Published in:
01-12-2014 | Original Scientific Report
Evaluation of the Open Abdomen Classification System: A Validity and Reliability Analysis
Authors:
Thordur Bjarnason, Agneta Montgomery, Stefan Acosta, Ulf Petersson
Published in:
World Journal of Surgery
|
Issue 12/2014
Login to get access
Abstract
Background
Classification of the open abdomen (OA) status is essential for clinical studies on the subject and may help to improve OA therapy. This is a validity and reliability analysis of the OA classification proposed by the World Society of the Abdominal Compartment Syndrome in 2013.
Methods
Prospective data on 111 consecutive OA patients treated with vacuum-assisted wound closure and mesh-mediated fascial traction (VAWCM) was used. For validity analysis, OA grades were compared with fascial closure and mortality. For reliability analysis, operative reports were graded by three external raters on two different occasions and the results compared. Instructions for use of the classification were constructed and studied by the external raters beforehand.
Results
The in-hospital mortality rate was 30 % (33/111). The delayed primary fascial closure rate was 89 % (85/95). Most complex grade (p = 0.033), deteriorating grade (p = 0.045), enteric leak (p = 0.001), and enteroatmospheric fistula (p = 0001) were associated with worse clinical outcomes, while initial grade, grade 1A only, contamination, fixation, and frozen abdomen were not. A floor effect was observed, with 20 % of patients receiving the lowest grade throughout OA period. Inter-rater reliability, expressed as intra-class correlation coefficient (ICC), was 0.77, 0.76, and 0.88 (95 % confidence interval 0.66–0.84, 0.65–0.84, and 0.81–0.92, respectively) and test–retest reliability 1.0, 0.99, and 0.95, respectively.
Conclusions
More complex OA grades were associated with worse clinical outcomes. However, favorable clinical results with the VAWCM technique caused many patients to receive the lowest grade, thus causing a floor effect and lower validity. Inter-rater and test–retest reliability was ‘good’ to ‘very good’.