Moderator-variable effect on the Brief Symptom Inventory test-item endorsements of spinal cord injury patients

This study shows that results from the Brief Symptom Inventory (BSI), an instrument used in several US rehabilitation centers to assess psychological distress among patients, are significantly confounded by the somatic characteristics of physically disabled patients with a specific diagnosis: spinal cord injury. Our findings also illustrate the nature and the effect of this moderator variable of physical condition on BSI item endorsement and suggest a method for controlling the moderator's variable effect. The BSI item responses of 288 persons with spinal cord injury (SCI) were compared to a non-patient normative sample of 719 persons. Thirty-six items reliably differentiated the two groups (p < 0.001) and seven of these items were judged to be physical sequelae of SCI. Other highly endorsed items describe a frequent pattern of psychological response to SCI. This response pattern does not appear to reflect the latent variables which are claimed to underlie test items. Expected endorsement of these items by patients with SCI can lead to errors in test interpretation by rehabilitation professionals.