Underdiagnosis of dementia in primary care: Variations in the observed prevalence and comparisons to the expected prevalence

Objectives: Dementia is a major and growing health problem. Diagnosis is an important step in the access to care, but many dementia patients remain undiagnosed. This study investigated the magnitude and variation in the difference between 'observed' and 'estimated' prevalence of dementia in general practices. We also explored practice characteristics associated with observed prevalence rates. Method: Six Primary Care Trusts (PCTs) provided data on all general practices (N = 351) in their area in terms of number of doctors, patient list size, number of patients over 65 years of age, socio-economic deprivation status of practices and number of patients on dementia registers. Results: The average observed prevalence overall of dementia amongst patients 65 years and over was 3.0% [95CI 2.8, 3.2]. The observed prevalence was 54.5% [95CI 49.2, 58.9] lower than the prevalence observed in the epidemiological studies in the UK. For an average size general practice (list size of 5269 patients) approximately 27 [95CI 22, 32] patients with dementia may remain undiagnosed. Statistically significant differences in prevalence rates were found between the different PCTs (Wald chi-square = 103.8 p < 0.001). The observed prevalence of dementia was significantly lower among practices run by one GP compared to multiple GPs (p = 0.003), and in more affluent areas (p < 0.001). Conclusion: Just under a half of the expected numbers of patients with dementia are recognised in GP dementia registers. The underdiagnosis of dementia varies with practice characteristics, socio-economic deprivation and between PCTs, which has implications for the local implementation of the National Dementia Strategy.