OBJECTIVE: To compare health services utilization and cost among commercially insured users of lispro insulin and non-lispro regular insulin. METHODS: This study used administrative claims data to identify continuously enrolled patients using insulin between 1/1/98 and 12/31/99. During a six-month identification period, patients receiving one or more prescriptions for lispro were categorized as lispro patients. Non-lispro patients were then matched to lispro patients using a Propensity Score (PS) derived from baseline characteristics. The PS model included variables such as age, gender, comorbidities, oral hypoglycemic use, dominant physician specialty, health plan location and baseline costs. After 1:1 matching, 12 months of follow-up cost and utilization data were then compared using unmatched t-tests. RESULTS: Of 11,443 patients, 3,341 (29.2%) had at least one prescription for lispro insulin, while the remaining 8,102 (70.8%) had at least one regular (non-lispro) insulin prescription. At baseline, lispro patients tended to be younger, were more often Type 1 with a history of insulin use, had fewer comorbidities, visited endocrinologists over family practice physicians and had lower total costs. PS balancing assured that only the 1,832 most appropriate patients of each type were then used in outcomes comparisons. Lispro patients had significantly higher average office visits (p=0.0022) and pharmacy prescriptions (p=0.0165) but lower inpatient hospital visits (p=0.0028) compared to non-lispro patients. Cost results were similar with lispro insulin patients having significantly higher average office visit costs (p=0.0237) and pharmacy costs (p<.0001) but lower inpatient hospital costs (p=0.0227). Total costs were not significantly different (p=0.5266). CONCLUSIONS: Lispro insulin and non-lispro regular insulin patients did not incur significantly different total costs. While lispro patients had higher pharmacy and office visit costs, they did not incur higher total medical costs. More intensive ambulatory care in such chronic disease patients does not appear to be associated with higher overall medical costs.