Jump to comment:
- Page navigation anchor for Cognitive tests to help diagnose dementia in symptomatic people in primary care and the communityCognitive tests to help diagnose dementia in symptomatic people in primary care and the communityWe thank Drs Larner and Fisher for their valuable and learned response to our article. We agree that a critical question for healthcare systems, commissioners, GPs and families is the issue of an acceptable false positive rate and threshold for referral. Sam Creavin would be delighted to hear the views of interested parties on this matter by email.Competing Interests: None declared.
- Page navigation anchor for Cognitive tests to help diagnose dementia in symptomatic people in primary care and the communityCognitive tests to help diagnose dementia in symptomatic people in primary care and the communityDr Creavin and colleagues question which cognitive tests might be used to identify dementia and cognitive impairment in primary care.1 As they show, the evidence base for these decisions is narrow, and the results heterogeneous. The authors’ preference for the General Practitioner Assessment of Cognition (GPCOG) is based on theoretical rather than empirical grounds.Although allusion is made to the long “menu of choices” for cognitive screening, there is no specific mention of the Six-item Cognitive Impairment Test (6CIT). Specifically designed for use in primary care, there is some (indirect) evidence that 6CIT is more frequently used than either MMSE or GPCOG. However, its negative scoring (lower score better) can be confusing, leading to scoring errors.2 An online version (www.patient.co.uk/doctor/six-item-cognitive-impairment-test-6cit) which scores the test and categorises the result as normal or significant may overcome some of the potential difficulties.Diagnostic test accuracy in primary care might be examined by using scores of cognitive screeners administered to patients who are subsequently referred to secondary care cognitive clinics, using the consensus diagnosis from the latter as reference standard and accepting inevitable selection bias in the patient cohort. Using this approach, 6CIT was found to have only modest sensitivity (>0.70), specificity (...Show MoreCompeting Interests: None declared.