Identifying people with dementia for epidemiological analysis in electronic primary care records in the UK

Kang, Sujin (2018) Identifying people with dementia for epidemiological analysis in electronic primary care records in the UK. MRes thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only until 13 December 2020. Subsequently available to Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (3MB)


There are many epidemiological questions that need to be answered regarding dementia, including the causes, prognosis, comorbidities, and treatment of the condition and complications. Routine data sources including a large electronic primary care dataset provide a way of investigating some of the complicated factors with rich information. However, there has been concern regarding whether dementia is recorded well enough. It has been reported that dementia is often poorly recorded in primary care and so it would be expected to be potentially difficult to identify people with dementia. Before the data sources can be used satisfactorily, it will be important to understand how dementia is recorded, and whether the records on dementia, related symptoms or use of related medications are the most appropriate ways of identifying people with dementia.

This study has therefore explored this by conducting a systematic review to understand how dementia has been identified previously in primary care databases in the UK, and added to this by exploring additional terms and symptoms, and medications that might be helpful in identifying people with dementia. The codes from the literature were not very consistent, taking into account the differing study objectives and populations.

The incidence rates from the longitudinal study which used reliable measurement techniques to identify dementia with a verified process and a recruitment strategy were higher than the rates from the study using primary care databases in the UK. This implies the method of dementia diagnosis may influence the rates. To identify people with dementia in primary care records, a careful alternative identification method for the study population seems to be required.

This study estimated and compared the prevalence and incidence of dementia in the UK using the alternative ways of identifying people with dementia in The Health Improvement Network (THIN) database. The method of comparison of the estimated rates with other studies was employed using the comprehensive list of diagnostic codes.

Since the Quality and Outcomes Framework (QOF) was introduced in 2004, there were slight and steady increasing trends in both prevalence and incidence rates up to a recent date. The estimated incidence rates per 1,000 person-years for the 60+ age group who had any of the first diagnosis among the QOF defined codes, Other diagnoses, Dementia symptoms and Prescribed medication were 2.5, 4.7, and 15.9 in 1995, 2004, and 2015 respectively. The estimated prevalence rates per 1,000 persons for those aged 60+ who had any of the first diagnosis among the QOF defined codes, Other diagnoses and Prescribed medication were 2.8, 3.2, and 10.2 in 1995, 2004, and 2015 respectively. The prevalence rates with those codes including the Dementia symptoms were 3.0, 4.3 and 19.4 in 1995, 2004 and 2015 respectively.

At least, using of the other dementia diagnostic codes (Other diagnoses) in addition to the QOF defined ones, (and treatments (Prescribed medication) for Alzheimer's disease or various other memory impairments for verification) was evidenced that will not missing out a number of people with dementia, in the comparison of the incidence and prevalence rates of dementia with the other estimates. The codes related to dementia symptoms (including memory loss/forgets, cognitive decline/impairment) seem to cover a broad definition of dementia or pre-existing dementia population in primary care records in the UK which need to be considered furthermore for study designs in epidemiology. They seem to identify the very elderly population with dementia, implying the codes related to dementia symptoms may cover important cases or the related information of dementia in primary care records. Nevertheless, the codes about non-specific or temporary symptoms may need to be used carefully as it is possible for memory or cognitive function to be impaired temporarily due to stress or medications.

Item Type: Thesis (University of Nottingham only) (MRes)
Supervisors: Lewis, Sarah
Gladman, John
Wardt, Veronika
Gibson, Jack
Keywords: Dementia, Diagnosis of dementia, Primary care records, Codes to identify dementia, Prescribed medication, Incidence and prevalence of dementia, Hypertension, Study designs
Subjects: W Medicine and related subjects (NLM Classification) > WM Psychiatry
W Medicine and related subjects (NLM Classification) > WT Geriatrics. Chronic disease
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 55273
Depositing User: Kang, Su
Date Deposited: 24 Jan 2019 08:55
Last Modified: 04 Mar 2019 16:33

Actions (Archive Staff Only)

Edit View Edit View