Diagnostic downshifting: the hidden dangers of testing
Diagnostic downshift is a term that describes the phenomenon of shifting specialist diagnostic tests from secondary care (hospital settings) to primary care (general practice settings).
This may be driven by various factors, such as the desire to detect diseases earlier, the pressure to reduce referrals and waiting times, or the availability of new technologies. However, diagnostic downshift may have unintended consequences, such as increased false positives, overdiagnosis, overtreatment, patient anxiety, clinician workload, and healthcare costs. Therefore, diagnostic downshift should be carefully evaluated and implemented, considering the different characteristics and needs of primary care populations and settings.
One possible example of diagnostic downshift in the NHS is the use of faecal calprotectin (FC) testing in primary care to rule out inflammatory bowel disease (IBD) in patients with lower gastrointestinal symptoms. FC is a protein that is released by white blood cells in the gut and can indicate inflammation. FC testing was originally developed and validated in secondary care settings, where the prevalence of IBD is higher and the patients have more severe symptoms. However, FC testing has been increasingly adopted in primary care as a way to reduce referrals and colonoscopies for patients with suspected IBD.
A brief case history to illustrate this is:
A 45-year-old woman presents to her GP with a 6-month history of intermittent diarrhoea, abdominal pain and bloating. She has no red flag symptoms, such as rectal bleeding, weight loss or anaemia. She has a family history of irritable bowel syndrome (IBS), but not of IBD or colorectal cancer. She is otherwise healthy and takes no regular medications.
The GP suspects IBS, but decides to order a FC test to rule out IBD, following the local guidelines. The test result comes back as 150 μg/g, which is above the cut-off of 50 μg/g for a negative test. The GP interprets this as a positive result and refers the patient to gastroenterology for further investigation.
The patient is anxious about the possibility of having IBD and undergoes a colonoscopy, which shows normal findings. The gastroenterologist diagnoses her with IBS and advises her to follow dietary and lifestyle modifications. The patient is relieved, but also feels frustrated by the unnecessary invasive procedure and the delay in receiving appropriate treatment.
This case illustrates some of the potential pitfalls of diagnostic downshift, such as:
The lower accuracy of FC testing in primary care, due to spectrum bias and the false-positive paradox. In low-prevalence settings, such as primary care, most positive results are likely to be false positives, especially if the test is not very specific. FC testing has a specificity of about 80% for IBD, which means that 20% of patients without IBD will have a positive result. Therefore, a positive FC test in primary care does not necessarily indicate IBD, but rather a need for further evaluation.
The indication creep and overdiagnosis of FC testing in primary care FC testing may be used for patients who do not have a high clinical suspicion of IBD, such as those with mild or non-specific symptoms, or those who have alternative diagnoses, such as IBS. This may lead to overdiagnosis of IBD, which is the detection of a condition that would not have caused harm or required treatment if left undetected. Overdiagnosis may expose patients to unnecessary harms from further tests, treatments and labelling.
The downstream consequences and costs of FC testing in primary care. FC testing may generate more referrals, colonoscopies, treatments and follow-ups for patients with positive results, which may increase the workload, anxiety and costs for both patients and clinicians. Moreover, FC testing may not reduce the demand for secondary care services, as some patients with negative results may still require referral due to persistent or worsening symptoms. Therefore, FC testing may not provide value for money or improve outcomes for patients with lower gastrointestinal symptoms in primary care.
Diagnostic downshift should be carefully evaluated and implemented in primary care settings, taking into account the different characteristics and needs of primary care populations and processes. Test formularies should be scrutinised in view of the setting of care, with interventions to focus rational testing towards those with higher pretest probabilities, while improving interpretation and communication of results.
Explainers
Spectrum bias is a type of bias that can occur when the performance of a diagnostic test or treatment is evaluated in a population that is not representative of the population in which the test or treatment will be used. This can lead to inaccurate estimates of the test's or treatment's performance, which can have negative consequences for patients.
The false positive paradox is a statistical phenomenon that occurs when a test has a high false positive rate but a low prevalence of the condition it is testing for. This means that the test is more likely to incorrectly indicate that a person has the condition than to correctly indicate that they do.
The false positive paradox is often illustrated with the following example:
Suppose there is a test for a disease that has a false positive rate of 1% and a prevalence of 0.1%. This means that out of 1,000 people tested, 10 will test positive, but only 1 of those 10 will actually have the disease.
If we look at the test results of 10 people who test positive, we see that only 1 of them actually has the disease. This means that there is a 90% chance that a person who tests positive for the disease does not actually have it.
This example shows that a test with a high false positive rate can be very misleading, especially if the prevalence of the condition it is testing for is low.
Sajid, I., Frost, K. and Paul, A.K. (2021). ‘Diagnostic downshift’: clinical and system consequences of extrapolating secondary care testing tactics to primary care. BMJ evidence-based medicine, [online] 27(3), pp.141–148. doi:https://doi.org/10.1136/bmjebm-2020-111629.