Re: Clinicians’ gut feeling about serious infections in children: observational study
Reading the paper by Van den Bruel et al.(1) elicited a intuitive feeling, perhaps even gut feeling, that something in this paper is wrong, or at least not quite right.
Van den Bruel et al. propose that gut feelings of clinicians should not be neglected, even if the clinical impression of the patient/child under their care suggests that there is no serious illness.(1) In the manuscript they present the diagnostic characteristics (sensitivity, specificity and positive likelihood ratio) of having a gut feeling the child has something serious.
Diagnostic tests need to be standardized. Western medicine's scientific framework requires that information be documented in a manner which is understood and reproducible, allowing knowledge to accrue, enabling the community of medical professionals to share a collective experience. This report does not really contribute such knowledge. The authors used a somewhat circular definition to “gut feeling”: “a feeling that something is wrong even if the clinician was not sure why”. The study's working definition is further obscured when we subtract the lower part if table 1 (children in whom clinical impression was of a non-serious illness) from the upper part (all children). This results in 63 children for whom the clinical impression was of a serious illness and the clinician had a gut feeling something was wrong.
Another feature of diagnostic tests is that, in practice, they are used both for diagnosis and for exclusion. The D-dimer test has been repeatedly demonstrated to serve as a rule-out test in patients suspected to be suffering from pulmonary embolism, who have a low a-priory risk of pulmonary embolism. Inappropriate interpretation of a positive D-dimer is all too common.(2,3) Proposing that gut feelings have clinical utility implies that the absence of such feeling also have clinical utility. Again, subtracting the lower part of table 1 from the upper part reveals that the absence of a gut feeling has a specificity of 88% among children in whom the clinical impression was that they do have a serious problem (negative LR 0.3). This result is less attractive and clearly cannot translate into clinical recommendations.
Unlike diagnostic tests, gut feeling can relate to any clinical suspicion. Interestingly, there was no report about gut feelings that non-infectious serious conditions exists – such as child abuse and neglect, developmental disorders and social dysfunction. These are universally known to be under-diagnosed all too often and are far more challenging to objectively diagnose, compared with infectious diseases.
I also wish to stress the points raised by Graham L Briars et al., who provided an earlier rapid response to this paper. In their previous 2007 analysis of this same patient sample, Van den Bruel et al. explicitly detailed the advantages of the CART analysis over the statistical methods they used in the current publication.(4) This inconsistency demands explanation. However, this is the least important problem. As stated by Graham L Briars et al., the current analysis in missing 91 children which were included in the previous report. This analysis is also missing 10 cases of serious illness which were reported in the previous publication, and some of the clinical diagnoses seem to have changed.
The main topic discussed in the current paper was the documentation of a gut feeling and the merits of such a decision/thought/feeling. In the methods section the authors state that doctors were explicitly asked to document intuitive feelings that might arise from the condition of the child or the behavior of the parents. In the previous paper from 2007, the authors stated that doctors were asked to state whether they “thought something was wrong”, again, circularly define with a negative phrase – “a subjective feeling of the physician that things were not right”. Worst of all, in the 2007 paper there were 137 patients among whom there was “a feeling that something was wrong”; however, in the current 2012 paper, there were 120 patients among whom there was a gut feeling something was wrong. I feel that the authors and BMJ should be concerned that these discrepancies exist.
Seeking complete diagnostic accuracy strategies is a Holy Grail quest. We should acknowledge that clinicians are sometimes /often mistaken in their clinical judgment. We should embrace and improve established clinical skills and better communicate what we know and what we do not know with both patients, parents, other doctors and nursing staff. These skills require acquaintance with up-to-date knowledge of the rational medical history taking and clinical examination.(5)
There are, and should be, some unalienable truths in the practice of medicine. They constitute values which, combined, draw an image of what we want medicine to be and how it should be practiced. They are not dependant on evidence which proves or disproves their utility or benefit. Among these are the capability to communicate with patients in compassion and care and an inherent effort to do good for the patient. If our instincts/feeling/intuition tell us that something is wrong with the patient, we should work on that assumption and refer and/or test and/or consult with others – even if we are only following a gut feeling. It is our calling and requires no evidence base.
1. Van den Bruel A, Thompson M, Buntinx F, Mant D.Clinicians’ gut feeling about serious infections in children: observational study. BMJ 2012;345:e6144
2. Jones P, Elangbam B, Williams NR. Inappropriate use and interpretation of D-dimer testing in the emergency department: an unexpected adverse effect of meeting the ‘‘4-h target’’. Emerg Med J 2010;27:43–47. doi:10.1136/emj.2009.075838
3. Yin F, Wilson T, Della Fave A, Larsen M, Yoon J, Nugusie B, Freeland H, Chow RD. Inappropriate Use of D-Dimer Assay and Pulmonary CT Angiography in the Evaluation of Suspected Acute Pulmonary Embolism. Am J Med Qual. 2012 Jan-Feb;27:74-9
4. Van den Bruel A, Aertgeerts B, Bruyninckx R, Aerts M, Buntinx F. Signs and symptoms for diagnosis of serious infections in children: a prospective study in primary care. Br J Gen Pract2007;57:538-46.
5. Simel D, Rennie D. The Rational Clinical Examination: Evidence-Based Clinical Diagnosis. McGraw-Hill Professional; 1 edition (August 25, 2008)
Competing interests: No competing interests