Introducing a new classification scheme, Dr Gerada and Dr Hillson say new research and genetic studies have identified many variants of diabetes.
'While general care can always be provided, a proper diagnosis is needed before any diagnosis or condition can be fully treated,' they say.
Dr Hillson says that, as knowledge of the genetic basis of diabetes grows, the importance of classification will be further refined.
Professor Kamlesh Khunti, professor of primary care diabetes at the University of Leicester, says that the development of new drugs will also make correct classification increasingly important. Many new treatments are only suitable for patients with specific forms of diabetes, he says.
The RCGP and NHS Diabetes also point out that misdiagnosis can worry patients and undermine trust in clinicians.
International groups have attempted to develop classification schemes to make sense of these new types and subtypes. But these have been complex and based on information from tests not usually available in a clinical setting.
A group set up by the RCGP and NHS Diabetes decided that a practical classification scheme that non-experts could use needed to be developed.
In researching the new classification scheme, the RCGP and NHS Diabetes carried out an audit that found 'substantial evidence of the miscoding and misclassification of diabetes'.
Analysis of one million patient records showed that 40% of diabetes patients had classification anomalies in their data. Some errors were simple mistaken entries in records, even though clinicians and the patients knew the correct classification. Others were from a lack of information or understanding by clinicians.
The audit found the most widespread misunderstanding was changing a diagnosis from type-2 to type-1 diabetes when patients began insulin therapy.
Dr Khunti: QOF must be updated (Photograph: P Hill)
Professor Khunti says that this confusion may partly arise from clinicians retaining the mindset of the 'non-insulin-dependent' and 'insulin dependent' terminology used before 'type-1' and 'type-2' became standard. He adds that patients may not have been coded correctly when the DM19 indicator was added to QOF in 2006.
DM19 requires GPs to classify patients as having either type-1 or type-2 diabetes. But GPs were not given any guidance on how patients should be classified into these two types, he says. QOF now needs to be updated to include the new classification system, he adds.
'It is not a complex algorithm, so you could easily have that in QOF. We would put them in a category of genetic or other. There are genetic codes, but they are not counted towards QOF.'
To help determine which patients may have been misclassified, the RCGP and NHS Diabetes have developed an audit tool. Available online, it can extract data from all the main practice IT system suppliers.
In a trial in five practices, around 40 patients in each practice were found to need their records checking.
Professor Simon de Lusignan of the University of Surrey helped develop the audit tool. He says the tool and supporting guidance give a 'blow by blow' account of how GPs can identify misclassified patients.
Most of these patients can be quickly and easily assessed and their records amended after a review of their notes, with only a few patients requiring more thorough analysis, he says.
These patients could then be reassessed at their regular review appointments, he says.
He also stresses that, for most people who are mislabelled, this would not mean they have been given worse treatment. Treatment decisions would have been guided by the needs of each individual, he says.
Correctly classifying patients will mean they are called for reviews at the correct intervals and that audits of care across patients will be more accurate.
'Overall the system will produce higher quality data,' he says. 'This is UK general practice being ahead of the curve, getting ahead, and wanting to "fine-tune" the process.'
The RCGP and NHS Diabetes have developed a new classification scheme to account for evidence that diabetes cannot simply be divided into type-1 and type-2. It suggests the following categories:Diabetes