To compare 2-year glycaemic control, hypoglycaemia and healthcare expenditures following insulin glargine (glargine, n = 2105) or neutral protamine Hagedorn (NPH) insulin (NPH, n = 734) initiation in patients with type 2 diabetes (T2D). Retrospective cohort study using an integrated US health insurance administrative database was conducted. Individuals with a diabetes diagnostic claim and initiated basal insulin therapy with glargine or NPH from 2001 to 2005 dispensed at least one oral antidiabetic drug prescription during 6 months prior to basal insulin initiation and enrolled in the same health insurance plan from 6 months before to 12 months or more after insulin initiation were identified. Repeated measures mixed-effects models evaluated glycaemic and financial outcomes to account for factors potentially contributing to selection of insulin therapy, that is, age, gender, baseline HbA1c level, health expenditures, co-morbidities, healthcare utilization, pharmacy co-payment and follow-up antidiabetic medications. Adjusted mean HbA1c value in the first year following insulin initiation was significantly lower for glargine versus NPH initiators (Δ = -0.43, p = 0.006); this difference diminished in the second year (Δ = -0.16, p = 0.375). First-year adjusted quarterly hypoglycaemia incidence rates were lower for glargine (2.1%) versus NPH (2.4%) (p = 0.02) as was the second-year quarterly rate (1.8 vs. 2.2%; p = 0.01). Both the first- and second-year adjusted total healthcare expenditures were lower in the glargine versus NPH group (year 1: $18,720 vs. $19,996, p = 0.005; year 2: $15,008 vs. $17,336; p < 0.001). Glargine therapy may be an effective long-term option for improving glycaemic control, with lower rates of hypoglycaemia and healthcare costs in patients with T2D. © 2011 Blackwell Publishing Ltd.