STUDY OBJECTIVE: The practice patterns of US emergency departments in the treatment of patients with isolated benign headache have been recently described. How treatment varies among EDs has not been reported. To assess institutional variability in the pharmacotherapy of patients with benign headache, we describe and analyze the practice patterns of 3 US EDs. METHODS: This health records survey included a cohort sample of consecutive adult patients aged 16 to 65 years treated with parenteral medication for isolated benign headache at 3 nonaffiliated US EDs: a large, group-model health maintenance organization, a tertiary-care academic center, and a rural community hospital. Patients who underwent a diagnostic search for intracranial pathology, who had any nonheadache secondary diagnosis, or who had coexistent trauma, fever (temperature of > or =38.0 degrees C [100.4 degrees F]), or known pregnancy were excluded from study analysis. Demographic, clinical, and pharmacotherapeutic variables were collected for each ED visit. Descriptive analyses were performed; comparisons were made with t tests. RESULTS: Of the 490 eligible patients treated during the 4-month study period, the mean age was 36.4 years, and 374 (76%) were women. During their 629 visits, 364 (58%) received a migraine diagnosis, and 258 (41%) received a nonspecific headache diagnosis. Polypharmacy was common: 515 (82%) received 2 or more medications, and 154 (25%) received 3 or more medications. Pharmacotherapy varied greatly among the EDs. Use of opioid agonists showed the widest variation (16% to 72%), although use of dihydroergotamine (5% to 16%), prochlorperazine (32% to 59%), and adjunct diphenhydramine with prochlorperazine (42% versus 88%) also varied. CONCLUSION: Great institutional variability exists among US EDs in the parenteral treatment of patients with isolated benign headache.