Psychological outcome measures of Well-being and Treatment Satisfaction have been designed and developed for people with tablet-treated Type 2 diabetes. The Well-being scale includes three six-item sub-scales to measure Depression, Anxiety, and Positive Well-being. A prime consideration when selecting items for the psychological well-being measures was to minimize the confounding of diabetic symptomatology with the somatic symptoms of depression and anxiety. Cronbach's alpha indicated that each of the Well-being sub-scales and the Treatment Satisfaction scale was internally reliable (alphas ranged from 0.70 to 0.88) and evidence for construct validity was provided by predicted associations with other variables collected at the time of the study. For example, lower Well-being scores were associated with being overweight (Depression: p less than 0.05; Anxiety: p less than 0.001) while greater Satisfaction with Treatment was associated with lower HbA1 levels (p less than 0.001) and lower percent ideal body weight (p less than 0.01). These scales should prove particularly useful where measures of quality of life are required to complement metabolic variables when evaluating new treatments, education programmes, and other interventions, or in the routine auditing of established methods of treatment.

The psychometric properties of a diabetes-specific treatment satisfaction scale were examined with responses from 128 adults with insulin-dependent diabetes who had used one of three treatment options for a period of 12 months. The reliability of the seven-item measured was found to be satisfactory (Cronbach's Alpha = 0.76) and factor analyses indicated three useful sub-scales (Perceived General Management; Perceived Compatibility with Lifestyle; Perceived Frequency of Hypo/hyperglycaemia). Use of the treatment satisfaction measure in a feasibility study of continuous subcutaneous insulin infusion (CSII) demonstrated the measure's ability to distinguish between three treatment groups (CSII, intensified conventional therapy and conventional therapy). People choosing to use CSII reported significantly greater improvements in satisfaction than those choosing either from of conventional therapy (F = 36.6; df 2, 125; p less than 0.001). If used in conjunction with measures of blood glucose control, the Treatment Satisfaction measure offers the opportunity for a more holistic appraisal of outcomes in studies evaluating and comparing treatments for insulin-dependent diabetes.

Glycemic control improved with both methods of insulin treated patients achieved satisfactory glycemic control (HbA1 less than 50 mmol hydroxymethylfurfural/mol Hb), whereas only 3 of 10 CIT-treated patients achieved this CSII. Patients' satisfaction with treatment improved during insulin therapy.

The design and development of scales to measure diabetes-specific health beliefs of patients treated with oral hypoglycaemic agents and diet are described. The scales, which were developed from the responses of 187 tablet-treated patients, measure perceived Benefits of, and Barriers to Treatment, and perceived Severity of, and Vulnerability to Complications of diabetes and other disorders unrelated to diabetes. Cronbach's alpha indicated that each scale was internally reliable (alpha = 0.67 to 0.89) and construct validity was indicated by expected correlations between the scales and other variables including measures of blood glucose control, body weight, and psychological well-being. Thus, patients who were overweight and/or had higher glycosylated haemoglobin (HbA1) levels perceived their treatment to be less 'Cost-effective' (Benefits less Barriers), rated Complications as being more severe, and reported greater Vulnerability for themselves and the 'average person' with diabetes to these Complications (p less than 0.05 to p less than 0.01). Furthermore, respondents with higher Depression and Anxiety scores and lower Positive Well-being and Treatment Satisfaction scores reported greater Vulnerability to Complications for themselves and the 'average person', and also felt that their treatment was less 'Cost-effective' (p less than 0.01 to p less than 0.001). The measures are likely to be particularly useful in studies which assess the efficacy of interventions in modifying health beliefs with a view to achieving desired behaviours. They may also be useful as an instrument of audit if indicators about the suitability of possible interventions are sought.