Early recognition of dementia with 67 per cent of cases diagnosed is an aim of NHS England. To achieve this, incentives, education and specialist clinics have been introduced to promote dementia recognition and diagnosis in primary care.

However, the process of diagnosis is often delayed, protracted or complicated and a significant proportion of people do not appear to receive diagnosis. Understanding the factors underlying this using robust research data could help us improve people’s access to timely diagnosis. In order to determine the true prevalence of dementia or correlates of dementia not known to primary care it is necessary for diagnoses made in primary care to be linked to independent and objective dementia diagnoses at the patient level.

We are working with researchers from the Cognitive Function and Ageing Study II (CFAS II; www.cfas.ac.uk), a nationally representative study of dementia led by Prof Carol Brayne at the University of Cambridge.  CFAS II provides this independent and objective dementia assessment to which we are adding key data on primary care diagnoses enabling us to address the questions below.

Key Research Questions

  • What was the prevalence of undiagnosed dementia between 2008 and 2012? How many undiagnosed people with dementia had a GP record of cognitive complaint/impairment, how many had been referred to specialist services and how many had no record of any cognitive impairment
  • To what extent do general practice dementia registers mandated by the quality outcomes framework (QOF) reflect GPs records of cognitive impairment among their patients? In other words, how many people with dementia are known to GPs as having a cognitive impairment but are not recorded on dementia registers
  • What are the social and clinical characteristics of the group without a study diagnosis of dementia but with a dementia diagnosis?
  • What are the associations between social and clinical characteristics of a person with dementia and their risk of being undiagnosed? Hence what are the independent predictors of help seeking and diagnosis among people within dementia?
  • What is the incidence of new diagnosis among the population with undiagnosed dementia?
  • If there are sufficient follow-up assessment data for undiagnosed people we will also try to explore the question: How are clinical, psychosocial and healthcare utilisation trajectories over a two year follow-up period associated with diagnosis status among people with dementia? Hence we will provide an indication of the benefits and harms of diagnosis.

Research Design

The Cognitive Function and Ageing Study II (CFAS II) Dementia Diagnosis Study (CADDY) which aims to determine prevalence, causes and consequences of undiagnosed dementia is a sub-study of CFAS II. During 2008-2011 CFAS II recruited a population-representative sample of 7500 people aged 65 and older from three geographical areas of England. Knowledgeable informants were also recruited for a subsample of participants. Participants completed detailed cognitive, health and socioeconomic assessments. They were conducted by trained researchers at baseline and at two-year follow up assessment. For each participant a study dementia status, previously validated against DMS-III-R, was established. 

Under CADDY, primary care records were sought for 598 CFAS II participants consenting to health service data linkage (445 with a CFAS II diagnosis of dementia and 153 without dementia but weighted towards those with cognitive impairment). Dementia diagnoses, cognitive complaint and referrals to specialist services were extracted from primary care records and these data were linked with CFASII data. 

Linking CFASII assessments and diagnoses with records held in primary care will provide a rare opportunity to estimate: (i) the true proportion of people with dementia with a formal diagnosis; (ii) social and clinical predictors of becoming diagnosed; and (iii) outcomes among diagnosed and non-diagnosed populations. Hypothesised predictors include social circumstances, severity of cognitive and behavioural symptoms, comorbidities and service use outcomes.

More than 90% of the requested primary care records have been returned. We will determine the prevalence of diagnosis within primary care with respect to dementia severity, recorded help-seeking or referral to secondary services, and compare the characteristics and levels of service use for diagnosed and undiagnosed people with dementia.

The Research Team

Principal Investigator: Clare Aldus (University of East Anglia)

Co-Investigators: Prof Antony Arthur (University of East Anglia); Prof Carol Brayne (University of Cambridge); Prof Tom Dening (University of Nottingham); Prof Chris Fox (University of East Anglia); Prof Fiona Matthews (Newcastle University); Prof Louise Robinson (Newcastle University); Dr Blossom Stephan (Newcastle University); Linda Barnes (University of Cambridge); and Dr George Savva (Quadram Institute)

Additional Support & Guidance: Judy Henwood (Sponsor Representative); Katy Blakely (Independent Local Commissioner); Gavin Terry (Policy Section, Alzheimer's Society); and Clare Symms (Sponsor Representative)

The CADDY PPI Advisory Group comprises 12 people across inspire, PPIRes & Alzheimer's Society.

The study is sponsored by NHS South Norfolk CCG and supported by Norwich Clinical Trials Unit.


The project was awarded £353,923 by National Institute for Health Research (NIHR). The study commenced on 1 January 2016 and will end on 31 December 2017. The final report is estimated to be published in July 2018. CADDY is registered with the NHS Central Portfolio Management System (CPMS 530655). Full approval for this research was granted 6 May 2016.

This is independent research funded by the NIHR Health Services and Delivery Research (HS&DR) programme (14/70/96; Dementia undetected or undiagnosed in primary care: the prevalence, causes and consequences). The views expressed here are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.