Development of a decision tree for differential nursing diagnosis: ineffective breathing pattern & ineffective airway clearance

NANDA International 2012 Conference Abstract
Click here to view 2012 Conference Presentations (Member Login is Required)

Session 2.6
Clinical Judgement Track
Oral Poster

Viviane M. Silva
Daniel B. Resende Chaves
Marcos Venícios de Oliveira Lopes
Lívia Zulmyra C. Andrade
Lívia M. Pascoal
Karine L. Rabelo

To develop a Decision Tree (DT) for differentiating between two nursing diagnoses (ND) in children with Acute Respiratory Infection (ARI):Ineffective breathing pattern (IBP) and Ineffective airway clearance (IAC). Decision trees help nurses in the process of diagnostic differentiation? Which defining characteristics (DC) contribute to the process of diagnostic differentiation?

Results / Findings
The DT developed using the CRT method had the best predictive power. In 86.4% of the cases this DT performed correct inferences. This DT presented six terminal nodes. Three levels of depth were used to express the interactions between the DC. Four DC were relevant for the inference of IAC and IBP: Dyspnea, Ineffective cough, Alterations in depth of breathing and Adventitious breath sounds. For IAC, the sequence of DC with greatest predictive power (97,1%) was: the absence of Dyspnea associated to the presence of  Ineffective cough and Adventitious breath sounds. For IBP, the sequence of DC which presented best predictive power (28,6%) was: the presence of Dyspnea and Alterations in depth of breathing, associated to de absence of Ineffective cough.

Discussion / Conclusion
The DC Ineffective cough and Adventitious breath sounds were mentioned in other studies as representative of IAC, and also highlighted by their high prevalence. These characteristics also have high sensitivity and moderate specificity for the occurrence of IAC (Silveira, Lima, & Lopes, 2008). Similar behavior occurs with the DC related to IBP. The literature reports that the DC Dyspnea has high sensitivity and moderate specificity to predict the occurrence of IBP (Cavalcante, Mendes, Lopes, & Lima, 2010).

This is a cross-sectional study. It was evaluated 117 children with ARI aged between 0 and 5 years. The DC of the IAC and IBP were evaluated by the researcher by performing a respiratory examination. CRT (Classification and Regression Trees) algorithm, CHAID (Chi-square Automatic Interaction Detection) and QUEST (Quick, Unbiased, Efficient Statistical Tree) were the methods used to develop the DT. The DT generated were submitted to a cross-validation process. For each tree, the risk of misclassification was calculated  through tests for each of the trees in samples previously reserved. As a final result, a single DT was presented, in which the estimated risk was calculated by the average of risk of all trees together.

© NANDA International, Inc. All Rights Reserved
Quick Jump Menu