The cognitive drug research computerized assessment system in the evaluation of early dementia‐is speed of the essence?

The cognitive drug research computerized assessment system (COGDRAS) was evaluated in 98 unselected Hammer‐smith Hospital Memory Clinic patients (mean age 64.9, range 28‐83 years). They were divided into five groups (worried well, depressed, demented, minimally cognitively impaired and other brain disorders) on clinical assessment, Mini Mental State Examination (MMSE) and Cambridge Cognitive Examination (CAMCOG) scores. All but one patient completed the computer package, confirming its acceptability. The results of the COGDRAS in the five groups were analysed blind to diagnosis. The performance of the demented group was significantly impaired in comparison with the worried well group, showing that the COGDRAS is valid in mild dementia (mean MMSE 21.5, mean CAMCOG 73.9). The depressed group tended to perform slightly less well than the worried well, and the ‘other’ group showed a wide range of scores consistent with its diversity. The minimally impaired had scores intermediate between the demented and worried well, but an interesting speed/accuracy tradeoff was seen‐patients appeared to maintain accuracy by taking longer to perform tasks. The heterogeneity of the minimally impaired group and the role of the measurement of task completion time in the diagnosis of early dementia are discussed.