In the window of the Wellcome Collection in London artists work to interpret and explain science: it’s an impressive experience to the irregular visitor. When faced with the presenting problems of a child & family, we are faced with trying to do the reverse. We have the sometimes inaccurate recollections of history, the variable responses of clinical examination and our own bias-riddled minds to bash, through the ‘art of diagnosis’ into a suitable explanation for the predicament and onwards into a management strategy. Can this really be evidence based?
The evidence based of diagnosis has been discussed in this column before [here and here], but usually in the setting of a single item, for example “How good is Bobbinogs Sign at predicting Rhotic Consonant pronunciation?”. What about clinical decision rules – simple combinations of clinically & laboratory parameters which are intended to more accurately and reliably make or exclude a diagnosis?
There are a few critical elements to bear in mind when appraising a clinical decision rule: it’s method of derivation, it’s usefulness when applied in different settings, and it’s predictive ability. A decision rule is usually developed by taking a set of data and seeing which are the simplest rules that lead to defining groups at lowest & highest risk of a disease. This method is entirely dependant on the data it’s made from – any chance associations will be impossible to identify from this. The rule needs to be tested again, either in the same place (OK) or in a different setting (better). (Different doctors, types of patients and situations can make a rule unusable. This may be particularly true of rules including elements of physical examination or history taking.) Confidence in a rule should be even better still if it had been tested in many different areas, and actually proven to make a difference to patient important outcomes. Finally, a judgement needs to be made about its predictive ability. Do you need it to make a diagnosis or exclude one? (Most rules are developed to do one, or the other, but not both.) How certain do you need to be of this? These are factors that require assimilation of the risks, patient preferences and healthcare structures in your own location.
McGinn, T.G. et al., 2000. Users’ Guides to the Medical Literature: XXII: How to Use Articles About Clinical Decision Rules. JAMA, 284(1), p.79-84.
Phillips, R. Disease spectrum vs. disease prevalence. Arch Dis Child 2008;93:628
Phillips, R. Test/don’t test?. Arch Dis Child 2005;90:1308
Roukema J, Steyerberg EW, Lei Jvd, Moll HtA. Randomized Trial of a Clinical Decision Support System: Impact on the Management of Children with Fever without Apparent Source. Journal of the American Medical Informatics Association : JAMIA 2008;15:107-113
Stiell, I. et al., 1995. Multicentre trial to introduce the Ottawa ankle rules for use of radiography in acute ankle injuries. BMJ, 311(7005), p.594-597.