Skip to main content Accessibility help
×
Home

Information:

  • Access
  • Cited by 4

Actions:

      • Send article to Kindle

        To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Variations in dementia diagnosis in England and association with general practice characteristics
        Available formats
        ×

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Variations in dementia diagnosis in England and association with general practice characteristics
        Available formats
        ×

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Variations in dementia diagnosis in England and association with general practice characteristics
        Available formats
        ×
Export citation

Abstract

Objectives

Improving dementia diagnosis rates in England has been a key strategic aim of the UK Government but the variation and low diagnosis rates are poorly understood. The aim of this study was to explore the variation in actual versus expected diagnosis of dementia across England, and how these variations were associated with general practice characteristics.

Method

A cross-sectional, ecological study design using secondary data sources and median regression modelling was used. Data from the year 2011 for 7711 of the GP practices in England (92.7%). Associations of dementia diagnosis rates (%) per practice, calculated using National Health Service England’s ‘Dementia Prevalence Calculator’ and various practice characteristics were explored using a regression model.

Results

The median dementia diagnosis rate was 41.6% and the interquartile range was 31.2–53.9%. Multivariable regression analysis demonstrated positive associations between dementia diagnosis rates and deprivation of the population, overall Quality and Outcomes Framework performance, type of primary care contract and size of practice list. Negative associations were found between dementia diagnosis rates and average experience of GPs in the practice and the proportion of the practice caseload over 65 years old.

Conclusion

Dementia diagnosis rates vary greatly across GP practices in England. This study has found independent associations between dementia diagnosis rates and a number of patient and practice characteristics. Consideration of these factors locally may provide targets for case-finding interventions and so facilitate timely diagnosis.

Introduction

Current National Health Service (NHS) policy promotes the potential benefits of early identification of dementia and the appropriate treatment and support to those who experience this condition (Department of Health, 2009). However, in 2012, NHS England estimated that only 47% of people with dementia had been diagnosed (Department of Health, 2012a). NHS England has been aiming to improve the diagnosis rates for dementia, following the Prime Minister’s Challenge to achieve two thirds of people with dementia being diagnosed by April 2015 (Department of Health, 2012b). ‘Diagnosis rate’ has been used as a term by the UK Government to describe the proportion of people predicted to have dementia in the population, who actually have a recorded diagnosis on electronic medical records in primary care.

Diagnosing dementia in primary care is not straightforward and requires the recognition of a syndrome which is challenging (Wilcock et al., 2015). Large variation in the diagnosis rates has been reported, with some known factors associated with dementia diagnosis. Estimates of diagnosis rates indicated they ranged from 39 to 75% across areas of England in 2012 (Department of Health, 2013). One previous study (Connolly et al., 2011) explored dementia diagnosis rates in Greater Manchester, England. The study found that singlehanded practices and practices in more affluent areas had significantly lower rates of dementia diagnosis.

Variation in diagnosis rates has been more extensively explored for other diseases, such as coronary heart disease, cancer and hepatitis. Several characteristics of general practices have been found to be associated with variations in diagnosis including deprivation of the practice population (Saxena et al., 2007; Bottle et al., 2012), singlehanded practices (Coupland et al., 2006), practice size (Saxena et al., 2007), size of caseload (Saxena et al., 2007), patient access (Bottle et al., 2012), doctor characteristics (Coupland et al., 2006; Bottle et al., 2012), Quality and Outcomes Framework (QOF) performance (Bottle et al., 2012) and financial factors related to GP contracts (Morgan and Beerstecher, 2006). However, similar associations with dementia diagnosis have not previously been explored in the same way, other than the Greater Manchester study. We chose variables within our study based on these findings from studies in other diseases.

Using publically available data from across England we explored the association between dementia diagnosis rates of GP practices with a range of patient and practice characteristics to identify possible targets for intervention.

Methods

Study design and sample

We used a cross-sectional, ecological study design that sourced publicly available data from various sources for the year 2011 across all GP practices in England. Our outcome was the reported number of patients with a recorded dementia diagnosis per practice as a proportion of expected diagnosis to reflect the UK Government definition of ‘diagnosis rate’. Reported dementia diagnosis was taken from primary care disease registers of the QOF. Expected dementia diagnosis was calculated using the prevalence calculator commissioned and used by NHS England (NHS England, 2013). The dementia prevalence estimates used in the calculator model are taken from an expert panel commissioned by the Alzheimer’s Society, United Kingdom (Knapp et al., 2007). The model has been constructed using estimates of the prevalence of dementia associated with the age and gender profiles of the GP populations and the proportion of people in residential care homes.

We chose the independent variables to explore their association with dementia diagnosis rates that had been previously identified in the diagnosis of other diseases in England. The type of primary care contract was taken from the Health and Social Care Information Centre (HSCIC) (General Medical Services contract (GMS) is a standard national contract, Personal Medical Services contract (PMS) and Alternative Provider Medical Services contract (APMS) contracts are individually negotiated by practices and local funders and Primary Care Trust Medical Services contract (PCTMS) is a standard localised contract). Practice list size, overall QOF score and GP experience data were also taken from HSCIC. Index of Multiple Deprivation (IMD) scores for each practice were sourced from Public Health England data. The IMD estimates an average deprivation score proportionally across a practice population. Patient-reported access measures were taken from the National General Practice Patient Survey, which is weighted to reflect the demographics of the practice population. Further details of each variable and data source are provided in Table 1.

Table 1 Data sources of patient and practice variables

NHS=National Health Service; GMS=General Medical Services contract; PMS=Personal Medical Services contract; APMS=Alternative Provider Medical Services contract; PCTMS=Primary Care Trust Medical Services contract.

Statistical analysis

As our outcome measure, dementia diagnosis rates, was positively skewed we used median (quantile) regression models to explore associations between diagnosis rates and selected variables (Kroenker and Hallock, 2001; Katz, 2011). We performed univariate analysis and included any variables with statistically significant associations (P<0.05) into a multivariable model.

We tested for correlation between factors within the multivariable model and linearity in continuous variables. None of the correlations exceeded our threshold of >0.7 and P<0.05 and so no interaction terms were included in the multivariable model. We found that only ‘QOF performance’ did not have a linear relationship with diagnosis rate and so this variable was included in the model as quintiles. All data were collated and analysed using STATA version 12.

Results

Full data were available for 7711 general practices representing 92.7% of the practices in England in 2011 (n=8316). Practices excluded due to incomplete data were more likely to be smaller practices than those included for analysis (mean patient list size 2595 versus 6685). There was some geographical variation in excluded practices, for instance over 10% of London practices were excluded compared with 4% of practices in the South West region of England.

We found wide variation in reported dementia diagnosis rates across the 7711 practices included in this study. The median dementia diagnosis rate was 41.6% of expected rate, and interquartile range was 31.2–53.9%.

All variables except patient-reported access measures were found to be significantly associated with diagnosis rate in univariate analysis (Table 2) and so were included in the multivariable median regression model.

Table 2 Univariate and multivariable regression analysis of primary care practice characteristics and difference in dementia diagnosis rates

QOF=Quality and Outcomes Framework; GMS=General Medical Services contract; PMS=Personal Medical Services contract; APMS - Alternative Provider Medical Services contract; PCTMS=Primary Care Trust Medical Services contract.

Patient factors associated with dementia diagnosis rates included age and deprivation. Each percentage point increase in the proportion of the practice population over 65 years of age was associated with a 0.3% reduction in dementia diagnosis rate [−0.3%; 95% confidence interval (95% CI) −0.4 to −0.2], on adjustment for other factors. Therefore where the proportion of practices list over 65 differ by 10%, a difference of 3% would be expected in diagnosis rates. Diagnosis rates of dementia were 8% higher in the most deprived quintile of practices, compared to the least deprived (8.2%; 95% CI 6.3–10.0).

Practice factors associated with dementia diagnosis rates included practice size, average GP experience, contract type and QOF score attainment. Compared with the average diagnosis rate in the smallest practices (list size 0–2999 patients), the average diagnosis rate of practices between 8000 and 9999 list size was 4.3% higher (95% CI 2.3–6.4) and the average diagnosis rate of practices over 10 000 list size was 3.5% higher (95% CI 1.7–5.4).

Average GP experience per practice was negatively associated with diagnosis rates. When the average experience increases by a decade, the diagnosis rate in practices decreases by 2.1% (95% CI −2.9 to −1.3).

Compared with diagnosis rates of practices with GMS contracts, practices with PMS contracts had a diagnosis rate that was 1.8% higher (95% CI 0.7–2.9). Practices with other types of contract (PCTMS and APMS) had a diagnosis rate that was 2.4% higher (95% CI −0.9 to 5.7), although there were relatively small numbers of practices in these categories, and this association was not statistically significant in the multivariable model.

Diagnosis rate was higher with increasing QOF performance. The highest quintile of practices by QOF performance had a diagnosis rate that was 5.3% higher (95% CI 3.6–7.1) than the lowest quintile.

As we were concerned about the potential impact of outliers, we tested the effect of outliers on our results. We excluded practices beyond three standard deviations (n=82) and beyond two standard deviations (n=174) and neither the coefficients nor overall factors associated with diagnosis rate were substantially altered.

Discussion

We found patient demographics; age and deprivation, and GP characteristics; experience, list size and QOF performance are associated with variation in dementia diagnosis.

Our results show that practices in more deprived areas have a higher diagnosis rate, confirming findings of a previous study (Connolly et al., 2011). This consistent finding suggests an amendment to the NHS England prevalence model calculator may be appropriate. Dementia prevalence and deprivation is a complex relationship. Several studies have found an association between dementia prevalence and higher levels of deprivation (Wilson et al., 1999; Versporten et al., 2005; Basta et al., 2008; De Deyn et al., 2011). This may be explained by higher cardiovascular risk factors (Banerjee, 2013) and the cognitive reserve hypothesis (Brayne et al., 2010; Meng and D’Arcy, 2012; Matthews et al., 2013) where the protective influence of active cognitive function delays the clinical presentation of dementia. Conversely less deprived populations are more likely to live into old age and therefore be at a higher risk of developing dementia, increasing the prevalence. Further exploration of this complex relationship between deprivation and dementia diagnosis rates may assist in improving the prevalence model and in identifying population-based interventions.

Fewer patients with dementia are being diagnosed in smaller practices and in practices with larger proportions of elderly patients. These results suggest that practices with a larger proportion of older patients may not be successfully identifying cases of dementia in amongst the natural deterioration in cognitive functioning of old age. In addition, it may not confer clinical benefit to record dementia diagnosis for patients already in care homes which may explain this association. These factors do, however, provide an opportunity to target resources in these practices to optimise case finding.

A cohort effect could explain practices with older and more experienced GPs less likely to record dementia in electronic medical records. Larger list size may increase documentation of dementia due to increased clinical specialisation in larger practices. The association found with PMS contract type may indicate innovative and proactive approaches exist within these GP practices relating to both contractual and clinical processes that enhance dementia recording. The lack of association with patient access variables is consistent with findings from a study of educational interventions that found patients and their carers had minimal contact with their GP practice (Downs et al., 2006).

This study has a number of limitations. First, despite the statistical significance of the associations identified in this study causal relationships cannot be established within this cross-sectional design. Second, we have relied heavily on the dementia prevalence model developed by NHS England to explore dementia diagnosis rates. There are a number of assumptions built into this model that may reduce its accuracy in predicting dementia in the population. The associations we identify could relate more to the prevalence model than to the diagnosing of dementia in GP practices which may well over-estimate prevalence (Matthews et al., 2013). Third, although we included data from the majority of GP practices in England (92.8%) some systematic differences between the practices included and excluded do exist, such as list size and geographical variations. The potential for these and other unknown structural factors to have an impact on our results cannot be ruled out. Fourth, data coding in clinical systems for dementia is variable and in some cases quite poor (Russell et al., 2013). Our study analyses the number of patients who have a recorded diagnosis of dementia on GP electronic patient records. We cannot demonstrate how reliable this coding is. There may be structural differences in coding practice that could affect these results and the way in which dementia diagnosis was recorded.

Rurality of the practice population may also influence access to the practice. However data on the rurality of GP practices are based on the location of the practice rather than the patients registered with that practice and were therefore not included.

Practice size, age profile, deprivation, QOF performance and GP contract type are all associated with dementia diagnosis. This wide variation we found in dementia diagnosis rates in English GP practices is reflected internationally. Debate about the variation in the detection of dementia in primary care across Europe (De Lepeleire et al., 2008) has highlighted the need to better understand the factors associated with this variation in order to incorporate them into systematic case-finding strategies. This study adds to our understanding of this complex process in England. The identification of specific patient factors associated with variations in dementia diagnosis provide potential population level targets for local government and public health organisations (eg, Public Health England) to identify patient groups at risk of dementia. Specific practice factors could be used by commissioners (eg, Clinical Commissioning Groups) to engage with practices where the case finding is lower than expected.

Since 2011, there have been a variety of national targets and incentives to improve dementia diagnosis rates in the United Kingdom. However, these have been widespread schemes, and do not appear to have considered some of the identified factors associated with dementia diagnosis in this study. Our analysis may provide a benchmark level of diagnosis rate by which the recent schemes can be measured both in the United Kingdom and internationally.

Acknowledgements

No funding was received for this research. No conflicts of interest to declare. The authors are grateful to NHS England for giving permission to use data from their Dementia Prevalence Calculator for this study.

References

Banerjee, S. 2013: Good news on dementia prevalence – we can make a difference. The Lancet 382, 13841386.
Basta, N.E., Matthews, F.E., Chatfield, M.D., Brayne, C. and Mrc, C. 2008: Community-level socio-economic status and cognitive and functional impairment in the older population. Eur Journal of Public Health 18, 4854.
Bottle, A., Tsang, C., Parsons, C., Majeed, A., Soljak, M. and Aylin, P. 2012: Association between patient and general practice characteristics and unplanned first-time admissions for cancer: observational study. British Journal of Cancer 107, 12131219.
Brayne, C., Ince, P.G., Keage, H.A., Mckeith, I.G., Matthews, F.E., Polvikoski, T. and Sulkava, R. 2010: Education, the brain and dementia: neuroprotection or compensation? Brain 133, 22102216.
Connolly, A., Gaehl, E., Martin, H., Morris, J. and Purandare, N. 2011: Underdiagnosis of dementia in primary care: variations in the observed prevalence and comparisons to the expected prevalence. Aging and Mental Health 15, 978984.
Coupland, C., Hippisley-Cox, J., Smith, S., Irving, W., Pringle, M., Ryder, S., Neal, K., Cater, R., Thomson, B., Pugh, S., Bicknell, M. and Bullock, D. 2006: General practice characteristics associated with rates of testing and detection of hepatitis C: cross-sectional study in Nottingham and Derbyshire. British Journal of General Practice 56, 620623.
De Deyn, P.P., Goeman, J., Vervaet, A., Dourcy-Belle-Rose, B., Van Dam, D. and Geerts, E. 2011: Prevalence and incidence of dementia among 75-80-year-old community-dwelling elderly in different districts of Antwerp, Belgium: the Antwerp Cognition (ANCOG) Study. Clinical Neurology and Neurosurgery 113, 736745.
De Lepeleire, J., Wind, A., Iliffe, S., Moniz-Cook, E., Wilcock, J., Gonzalez, V., Derksen, E., Gianelli, M. and Vernooij-Dassen, M. 2008: The primary care diagnosis of dementia in Europe: an analysis using multidisciplinary, multinational expert groups. Aging and Mental Health 12, 568576.
Department of Health 2009. Living well with dementia: a national dementia strategy. London: Department of Health.
Department of Health 2012a. The Mandate – a mandate from the Government to the NHS Commissioning Board: April 2013 to March 2015. London: Department of Health.
Department of Health 2012b. Prime Minister’s challenge on dementia: delivering major improvements in dementia care and research by 2015. London: Department of Health.
Department of Health 2013. Dementia: a state of the nation report on dementia care and support in England. London: Department of Health.
Downs, M., Turner, S., Bryans, M., Wilcock, J., Keady, J., Levin, E., O’Carroll, R., Howie, K. and Iliffe, S. 2006: Effectiveness of educational interventions in improving detection and management of dementia in primary care: cluster randomised controlled study. British Medical Journal 332, 692696.
Katz, M.H. 2011. Multivariable analysis: a practical guide for clinicians and public health researchers. Cambridge: Cambridge University Press.
Knapp, M., Prince, M., Albanese, E., Banerjee, S., Dhanasiri, S., Fernandez, J., Ferri, C., Snell, T. and Stewart, R. 2007: Dementia UK: report to the Alzheimer’s Society. Kings College London and London School of Economics and Political Science.
Kroenker, R. and Hallock, K.F. 2001: Quantile regressions. Journal of Economic Perspectives 15, 143156.
Matthews, F.E., Arthur, A., Barnes, L.E., Bond, J., Jagger, C., Robinson, L. and Brayne, C., Medical Research Council Cognitive Function and Ageing Collaboration 2013: A two-decade comparison of prevalence of dementia in individuals aged 65 years and older from three geographical areas of England: results of the Cognitive Function and Ageing Study I and II. Lancet 382, 14051412.
Meng, X. and D’Arcy, C. 2012: Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta-analyses and qualitative analyses. PLoS One 7, e38268.
Morgan, C.L. and Beerstecher, H.J. 2006: Primary care funding, contract status, and outcomes: an observational study. British Journal General Practice 56, 825829.
NHS England 2013. Dementia prevalence calculator: introduction to the online tool for improving dementia diagnosis and diagnosis pathways. London: NHS England.
Russell, P., Banerjee, S., Watt, J., Adleman, R., Agoe, B., Burnie, N., Carefull, A., Chandan, K., Constable, D., Daniels, M., Davies, D., Deshmukh, S., Huddart, M., Jabin, A., Jarrett, P., King, J., Koch, T., Kumar, S., Lees, S., Mir, S., Naidoo, D., Nyame, S., Sasae, R., Sharma, T., Thormod, C., Vedavanam, K., Wilton, A. and Flaherty, B. 2013: Improving the identification of people with dementia in primary care: evaluation of the impact of primary care dementia coding guidance on identified prevalence. British Medical Journal Open 3, e004023.
Saxena, S., Car, J., Eldred, D., Soljak, M. and Majeed, A. 2007: Practice size, caseload, deprivation and quality of care of patients with coronary heart disease, hypertension and stroke in primary care: national cross-sectional study. BMC Health Services Research 7, 96.
Versporten, A., Bossuyt, N., Meulenbergs, L., Baro, F. and Van Oyen, H. 2005: The incidence of dementia: relationship with educational attainment. Archives of Public Health 63, 279292.
Wilcock, J., Jain, P., Griffin, M., Thune-Boyle, I., Lefford, F., Rapp, D. and Iliffe, S. 2015: Diagnosis and management of dementia in family practice. Aging and Mental Health 20, 18.
Wilson, K.C., Chen, R., Taylor, S., Mccracken, C.F. and Copeland, J.R. 1999: Socio-economic deprivation and the prevalence and prediction of depression in older community residents. The MRC-ALPHA Study. British Journal of Psychiatry 175, 549553.