Development and validation of the Glasgow Epilepsy Outcome Scale (GEOS): a new instrument for measuring concerns about epilepsy in people with mental retardation.

PURPOSE To develop a measure for use with adults with epilepsy and mental retardation, capable of assessing both clinical and care concerns and of quantifying treatment outcomes. METHODS Extensive validational and other psychometric evaluation was undertaken, comprising initial scale development work with 48 carers and 46 health practitioners, followed by formal field testing on a sample of 186 patients, using 384 respondents (160 clinicians, 141 staff, 83 family). Recognised qualitative methods were applied to identify central themes, and psychometric procedures generated data on validity, reliability, and component structure. RESULTS A total of 1,007 items of concern was generated, which was reduced systematically to a representative set of 90 items. The GEOS-90 comprises four subscales: concerns about "seizures," "treatment," "caring," and "social impact," each explaining approximately 70% of variance. Subscales and factor scales had strong internal consistency (alpha > or = 0.82). Stepwise linear regression was applied to derive a short-form version with similar structure. Thirty-five items were retained (GEOS-35; alpha > or = 0.89). Both scales discriminated moderately on clinical variables (number of seizure types, mono- vs. polytherapy, seizure frequency; all values of p < 0.05) and demonstrated concurrent validity with interview ratings from the ELDQOL (p < 0.05). CONCLUSIONS The GEOS scales appear valid and reliable for use with clinical populations of people with mental retardation.