There is substantial variation in the use of tests in general practice before patients are diagnosed with cancer, researchers have found.
A study of 53,000 patient records identified a fourfold difference in the use of any investigation, blood tests and imaging, before a cancer diagnosis.
Overall, 56% of patients in the study had at least one investigation, although this varied among practices.
The most common was a blood test, given to 44% of patients, followed by imaging at 23%. The least frequent was endoscopy at 2% which may relate to access, the researchers said.
The variation could not really be explained by practice characteristics, including size, and there was even greater variation when case mix was taken into account because some types of cancer were over-represented in practices with lower-than-average rates of investigation.
They did find that rural practices had lower use of any investigation and those practices with more older age patients were more likely to use all types of investigations.
Higher levels of comorbidity were found to be associated with lower use of investigations, which may reflect the fact that recent use of investigations, such as blood tests, is higher in this patient group, so GPs may be less inclined to order new ones.
Reporting the findings in BMJ Quality and Safety, the researchers from the University of Exeter Medical School, said both over-testing and under-testing may be happening.
Yet with no robust evidence on what the right level of testing is, doctors are ‘working in the dark’ about what investigations are appropriate.
While investigations can lead to an earlier diagnosis, they could also delay access to some specialist testing so guidance for practices was needed once more evidence was available, they said.
As happens with prescribing volume, it would be helpful for practices to receive regular updates on their relative use of investigations, such as blood tests, compared with other practices to ‘support reflective practice’.
GP researchers found last year that a quarter of blood tests done in general practice may be unnecessary and would not lead to a change in management.
Study leader Professor Gary Abel, deputy director of the Exeter Collaboration for Academic Primary Care told Pulse that the substantial variation in testing found between practices was much larger than any variation seen between different types of patients once the type of their cancer was accounted for.
‘In some ways this is not surprising as we do not have good evidence on what the right level of testing is, and doctors are often working in the dark as there is no robust evidence to guide their decisions about what investigations are appropriate,’ he said.
‘Whilst investigations can help facilitate an earlier diagnosis of cancer in many cases, with an improved outlook and treatment options available for earlier stage disease, most of those who are investigated will not have cancer.
‘It is also likely that overuseof these particular investigations can delay access to more specialist investigation and may result in worse outcomes for some patients.
‘Novel research is required to find this sweet spot between over and under-testing, and how to target investigations appropriately at the right patients at the right time.’
