The burden of Alzheimer's disease (AD) and other dementia is huge and still growing. The development of dementia/AD is a life-long process, where brain disease may start over 20 years before symptoms appear. Current drugs do not cure dementia, or stop it worsening. More focus is needed on preventing dementia in middle age. Recent results from the Finnish FINGER trial show that a prevention programme targeting physical and mental exercise, and a healthy diet, may slow down the development of dementia. The EU-funded MULTI-MODE project aims to help a European partnership develop lifestyle changes based on FINGER trial to prevent dementia. The main aim of this data analysis is check that a score used in the FINGER study to find people likely to get dementia also works in the UK population. The score will be used to test a pilot UK dementia prevention programme using a website and a mobile app. The pilot programme is a linked but separate project, which will use the outputs from this CPRD project, not CPRD data directly. We will also compare the score with a similar score recently developed in the UK, and check that no other improvements can be made to it.
This will be an open retrospective longitudinal cohort study. Data on dementia risk factors (including CVD outcomes) will be analysed. Our outcome of interest will include all participants with either a diagnosis of dementia or probable dementia based on a diagnostic algorithm including cognitive test results, prescribing and test data. The aim is to estimate the incidence and prevalence of diagnosed and probable dementia, to validate two existing risk scores, and to compare their discrimination and calibration to score developed using a broader definition of dementia and additional risk factor data, including for the first time ethnicity.
We will draw upon the TRIPOD guideline and previous CPRD validations, and using Cox regression we will apply the algorithm for each score to eligible patients in the CPRD study cohort to obtain predicted risks for each of the relevant clinical outcomes. We will then test the performance of each score in the CPRD cohort and compare it with the published results from the original validation cohorts using C- and D-statistics and an R2 statistic. The results will be utilised almost immediately by the MULTIMODE project, which will adopt the best performing score for use in identifying patients to be offered participation in the pilot programme.
Dementia diagnosis
Michael Soljak - Chief Investigator - Imperial College London
Anita Kulatilake - Collaborator - Imperial College London
Azeem Majeed - Collaborator - Imperial College London
Bang Zheng - Collaborator - Imperial College London
Bowen Su - Collaborator - Imperial College London
Lefkos Middleton - Collaborator - Imperial College London
Mahsa Mazidi - Collaborator - Imperial College London
Roger Newson - Collaborator - Imperial College London
HES Admitted Patient Care;HES Admitted Patient Care;HES Outpatient;HES Outpatient;ONS Death Registration Data;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation