A comparison of measures of disability and health status in people with physical disabilities undergoing vocational rehabilitation.

BACKGROUND The aim of the study was to test among people undergoing vocational rehabilitation six measures of disability and health status commonly used in medical rehabilitation. METHOD A cross-sectional survey was carried out on 30 people with disabilities on an occupational programme at a non-medical rehabilitation facility in England. Measures used were the Barthel index, Extended Activity of Daily Living (EADL) scale, the Office of Population Censuses and Surveys (OPCS) disability questionnaire, the Functional Independence Measure (FIM), the Nottingham Health Profile (NHP) and the Medical Outcome Study Short Form 36 (SF36). The main outcome measures were item non-response and time for completion of each measure, profile of disability and health status described by each measure, including pertinent domains detected or missed, and floor and ceiling effects. RESULTS Item non-response was very low with all the instruments; the Barthel index was on average the quickest to complete (mean time 2.2 minutes) and the SF36 the longest (mean time 9.1 minutes). The study group were characterized as having problems in mobility or locomotion and bladder or bowel control, but some of the instruments were insensitive, detecting no disability in many subjects (e.g. 33 per cent showed no disability on the Barthel index). The SF36 scores were the least affected by floor and ceiling effects; mean SF36 scores on all scales except physical functioning were similar to those for the general population of similar age. Some problems were detected by only one of the instruments (e.g. pain and sleep by the NHP, problems with intellectual functioning by the OPCS scale). CONCLUSION Disability measures commonly used in medical rehabilitation, such as the Barthel score and FIM, may be less useful in vocational rehabilitation where disabilities are less severe. Other measures show more promise but further testing is needed.