Uroflowmetry with pelvic floor electromyography: inter-rater agreement on diagnosis of pediatric non-neurogenic voiding disorders

Abstract

Uroflowmetry with electromyography (uroflow-EMG) is commonly used for evaluation of lower urinary tract (LUT) function in children. Diagnostic criteria based largely on uroflow-EMG findings have previously been proposed for several conditions collectively termed non-neurogenic voiding disorders (NNVDs). These include dysfunctional voiding (DV), idiopathic detrusor overactivity disorder (IDOD), detrusor underutilization disorder (DUD), and primary bladder neck dysfunction (PBND). It is unknown whether practitioners with varying levels of training and experience can apply the diagnostic criteria for these conditions with a high level of consistency. To assess inter-rater agreement on diagnosis of NNVDs using uroflow-EMG studies. Six raters performed post hoc evaluation of 84 uroflow-EMG studies and associated clinical data from children with symptoms of LUT dysfunction and no evidence of neurologic or anatomic abnormalities. Each rater was asked to categorize the uroflow-EMG studies as being consistent with DV, IDOD, DUD, PBND, or normal/unclassifiable. A consensus diagnosis was noted for studies on which at least four raters agreed. Inter-rater agreement was assessed via calculation of unweighted Fleiss' kappa statistics. Overall inter-rater agreement on NNVD diagnoses was moderate (kappa 0.46, 95% CI 0.38-0.54). Agreement between individual raters ranged from 0.33 (fair) to 0.74 (substantial) (Figure). There was no consensus on diagnosis for 20 patients (24%). Several factors may contribute to inter-rater disagreement on diagnosis of NNVDs. These include instances where patients satisfy one criterion for a particular diagnosis while missing others - or have findings consistent with more than one diagnosis. Strategies to address this may involve simplifying the diagnostic criteria, developing a clear algorithm that prioritizes certain criteria, and/or allowing assignment of multiple diagnoses. Practitioners could also benefit from standardized education regarding the diagnostic criteria for NNVDs. Potential limitations of this analysis included the use of just one uroflow-EMG study per patient in almost all cases. Also, the raters had variable levels of previous experience using the diagnostic criteria for NNVDs, and it is possible that they were not always applied as originally intended. If this were the case, it would support development of a standardized education tool to facilitate practitioner understanding and application of the criteria. Uroflow-EMG has shown promise for improving clinical management of NNVDs associated with pediatric LUT dysfunction. However, inter-rater agreement on NNVD diagnoses using current criteria is suboptimal. Various mechanisms should be explored to improve consistency in practitioners' diagnosis of NNVDs.