Diagnostic tests 4: likelihood ratiosBMJ 2004; 329 doi: https://doi.org/10.1136/bmj.329.7458.168 (Published 15 July 2004) Cite this as: BMJ 2004;329:168
All rapid responses
Two recently published articles in BMJ, one in endorsement of
likelihood ratios (LRs) (1) and the other in censure of SnNOut and SpPIn
properties of the diagnostic tests (2), have truly emphasized on the
shortcomings accompanying sensitivity and specificity and the need for
utilization of LRs. We believe, however, that the most crucial
characteristic of LRs, which have not been accentuated properly, is their
dynamicity and applicability in clinical settings.
In a static circumstance, we need tests just to rule in or rule out a
diagnosis. The most renowned directive for this purpose is use of signs
and symptoms with high sensitivity or specificity to SnNOut or SpPIn.
Nevertheless, getting a positive answer to a sensitive test, for instance,
has no effect on decision-making process and discontinues the progression.
Moreover, assuming that a diagnosis can be ruled in or ruled out with
confidence, when in reality it cannot, could have serious consequences for
In a dynamic state, however, we are dealing with probabilities and
thresholds that describe our understanding of the disease. In this
approach, we can make the diagnosis or start the therapy just after
surpassing the diagnostic or treatment threshold. In clinical
circumstances, typically full of uncertainty (3), we need tests to change
the degree of our uncertainty in order to surpass our estimate of
probability of a disease around a predefined threshold. LRs are dynamic
measures that have this strength. In clinical practice, using a nomogram
or even simple consideration of the significance of LR of a test, we can
re-estimate the probability of a disease.
The desire to completely rule in or rule out a diagnosis is almost a
plague induced by mathematical (intuitive) thinking in clinical medicine.
Contrary to this viewpoint, evidence-based diagnostic approach appreciates
existence, and physicians’ inability to get rid of, uncertainty in
clinical medicine. LRs are the most dynamic and the most applicable tools
for refining probability in clinical settings.
1) Deeks JJ, Altman DG. Diagnostic tests 4: likelihood ratios. BMJ
2) Pewsner D, Battaglia M, Minder C, Marx A, Bucher HC, Egger M. Ruling a
diagnosis in or out with "SpPIn" and "SnNOut": a note of caution. BMJ
3) Hunink M, Glasziou P. Decision making in health and medicine:
integrating evidence and values. Cambridge University Press, 2001.
Competing interests: No competing interests