Abstract
Objective: To determine test-retest reliability and interrater reliability for structured interviews for the Glasgow Outcome Scale (GOS) using in-person and telephone contact.
Methods: Study 1: Thirty head-injured patients were interviewed face-to-face and then reinterviewed by telephone a few days later by the same rater. Study 2: Fifty-six head-injured patients were interviewed by telephone and then face-to-face interviews were carried out by a different person up to 1 month later. Agreement between ratings on the GOS and the extended GOS (GOSE) in each of the studies was assessed using the [kappa] statistic weighted with quadratic weights.
Results: Values of [kappa]w for the test-retest reliability study were .92 for both GOS and GOSE, and for interrater reliability study were .85 for the GOS and .84 for the GOSE.
Conclusions: The findings indicate good test-retest and interrater reliability for the structured interviews. In most circumstances a structured interview over the telephone can provide a reliable assessment of the GOS, and can safely be substituted for in person contact.
The Glasgow Outcome Scale (GOS) is a global assessment of independent living and social reintegration that is widely used as an outcome measure in brain injury research.1 The GOS does not require a detailed psychologic or neurologic examination and can be administered by professionals from different backgrounds. As an assessment of activity and participation it is suited to busy clinical settings and large scale research studies because of its simplicity and speed of use. The GOS consists of five categories: dead, vegetative state, severe disability (conscious, but disabled), moderate disability (disabled, but independent), and good recovery. For the extended GOS (GOSE) the latter three categories are divided into upper and lower bands.
Information to assign a GOS is generally collected through face-to-face contact or telephone interview with the patient, a close family member, or other individual who has good knowledge of the status of the head-injured person. The GOS has also been assigned through information gathered from postal questionnaires, through neurosurgical case notes, or a combination of sources.2-7 Despite the practical importance of the issue, there is little information about the reliability of GOS ratings with different kinds of contact. Maas and colleagues have stated that assessment of outcome on the basis of information obtained by telephone or through letters may be highly unreliable.
We have developed a structured interview for the GOS and found very good interrater reliability for this format when the scale is applied in face-to-face interviews by a psychologist and research nurses.9 However, there is a need to further investigate the reliability of the approach under different circumstances. The current study had two aims: to examine the repeatability of the GOS structured interview, and to investigate if the structured interview format can be used over the telephone without reliability being compromised. We conducted two separate studies, each using telephone interviews. In the first study patients were interviewed in person and then reinterviewed by telephone a few days later. In the second study telephone interviews were conducted first and then face-to-face interviews were carried out by a different person up to 1 month later.