STUDY DESIGN: A prospective repeated measures design was used to produce a back-specific version of the Short Form-36 Physical Functioning scale (SF-36 PF) by Rasch analysis of a pool of items from the SF-36 PF, Oswestry Disability Questionnaire, and the Quebec Back Pain Disability Scale. OBJECTIVES: To identify items for a back-specific version of the SF-36 PF scale and to compare the psychometric properties of the new version with the original 10-item scale. SUMMARY OF BACKGROUND DATA: Adequate assessment of patient function requires the administration of a generic and a condition-specific questionnaire. A back-specific version of the SF-36 PF would facilitate comprehensive patient assessment in the clinical setting. METHODS: Consecutive patients with low back pain presenting for physiotherapy treatment were recruited at three public hospitals, three community health services, and four private practices. Patients completed questionnaires on two occasions 6 weeks apart. RESULTS: A scale of 18 items showed a better fit to the Rasch model than the original SF-36 PF scale. Items in the original scale that had a poor fit (INFIT/OUTFIT statistics outside the range 0.7-1.3) showed an acceptable fit in the new scale. The augmented scale had comparable reliability and improved responsiveness to the original 10-item SF-36 PF scale. The minimum detectable change (90% confidence) and the minimum clinically important difference were 12 points. Floor and ceiling effects were practically eliminated. The psychometric properties of the new scale were comparable to those of the Oswestry Disability Questionnaire. CONCLUSIONS: The Low-Back SF-36 PF18 comprises the 10-item SF-36 PF scale and four items each from the Oswestry and Quebec back pain questionnaires. The possible total score ranges from 0 to 100, with a higher score indicating better function. The new scale appears to offer advantages over the use of the original scale for the assessment of functioning in patients with low back pain.