for herpes zoster (HZ) through telephone interviews. METHODS: HZ patients (N=116) were recruited from a managed care organization in Boston. HZ-related hospitalizations, Emergency Department (ED) visits, outpatient visits and telephone calls were compared separately, with further differentiation among outpatient contacts with primary care providers and specialists. Medication use comparisons were made for analgesics, antivirals and other prescription medications. Judgments of consistency were based on intraclass correlation coefficients (ICC) unless data were sparse. RESULTS: Two patients with HZ hospitalizations were found in the electronic records versus three in the questionnaire. Four patients with ED visits were found in the electronic records versus 34 in the questionnaire. The ICC for outpatient visits was 0.46 (CI 0.06-0.85; nelectronic records=268, nZHEQ=265). The ICC for telephone calls was 0.51 (CI 0.06-0.96; nelectronic records=87, nZHEQ=110). By outpatient provider type, the ICC ranged from 0.27 for primary care visits to 0.84 for specialist visits. ICCs of 0.55, 0.64 and 0.56 were found for pain medications, antiviral medications, and other medications, respectively. CONCLUSION: Patient self-reported HCRU reasonably (ICC =0.4), but imperfectly matched the electronic records across most categories, with no systematic bias observed. Patient misclassification of urgent care visits as ED visits may account for some of the observed discrepancies between ED and primary care visits. This study, which represents one of the first efforts to evaluate the design of a patient self-report HCRU questionnaire for use in clinical trials, identified limitations to comparisons between the two types of data sources and offers insight into potential improvements for the design and validation of such questionnaires.