Reliability of a telephone-based Glasgow Outcome Scale assessment using a structured interview in a heterogenous population of patients and examiners.

A reliable telephone-based Glasgow Outcome Scale (GOS) assessment would be advantageous to both patients and investigators. Using a previously published structured GOS interview and scoring system, the aim of this study was to assess the reliability of telephone-based GOS scores compared to those obtained face-to-face in a heterogenous population of patients and examiners. Sixty-six patients hospitalized for a variety of acute neurological injuries underwent two GOS interviews approximately 90 days after injury. From a pool of six examiners, structured interviews were conducted using a standardized data form containing nine yes/no questions. One interview was conducted face-to-face and the other interview was conducted by telephone, the order being randomly selected. A different examiner conducted the second interview, and was unaware of the findings of the first interview. From this data, a separate investigator assigned GOS scores using standardized criteria. Concordant GOS scores were obtained in 71% (47/66) of patients and discordant scores in 29% (19/66); kappa = 0.56 +/- 0.08 (SE) (95% CI kappa = 0.40-0.73). Patient-, examiner-, and interview-related characteristics had no significant associations with GOS concordance, although patient sex had a significant association with discrepant responses to one specific question (work at previous capacity). When used by multiple examiners to assess patients with diverse neurological conditions, use of a structured GOS examination does not guarantee a reliable telephone-based GOS score. Determination of whether patient sex influences the validity of the structured face-to-face GOS interview is worthy of future study.