(submitted to the British Medical Journal on the 7th July 2013 by Dr Peter J. Gordon as a rapid response to the Editorial “Too much medicine; too little care” BMJ 2013;347:f4247)
Glasziou et al’s editorial “Too much medicine; too little care” suggests we ask the questions “Does this new test detect more or earlier ‘disease’? Do we understand the course of disease in these extra cases?” This immediately brought to mind concerns I have had for some time regarding “a new touch screen test for dementia.” CANTABmobile is a computer-based tool that has been widely promoted to the NHS by the commercial company Cambridge Cognition. Along with IXICO, a commercial brain-imaging company, Cambridge Cognition received a grant from the Government-funded Biomedical Catalyst to provide an“early diagnosis service with the potential to provide a paradigm shift in diagnosis and care.”
Cambridge Cognition has, independently of IXICO, promoted CANTABmobile to primary care services across the UK. This is where my concern lies.
I am fully behind scientific innovation; however it was misleading to widely promote this test to primary care as a 5-7 minute “test for dementia.” Given this test was being “piloted” in NHS Forth Valley, and in a practice in my catchment area, I wrote to Cambridge Cognition on the 27th November 2012 outlining various concerns that I had about their promotional claim. In particular I asked for the evidence behind the headline claim that this was a “test for dementia” and whether the company envisaged any potential harms associated with it
CANTABmobile is an isolated test of Paired Associates Learning (PAL) test and measures visual episodic memory. Visual paired-learning is useful in the assessment of patients presenting with memory problems, but in isolation cannot diagnose dementia. Reading the FAQ provided by Cambridge Cognition it is rather easy to get confused and come away with the understanding that used alone CANTABmobile is a test which is sensitive for “mild to moderate Alzheimer’s disease.” Cambridge Cognition does not mention in their promotional material that only 5-10% of those who get an amber or red light on their test will progress, over the following year, to dementia. The objective scientific reality is that even after five years, at the very most, only 50% of this group will ever develop dementia. The harm caused by this overdiagnosis has been ignored, and this seams to be particularly unbalanced when Cambridge Cognition have promoted CANTABmobile as a useful tool of reassurance for the worried well.
I note that, since I wrote to Cambridge Cognition, the marketing has been changed to “a new touchscreen test for memory impairment” (see screenshots). This was not before I saw patients referred to me on the basis of the original claim.
At a recent meeting on the Timely diagnosis of dementia on the 5th June 2013 Prof Sube Banerjee has been quoted as saying on early diagnosis: that it is “crucial to end toxic uncertainty for both patient and family” but surely it is equally valid to be worried about the toxic certainty of an isolated test that is promoted misleadingly and avoids discussion of uncertainty and harm?
 Business Weekly, Cambridge Cognition in £15 million aim IPO http://www.businessweekly.co.uk/biomedtech-/15199-cambridge-cognition-in-p15m-aim-ipo
 Gordon, P. Letters sent to Cambridge Cognition on CANTABmobile and replies received http://holeousia.com/category/medical-writings/dementia/cantabmobile/
 Fox, C., Lafortune, L., Boustani, M., & Brayne, C. The pros and cons of early diagnosis in dementia. British Journal of General Practice, Volume 63, Number 612, July 2013 , pp. e510-e512(3)
 Banarjee, S. as quoted in “Timely diagnosis of dementia: a personal overview from carer Ming Ho of a meeting held at the King’s Fund, 5 June 2013 http://www.dementiauk.org/assets/files/what_we_do/uniting_carers/11_june_amendments_timely_diagnosis_of_dementia_5_june_2013_meeting_report.pdf
 Strech, D. et al The full spectrum of ethical issues in dementia care: systematic qualitative review. BJP June 2013 202:400–406