Assessing outcome in head injury trials

It is often suggested that one reason that neuroprotection trials in head injury have not shown benefit is the insensitivity of the outcome measures employed. The review considers strengths and weaknesses of the main approaches to assessing outcome in head injury trials. Determination of a response to neuroprotection requires a measure that is both sensitive to differences in outcome and is also influenced by brain injury. Sets of tests have been proposed for use in clinical trials that include scales of disability and handicap and measures of neuropsychological impairment. The Glasgow Outcome Scale (GOS) covers the whole spectrum of outcome after head injury and is the most popular primary endpoint for head injury trials. Although the GOS is a simple scale, small changes in the distribution of outcome can be detected with appropriate sample sizes. Late outcome in survivors can be influenced by various personal and environmental factors in addition to initial brain pathology. Other rating scales and questionnaires do not appear to offer significant advantages over the GOS, but can be used to supplement the information collected. Neuropsychological tests have properties that make them attractive as outcome measures, although there are practical difficulties with using tests. Neuropsychological assessment is potentially a powerful way of testing specific hypotheses concerning the effects of treatment. All current outcome measures have limitations; choice of the most appropriate endpoint will depend both on the properties of the measures available, and on the anticipated effects of treatment.