A clinical prediction rule represents a distillation of measurable features, usually by regression modelling. It helps standardise the approach to diagnosis and, in theory, should reduce variation in diagnosis and inappropriate prescribing.
Our review compared Centor’s score with McIsaac’s score.1 These are the two most recommended prediction rules used for diagnosing GABHS-related pharyngitis in different national guidance.2,3 Although other scores such as Walsh’s score exist, they tend not to feature in international guidance.2,3
FeverPAIN does appear in UK guidance and was developed in the UK by one of the co-authors of the editorial.4 However, the derivation study is still the only study that has evaluated the rule and so meta-analysis was not possible.
Our review demonstrates for both scores that there is substantial variation in performance across different settings.1 Furthermore, the two studies that reported the most favourable receiver operating characteristic (ROC) curves for McIsaac’s score have McIsaac as first author. Both of these points reinforce the need for multiple independent validation studies of FeverPAIN before we conclude on its accuracy.
Clinical features, in whichever combination or weighting, are unlikely to be sufficient to rule in GABHS pharyngitis. Point-of-care (POC) tests vary in shape and form, with some more disruptive to the consultation than others. But this should be weighed against the likely benefit they can bring — we don’t hesitate in sending a patient to the loo to produce a urine sample if a subsequent dipstick test helps diagnose a urinary tract infection. So it may be in the future that a POC test augments one of the scores sufficiently to reduce diagnostic errors and the inappropriate prescribing of antibiotics.
Otherwise we may use clinical gestalt, which allows for less measurable, intangible features to be included in the diagnostic process. This is not without merits, but it is also more likely to vary between practitioners and be open to cognitive biases.
Notes
Competing interests
We are the authors of the meta-analysis that compared Centor’s score with McIsaac’s score.
- © British Journal of General Practice 2020
