Among the commonest tests ordered by doctors is a full blood count. The test presents signs of iron deficiency anaemia. The prevalence of that condition is reported as follows:
In Australia in 2011–12, around 760,000 people aged 18 years and over (4.5%) were at risk of anaemia, with women more likely to be at risk than men (6.4% compared with 2.5%). The risk of anaemia was highest among older Australians, with rates rapidly increasing after the age of 65 years. People aged 75 years and older were more likely to be at risk of anaemia than all other Australians, with 16.0% in the at risk range compared with 3.6% of Australians aged less than 75 years. Australian Health Survey
With regard to this blood test (AACC):
- Haemoglobin (Hb)—may be normal early in the disease but will decrease as anaemia worsens
- Red blood cell indices—early on, the RBCs may be a normal size and colour (normocytic, normochromic) but as the anaemia progresses, the RBCs become smaller (microcytic) and paler (hypochromic) than normal.
- Average size of RBCs (mean corpuscular volume, MCV)—decreased
- Average amount of haemoglobin in RBCs (mean corpuscular haemoglobin, MCH)—decreased
- Haemoglobin concentration (mean corpuscular haemoglobin concentration, MCHC)—decreased
- Increased variation in the size of RBCs (red cell distribution width, RDW)
- A guide to interpreting the test is here
Therefore among the pathognomonic features of established iron deficiency anaemia (IDA) is a low Mean Corpuscular Volume (MCV). The sensitivity and specificity of a low MCV for a diagnosis of iron deficiency anaemia are quoted as 42% and 93%. Assuming a prevalence of 3.6% in the under 75 year old age group this means that if 100 adults in Australia had a full blood count then 3-4 will have iron deficiency anaemia. Screening these people for IDA with this test 8.3% of people will be told they have an abnormal test i.e. 8 people. Of these only 1-2 will be a true positive for IDA. On the other hand 6-7 may be misled into thinking they might have iron deficiency. 91 will be told they have a normal test in this case 2 may be incorrectly reassured. Of course there are other significant conditions which present with a microcytosis ( low MCV) although ‘treatment’ is not necessary in many such cases and also screening for IDA involves other and more sensitive tests.
If the prevalence of the condition was 20%, then even the modest sensitivity and specificity of this test would identify more people at risk of IDA even though it will also miss people with the condition.
- Number of people with positive test: 14, correctly identified: 8
- Number of people with negative result: 86, incorrectly reassured: 11
In practice the sensitivity and specificity of tests may be assumed closer to 90% in each case. Given these figures the numbers of people from 100 people test and correctly identified, incorrectly reassured or told they are ill depends on the prevalence. The prevalence of most pathology in the community is low often well below 1%. The figures are presented in the infographic below.
Prevalence 0.005% ( 5 per 1000 people, e.g. hypothyroidism ) 2% ( e.g. diabetes) 20% (e.g. common and plantar warts).
From these figures it can be seen that testing is more fruitful in circumstances in which the prevalence is high. The prevalence is higher in those who have signs and symptoms of a condition. One could argue therefore that the ‘prevalence’ is much higher in those who choose to consult a doctor as opposed to the ‘prevalence’ in the community. For iron deficiency anaemia these circumstances are well known. Which means an effective consultation in which the patient is heard and examined is crucial to interpreting test results. As can be seen from the calculations there is a substantial risk of labelling people as ill, or requiring yet more tests given the modest prevalence of most conditions in the community and where there might be an indiscriminate use of tests.
It is hard to disagree with Campbell and colleagues who considered this issue and noted that:
1) Diagnosis is based on a combination of tests and clinical examination and there is little research based on the sensitivity and specificity of the combination of different examinations as opposed to a one-off test, which is why GPs are unlikely to know the values.
2) It is unclear what is meant by the prevalence of asthma or diabetes for these GPs. It is not the proportion of people in the population with the disease, but rather the proportion of people who come to consult who have the disease (perhaps with similar age and clinical history). This proportion is likely to be quite high and so the issue of overestimating the positive predictive value is less important.
3) The prevalence of the disease will also depend on the severity of the disease being tested for and so this also muddles the calculations.
We might however equally reasonably expect doctors to have an understanding of the issue if only because the practice of medicine involves the most crucial of ‘tests’ the history and the examination and this issue highlights the importance of that activity. Tests that are not appropriately interpreted can be harmful if only because they become a source of anxiety.
Picture by Aplonid