Self-completion instruments assessing subjective health are increasingly being used to measure patient outcome. However, there is very little evidence as yet of existing instruments' responsiveness to change. This paper describes a study to evaluate the responsiveness to change of a self-completion instrument for the measurement of clinical outcome in patients with diabetes. A prospective follow-up study of 48 patients with non-insulin-dependent diabetes commencing insulin therapy was carried out, with assessments at baseline (i.e. pre-intervention), 6 weeks and 3 months post-intervention. The outcome measures used were the changes in scores on the self-completion instrument for symptom level, physical function, energy, depression, psychological distress and barriers to activity, HbA1c, non-fasting serum cholesterol and the body mass index (BMI). There were significant improvements in HbA1c and non-fasting serum cholesterol and worsening of the BMI, as expected. The self-completion instrument detected significant improvements in patient-reported symptoms within 6 weeks of the intervention (p < 0.01) and in energy levels (p < 0.05). There is evidence from this study of the self-completion instrument's ability to respond to change and it has potential for use in evaluative studies.