This site is intended for health professionals only

GP review of dementia coding ‘could help close diagnosis gap’

Practices are able to substantially increase the number of patients on their lists recorded as having dementia simply by ‘cleaning up their records’, shows a new study that researchers claim highlights uncertainties over the Government’s promotion of GP case-finding to boost diagnoses.

The new study – published in BMJ Open last month – reveals that a review of practice records that would take one GP less than a day could raise the current national estimated diagnosis rate from 46% to nearly 55%.

This increase would provide a welcome boost to the Government’s national strategy on dementia, which has set a target of having two-thirds of people with dementia formally diagnosed by next year.

The study involved 23 practices across 19 boroughs in London, where GPs used a coding exercise protocol to review their practices’ medical records and update the QOF dementia register.

As a result, the overall number of patients on the register rose significantly, from 1,007 to 1,139 cases, meaning the identified prevalence of dementia among patients aged 65 and over went up from 5.1% to 5.8%. Based on an expected prevalence among Londoners this age of 8.1%, the researchers calculated that this would raise the proportion of people with dementia who are formally diagnosed by 8.8 percentage points, from 63% to 71.8%.

Furthermore, they said that applying the process on a national scale could raise the current estimated diagnosis rate from 46% to 54.8%.

The researchers found the miscoding of dementia was largely down to overly complex coding systems or unclear information about the diagnosis from secondary care, rather than because GPs had deliberately avoided ‘labelling’ patients with dementia because of uncertainty about the diagnosis.

They suggested the approach could offer ‘a simple and cheap’ approach to ‘closing the diagnosis gap’ that GPs have been asked to tackle through a controversial directed enhanced service (DES) to opportunistically screen at-risk patients for potential memory problems introduced last year.

Critics have argued the diagnosis of dementia is complex and the DES risks over-diagnosing large numbers of patients, while the estimated ‘true’ prevalence of dementia in the UK has recently been revised downwards.

Led by Professor Sube Banerjee, professor of dementia at Brighton and Sussex Medical School, the researchers found the average time a GP took to review each practice’s list was 4.7 hours. The authors conceded the GPs involved were ‘highly motivated’ participants in the NHS London dementia fellowship, who would likely be able to complete the recoding exercise more quickly than the average GP, but suggested the longest it would take would be a day of a GP’s time.

They concluded: ‘The results presented here suggest that completing this exercise could provide a simple, cheap and useful first step to improve accuracy of records. More accurate information can help to improve the management of patients and also help to close the diagnosis gap.’

Experts said this latest study highlighted uncertainties over the alleged degree of under-recognition of dementia by GPs and the continued emphasis on driving up diagnosis rates.

Dr Chris Fox, clinical senior lecturer at the University of East Anglia and a specialist in dementia management, said the findings highlighted the ‘unnecessary diversion of resources into increasing diagnosis rates and away from the development of services for patients diagnosed with dementia’ and argued that carrying out the exercise could exacerbate this problem.

Dr Fox said: ‘The cost of 4.7 hours per practice is unlikely to have a large impact but the extra cost associated with increased QOF payments could be significant if scaled nationally.’

Deputy GPC chair Dr Richard Vautrey said: ‘The study offers one explanation for a degree of the gap in diagnosis with dementia – but coding issues would similarly account for apparent underdiagnosis of other conditions, such as COPD and diabetes.

‘The issue is whether these patients have not been treated appropriately and that certainly does not seem to be the case.’

Dr Vautrey added that carrying out coding reviews would inevitably take too much time out of the average practice’s workload.

He said: ‘Practices do already review databases regularly, when there is time in the consultation GPs will update a patient’s records on the basis of new information. But if you carry out a review of your dementia register that is inevitably time you won’t be spending on your asthma or other chronic patients.’

Professor Steve Iliffe, professor of primary care for older people at University College London, who has questioned the Government’s estimates on the predicted prevalence of dementia and expressed concerns about over-simplifying the causes of under-diagnosis, said the study corroborated as-yet unpublished findings from the EVIDEM-ED trial, in which ‘preliminary analyses suggest that miscoding and undercoding were common’.

Professor Iliffe told Pulse: ‘A 9% increase is not particularly large, but also it does not make that much difference if about half of probable cases are not yet on the QOF register.

‘I still think the predicted case numbers are inflated, and that there are other factors operating too.’

An NHS England spokesperson said: ‘NHS England is clear about the importance of both timely diagnosis of patients with dementia and creating an accurate clinical record. The latter is simply good clinical practice and anything which stimulates this is to be applauded, especially something as straightforward as is described.

‘Timely diagnosis is a prerequisite to post diagnostic support and facilitates better planning for future care involving the patient’s family and principal carer(s) and  will facilitate the support and care we have heard people with dementia and their carers seek.’


Visit Pulse Reference for details on 140 symptoms, including easily searchable symptoms and categories, offering you a free platform to check symptoms and receive potential diagnoses during consultations.