Problems with the performance of the SF-36 among people with type 2 diabetes in general practice q Alison J. Woodcock 1 , Steven A. Julious 2,  , Ann Louise Kinmonth 3 & Michael J. Campbell 4 1 Department of Psychology, Royal Holloway, University of London; 2 Medical Statistics and Computing, University of Southampton; 3 Department of Public Health and Primary Care, University of Cambridge; 4 SCHARR, Institute of General Practice and Primary Care, University of Sheffield, UK   present address: Clinical Pharmacology Statistics, GlaxoSmithKline, Harlow, UK Accepted in revised form 15 October 2001 Abstract Objective: To validate the short form-36 (SF-36) among people with type 2 diabetes in general practice, and to make comparisons with the Audit of Diabetis Dependent Quality of Life (ADDQoL). Design: Postal survey with one reminder. Setting: Four general practices. Patients: One hundred and eighty-four eligible patients (30–70 years) with type 2 diabetes on 14 general practitioner lists. Measures: SF-36 response rates, distribution of dimension scores and internal consistency. Median scores in relation to sociodemography and self-reported health. Comparisons with ADDQoL scores. Results: One hundred and thirty-one pa- tients responded (71%). Distributions of SF-36 dimension scores were mostly skewed. Internal consistency and construct validity were acceptable, with predictable sociodemographic trends. People with illness related to or unrelated to diabetes scored significantly lower on most dimensions. SF-36 dimension scores correlated best with relevant diabetes-specific ADDQoL scores amongst respondents reporting no com- orbidity. Conclusions: Although valid and reliable, SF-36 scores are strongly affected by non-diabetic comorbidity in type 2 diabetes, supporting the complementary use of a diabetes-specific measure, providing information about the impact of diabetes specifically. Key words: Audit of Diabetes Dependent Quality of Life, General practice, Quality of life, Short form-36, Type 2 diabetes Introduction Comprehensive assessment of the costs and bene- fits of new technologies in chronic diseases in- cluding diabetes [1, 2] demands a range of measures, spanning the biomedical, psychosocial and economic. In preparation for such an assess- ment of general practice diabetes care [3], the ge- neric short form-36 (SF-36) health survey questionnaire [4–6] was validated and compared with the disease-specific Audit of Diabetes De- pendent Quality of Life (ADDQoL) [7]. The 36 questions in the SF-36 span functional health status and general health, currently and over the previous 4 weeks. Validated in a UK general population [6], it discriminates between perceptions of people with and without one or more chronic diseases [6], between older and younger people [6, 8] and between people with and without diabetes (type unspecified) [8]. In one study, those with diabetes reported poorer health in all dimensions except pain [9]. The ADDQoL is attracting interest as an out- come measure for diabetes [10, 11]. It measures q On behalf of the Diabetes Care From Diagnosis Group, University of Southampton. Quality of Life Research 10: 661–670, 2001. Ó 2001 Kluwer Academic Publishers. Printed in the Netherlands. 661