PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of brjgenpracRCGP homepageJ R Coll Gen Pract at PubMed CentralBJGP at RCGPBJGP at RCGP
 
Br J Gen Pract. 2007 August 1; 57(541): 678–679.
PMCID: PMC2099686

Diagnostic testing: the importance of context

Holm and colleagues' study of the Kryptor®-PCT assay1 illustrates the importance of undertaking diagnostic research in the appropriate setting. Their findings reveal the much lower discriminatory power of procalcitonin in primary care patients in comparison with hospitalised patients.

However, in interpreting the results of any such diagnostic research and assessing the importance of the findings, it is also helpful to consider two additional contextual factors: the other elements of the clinical assessment and the place of the new technology within a diagnostic processing pathway.

From the data provided by Holm et al, I have calculated positive and negative likelihood ratios (LRs) by comparing the blood results against the radiographic ‘reference standard’ (Table 1).

Table 1
Positive and negative likelihood ratios for pneumonia diagnosis.

The magnitude of the LR provides a measure of the predictive ability of a clinical indicant (for example, symptom, sign, or test finding). Clinical indicants with LRs greater than 1 increase the chances of disease: the larger the LR the more compelling the argument for disease. Conversely, clinical indicants that have LRs between 1 and 0 decrease the probability of disease: the closer the LR to zero, the more convincing the finding argues against disease. The adjectives ‘positive’ or ‘negative’ indicate whether the LR refers to the presence of the clinical information (positive) or the absence of the clinical information (negative). Positive LRs with the highest value argue most for disease when the clinical information is present; negative LRs with the value closest to zero argue the most against disease when that clinical information is absent.

In my recently published book Patient Centred Diagnosis2 I have assembled a number of LRs for clinical assessment. From this it seems that a duration of illness less than 24 hours before consulting a GP was the variable in the history with the highest positive LR for pneumonia diagnosis (Table 2).

Table 2
Promptness of consulting and pneumonia diagnosis.

The LRs for a number of more traditional clinical features used to determine whether an adult has a community-acquired pneumonia are shown in Table 3.

Table 3
Likelihood ratios for pneumonia diagnosis in adults.

Although some individual findings, such as raised respiratory rate, elevated temperature, dullness to percussion, and bronchial breath sounds, provide substantial positive LRs, clusters of findings are more powerful, especially as some individual findings may be unreliable. The combination of temperature of greater then 37.8°C, heart rate more than 100 beats per minute, crackles, and diminished breath sounds in a patient without asthma provides a positive LR of 8.2, while the absence of this combination produces a negative LR of 0.3. In the study by Holm et al the interquartile range for procalcitonin was 0.04–0.08 ng/ml (median 0.05 ng/ml); with this in mind, it is worth noting that a procalcitonin level of ≥0.06 ng/ml only provides a positive LR of 2.06.

In using LRs in the context of clinical practice, Bayes' theorem is a very helpful tool to assist in the understanding of diagnostic processing. It is most clearly expressed in the form:

Posterior Odds=LR×Prior Odds

This formula emphasises that the interpretation of the significance of any new information should depend on our existing knowledge about the probability of a disease (the prior probability or prior odds of disease). Thus, a patient who comes to see their primary care physician about a cough will already have a prior (existing) probability of pneumonia. This probability will be modified by additional information derived from the medical history to arrive at a new (post-history) probability of pneumonia. This probability may, in turn, be further adjusted by data derived from the clinical examination to produce a post-examination probability that, after a procalcitonin test, could then become a post-test probability. Thus, in an idealised form, the diagnostic processing pathway can be seen as a number of probability steps increasing the certainty of disease (or absence of disease; Figure 1).

Figure 1
Diagnostic processing pathway.

It may be that, in many circumstances, the disease probability after the history and examination (the post-examination probability) is such that undertaking an investigation is actually unnecessary whatever its LR!

REFERENCES

1. Holm A, Pedersen SS, Nexoe J, et al. Procalcitonin versus C-reactive protein for predicting pneumonia in adults with lower respiratory tract infection in primary care. Br J Gen Pract. 2007;57(540):555–560. [PMC free article] [PubMed]
2. Summerton N. Patient centred diagnosis. Abingdon: Radcliffe Publishing; 2007.

Articles from The British Journal of General Practice are provided here courtesy of Royal College of General Practitioners