Psychometric properties of a new instrument for assessment of treatment satisfaction in insulin-requiring diabetic patients, specifically adapted for evaluation of potential advantages of insulin analogs used either for basal, prandial or correctional purposes

Sensitive instruments are needed to assess the effects of minor treatment changes resulting e.g. from a change of an insulin preparation. To assess such small improvements in diabetic patients treated with Functional Insulin Treatment (FIT) targeting on full flexibility of food intake and life style by appropriate choice of insulin dosage for correctional, prandial or basal purposes, we extended the Status (S) version of the DTSQ (Diabetes Treatment Satisfaction Questionnaire; Bradley, 1994) with items measuring specific FIT components of short-acting insulin (Howorka and Bradley, 2000) and recently for the delayed-acting insulin (22-item version). This extended version (DTSQ-S/22/) was investigated in a validation study with 180 patients (age 46+16, diabetes duration 18+11yrs, women 59%, self-measurements 7+1/d, injections 6.3+1.3/d, HbA1c 7.3+1.0%). Cronbach's alpha coefficients were satisfactory for the classical scale /first 8 items/ (alpha=0.85) and for the 22-items extended scale (alpha=0.91). In factor analysis, principal component analysis of 22 items scale yielded 4 factors explaining 62% of the variance: Factor one described the dimension “satisfaction with correctional and prandial (short acting) insulin“ and explained 19% of the variance, factor two the dimension “satisfaction with predictability, stability and understanding of blood glucose course“, explained 16% of variance. Factor three dealt with more general “satisfaction with current treatment, wish-to-continue and satisfaction with convenience and flexibility“ and explained 14% of the variance. Factor four described “satisfaction with hypoglycaemia perception and treatment“ and explained 13% of the variance. Corrected item-total correlations for the new items related to basal replacement were all p<0.01. As an indication of criterion-related validity significant correlations were found between the sum score of the extended instrument and quality of life assessed by ADDQoL. Extended German DTSQ version shows high levels of internal consistency useful for assessment of minor treatment changes.