Easy Probability Estimation for Diagnosing Early Axial SpA
Easy Probability Estimation for Diagnosing Early Axial SpA
Objectives. Several sets of criteria for the diagnosis of axial SpA (including non-radiographic axial spondyloarthritis) have been proposed in the literature in which scores were attributed to relevant findings and the diagnosis requests a minimal sum of these scores. To quantitatively estimate the probability of axial SpA, multiplying the likelihood ratios of all relevant findings was proposed by Rudwaleit et al. in 2004. The objective of our proposal is to combine the advantages of both, i.e. to estimate the probability by summing up scores instead of multiplying likelihood ratios.
Methods. An easy way to estimate the probability of axial spondyloarthritis is to use the logarithms of the likelihood ratios as scores attributed to relevant findings and to use the sum of these scores for the probability estimation.
Results. A list of whole-numbered scores for relevant findings is presented, and also threshold sum values necessary for a definite and for a probable diagnosis of axial SpA as well as a threshold below which the diagnosis of axial spondyloarthritis can be excluded. In a diagram, the probability of axial spondyloarthritis is given for sum values between these thresholds.
Conclusion. By the method proposed, the advantages of both, the easy summing up of scores and the quantitative calculation of the diagnosis probability, are combined. Our method also makes it easier to estimate which additional tests are necessary to come to a definite diagnosis.
As a means of making the diagnosis of early axial SpA including non-radiographic axial SpA, (i.e. axial SpA without radiographic sacroiliitis), several sets of criteria were proposed in the literature in which scores from ½ to 1½ or from 1 to 3 have been attributed to the relevant findings. According to these criteria, the diagnosis of SpA can be made if the sum of relevant scores exceeds a certain threshold requested for this diagnosis.
The advantage of these proposals is their simplicity. However, they do not allow quantitative estimation of the probability of the disease. They allow only a qualitative yes or no answer to the question of whether the diagnosis is probable or not.
A method for quantitative estimation of the probability of axial SpA in a patient with chronic back pain, age at onset <45 years and no definite X-ray changes in the sacroiliac joints was proposed in 2004 by Rudwaleit et al.. They made use of the likelihood ratios (LRs) of test results relevant for the diagnosis. A LR is the likelihood of a given test result in a person with a disease compared with the likelihood of this result in a person without the disease. The LRs LR = sensitivity/(1 − specificity) for any positive finding (derived from the sensitivity and specificity of the tests) and LR = (1 − sensitivity)/specificity for any negative finding are multiplied, and the post-test probability P post is calculated from the pre-test probability P pre and the LR product ΠLR according to the following equation:
The objective of our proposal is to combine the advantages of both, i.e. to estimate the probability by summing up scores attributed to the clinical, laboratory and imaging features instead of multiplying their LRs.
Abstract and Introduction
Abstract
Objectives. Several sets of criteria for the diagnosis of axial SpA (including non-radiographic axial spondyloarthritis) have been proposed in the literature in which scores were attributed to relevant findings and the diagnosis requests a minimal sum of these scores. To quantitatively estimate the probability of axial SpA, multiplying the likelihood ratios of all relevant findings was proposed by Rudwaleit et al. in 2004. The objective of our proposal is to combine the advantages of both, i.e. to estimate the probability by summing up scores instead of multiplying likelihood ratios.
Methods. An easy way to estimate the probability of axial spondyloarthritis is to use the logarithms of the likelihood ratios as scores attributed to relevant findings and to use the sum of these scores for the probability estimation.
Results. A list of whole-numbered scores for relevant findings is presented, and also threshold sum values necessary for a definite and for a probable diagnosis of axial SpA as well as a threshold below which the diagnosis of axial spondyloarthritis can be excluded. In a diagram, the probability of axial spondyloarthritis is given for sum values between these thresholds.
Conclusion. By the method proposed, the advantages of both, the easy summing up of scores and the quantitative calculation of the diagnosis probability, are combined. Our method also makes it easier to estimate which additional tests are necessary to come to a definite diagnosis.
Introduction
As a means of making the diagnosis of early axial SpA including non-radiographic axial SpA, (i.e. axial SpA without radiographic sacroiliitis), several sets of criteria were proposed in the literature in which scores from ½ to 1½ or from 1 to 3 have been attributed to the relevant findings. According to these criteria, the diagnosis of SpA can be made if the sum of relevant scores exceeds a certain threshold requested for this diagnosis.
The advantage of these proposals is their simplicity. However, they do not allow quantitative estimation of the probability of the disease. They allow only a qualitative yes or no answer to the question of whether the diagnosis is probable or not.
A method for quantitative estimation of the probability of axial SpA in a patient with chronic back pain, age at onset <45 years and no definite X-ray changes in the sacroiliac joints was proposed in 2004 by Rudwaleit et al.. They made use of the likelihood ratios (LRs) of test results relevant for the diagnosis. A LR is the likelihood of a given test result in a person with a disease compared with the likelihood of this result in a person without the disease. The LRs LR = sensitivity/(1 − specificity) for any positive finding (derived from the sensitivity and specificity of the tests) and LR = (1 − sensitivity)/specificity for any negative finding are multiplied, and the post-test probability P post is calculated from the pre-test probability P pre and the LR product ΠLR according to the following equation:
The objective of our proposal is to combine the advantages of both, i.e. to estimate the probability by summing up scores attributed to the clinical, laboratory and imaging features instead of multiplying their LRs.
