Relation between Glasgow outcome score extended (GOSE) and the EQ-5D health status questionnaire after head injury
In this Journal Wilsonet al investigated aspects of the validity of the Glasgow outcome score (GOS) and the extended form (GOSE), particularly the relation between the GOS and subjective reports of health status.1 A potentially useful relation between the GOSE, measured using a standard interview at 1 month after head injury,2 and health status assessed using the Euroqol EQ-5D is now reported. The GOS is the most widely used method to describe overall outcome after head injury.3 The EQ-5D questionnaire is a validated tool that measures health status in five quality of life dimensions,4 with the most appropriate of three statements in each of these selected by the respondent. These responses are coded to give a single score. A visual analogue score of current health state is also recorded.
Local research ethics committee approval was granted to measure the GOSE on, and to administer the EQ-5D questionnaire to, patients attending the emergency departments of four hospitals in Manchester (United Kingdom). All patients aged over 18 presenting with a head injury of any initial severity (Glasgow coma score 3–15) within 6 hours of injury were eligible for enrolment. Patients with extracranial injuries or preinjury morbidity were not excluded. Follow up occurred at 1 month. The GOSE was measured by telephone interview and the EQ-5D by postal survey. One hundred and twenty one patients were recruited (76 men), of whom 100 (83%) had outcome assigned by GOSE, and 67 (55%) returned the EQ-5D questionnaire. Four of the second group omitted the visual analogue score assessment. The EQ-5D questionnaire only was assessed in nine patients (7%). Fifty eight patients underwent outcome assignment by both methods, and these results form the basis of subsequent statistical analysis.
A predominance of apparently minor head injury (initial Glasgow coma score (GCS) 14–15) was found in the study respondents (56/58 (97%)). The proportion of those 56 patients with more minor injuries who had less than a good recovery (GOSE<7) at one month was 23%. The two patients with GCS less than 14 experienced a good recovery.
A significant correlation between the EQ-5D score and GOSE was found (Spearman's r 0.427 p=0.001). The correlation between the visual analogue score reading and GOSE was also significant (Spearman's r 0.38 p=0.005). All 18 patients who replied “no problem” in each health state category of the EQ-5D questionnaire had a good recovery as measured by GOSE (>6) (Pearson's χ2 7.54 p=0.006).
The follow up rate for the GOSE, assessed by telephone interview, was superior to the EQ-5D postal survey. A few patients, however, responded only by post. In an unselected group of patients with predominantly minor head injuries it seems that EQ-5D health status score can be used to infer a good recovery after head injury if no problems are identified in the five domains of health status.
