Reproductibilité de l’Association américaine pour la chirurgie de la mise à l’échelle des traumatismes pour les lésions rénales : question méthodologique pour éviter une mauvaise interprétation

25 avril 2018

Auteurs : S. Sabour
Référence : Prog Urol, 2018, 5, 28, 239-240




 



Dear editor, I was interested to read the paper by Phan QB and colleagues published in Prog Urol 2017 Nov.1 The American Association for the Surgery of Trauma (AAST) Organ Injury Scale (OIS) is the most used classification for renal trauma. The purpose of the authors was to assess the subjectivity of AAST scaling. Computed Tomography (CT) scan were analyzed and injuries graded according to AAST OIS independently by a senior radiologist, a senior urologist who was blind to clinical data and a resident urologist. They reported that ninety-seven patients had 101 renal injuries: low grade in 58.4% (11.9% grade I, 17.8% grade II, 28.7% grade III) and high grade in 41.6% of cases (23.6% grade IV and 17.8% grade V). The agreement was fair with Kappa coefficient at 0.36. The agreement was moderate in severity sub-division analysis (low or high grade): Kappa coefficient at 0.59. There was a disagreement in 49.5% between the senior urologist's and the senior radiologist's ratings [1].


However, it is crucial to know that two important weaknesses of kappa are as follows. First, it depends on the prevalence in each category, which means it can be possible to have different k values having the same percentage for both agreement and disagreement cells. Table 1 shows that in both (a) and (b) situations, the prevalence of agreement cells are 90% and of disagreement cells, 10%; however, we get different kappa values (0.44 as moderate and 0.80 as very good, respectively). Kappa value also depends on the number of categories [2, 3].


They concluded that AAST OIS for renal trauma suffers from subjectivity. Such conclusion should be supported by the above mentioned methodological and statistical issues. Otherwise, misinterpretation of the finding cannot be avoided.


Disclosure of interest


The author declares that he has no competing interest.




Table 1 - Limitation of kappa for comparison of two observers' diagnoses with different prevalence in the two categories.
    Observer 1 
 
    Low grade  High grade  Total (%) 
Situation (a)          
Observer 2  Low grade  85  90 
Kappa=0.44  High grade  5  10 
  Total  90  10  100 
Situation (b)          
Observer 2  Low grade  45  50 
Kappa=0.80  High grade  45  50 
  Total  50  50  100 



Légende :
In both situations a & b, the prevalence of concordance cells are equal to 90% (Bolds), however, we get different kappa value.


References



Phan Q.B., Mourey E., Estivalet L., Delattre B., Bardet F., et al. Reliability and reproducibility of the American Association for the Surgery of Trauma scaling for renal injury and impact on radiologic follow-up Prog Urol 2017 ; 10.1016/j.purol.2017.09.013[pii: S1166-7087(17)30565-1].
Szklo M., Nieto FJ: Epidemiology beyond the basics  Sudbury, MA: Jones and Bartlett Publisher (2007). 
Sabour S. Spinal instability neoplastic scale: methodologic issues to avoid misinterpretation AJR Am J Roentgenol 2015 ;  204 (4) : W493






© 2018 
Elsevier Masson SAS. Tous droits réservés.