Even a small improvement in satisfaction with treatment for a chronic disease can be valuable. However, sensitive measurements instruments are needed to assess the effects of treatment changes in patients already well satisfied with baseline treatment. Such instruments were thought to be necessary to deal with ceiling effects in scores of satisfaction with functional insulin treatment (FIT) accommodating full flexibility of food intake and lifestyle in diabetes by manipulation of insulin (FIT; Howorka et al. 2000). The Status(S) version of the Diabetes Treatment Satisfaction Questionnaire (DTSQ, Bradley 1994) was extended with items measuring specific components of FIT and its psychometric properties investigated in a validation study with 171 FIT patients with diabetes. A new Change(C) version of DTSQ extended for FIT was used together with the DTSQ(S) by 52 participants in a subsequent randomised cross-over study involving new fast-acting lispro vs. regular insulin, where treatment satisfaction was the primary outcome variable. Insulin lispro use improved satisfaction (p < or = 0.001) both, on classical and extended versions of DTSQ(S) and (C), despite high baseline levels of satisfaction. However, the DTSQ(C) augmented the effects of lispro (p = 0.0001) apparent on DTSQ(S) scores and revealed significant treatment effects (p < or = 0.01) not significant with DTSQ(S). In parallel, glycaemic control improved with lispro (e.g., HbA1c, p = 0.023). Improved satisfaction with treatment was more apparent with DTSQ(C) than DTSQ(S) in patients who at baseline were at or near ceiling for treatment satisfaction.