From multidimensional neuropsychological outcomes to a cognitive complication rate: The International Subarachnoid Aneurysm Trial

Abstract

Abstract Background The neuropsychological arm of the International Subarachnoid Aneurysm Trial (N-ISAT) evaluated the cognitive outcome of 573 patients at 12 months following subarachnoid haemorrhage (SAH). The assessment included 29 psychometric measures, yielding a substantial and complex body of data. We have explored alternative and optimal methodologies for analysing and summarising these data to enable the estimation of a cognitive complication rate (CCR). Any differences in cognitive outcome between the two arms of the trial are not however reported here. Methods All individual test scores were transformed into z-scores and a 5th percentile cut-off for impairment was established. A principal components analysis (PCA) was applied to these data to mathematically transform correlated test scores into a smaller number of uncorrelated principal components, or cognitive 'domains'. These domains formed the basis for grouping and weighting individual patients' impaired scores on individual measures. In order to increase the sample size, a series of methods for handling missing data were applied. Results We estimated a 34.1% CCR in all those patients seen face-to-face, rising to 37.4% CCR with the inclusion of patients who were unable to attend assessment for reason related to the index SAH. This group demonstrated significantly more self and carer/relative rated disability on a Health Related Quality of Life questionnaire, than patients classified as having no functionally significant cognitive deficits. Conclusion Evaluating neuropsychological outcome in a large RCT involves unique methodological and organizational challenges. We have demonstrated how these problems may be addressed by re-classifying interval data from 29 measures into a dichotomous CCR. We have presented a 'sliding scale' of undifferentiated individual cognitive impairments, and then on the basis of PCA-derived cognitive 'domains', included consideration of the distribution of impairments in these terms. In order to maximize sample size we have suggested ways for patients who did not complete the entire protocol to be included in the overall CCR. ISAT trial registration ISRCTN49866681