Validação clínica de árvore de decisão para diferenciação entre os diagnósticos de enfermagem padrão respiratório ineficaz e desobstrução ineficaz de vias aéreas

Abstract

This study aimed to validate clinically a decision tree (DT) developed for differentiation of two respiratory nursing diagnoses: Ineffective breathing pattern (IBP) and Ineffective airway clearing (IAC). This decision tree was developed in a previous study with 249 children with acute respiratory infection. One of the generated, the DT developed with the Classification and Regression Trees algorithm was chosen to be clinically validated by presenting better overall prediction (86.4%). Methodological study was developed from records databases that served as subsidy for construction of two dissertations. It has been estimated a sample of 222 records, being these of 111 children with acute respiratory infection and 111 with asthma. Were structured three groups of diagnosticians; standard reference group, control group and case group. The case group had access to decision tree as a tool to aid in diagnostic inference. The inferences of the comparison groups (case and control) were analyzed by taking as reference parameter the inference held by the standard reference group. The inferences of diagnosticians were tabulated in a spreadsheet of Excel software and the data were analyzed in IBM SPSS version 21.0 for Windows and statistical package R. Of the 222 records analyzed, 55.9% were male, had as main family history Asthma (57.2%), followed by tuberculosis (17.6%). IAC was diagnosed in 80.6% and IBP in 68.5% of records assessed by the standard reference group. Younger children were more often diagnosed with both diagnoses. Day care attendance and higher frequency of time were more prevalent in children without the diagnoses. The defining characteristics with statistical association with IAC were: Dyspnea (84.68%, p <0.001), Adventitious respiratory sound (84.23%, p <0.001), Ineffective cough (72.07%, p <0.001), and Orthopnea (70.72%, p <0.001). In the IBP, there was association with Dyspnea (85.13%, p <0.001) Change in respiratory depth (69.81%, p <0.001), Orthopnea (69.81%, p <0.001), Tachypnea (64.86%, p <0.001) and use of accessory muscles to breathe (66.66%, p <0.001). The diagnosticians that used the DT obtained concordance rates with the standard reference group higher than those who did not use the support tool. Larger interventions were necessary in the diagnosticians group without DT to achieve equality of inferences. This data points to the ability of the tool to make inferences more homogeneous. Best rates of sensitivity, specificity, diagnostic odds ratio, positive and negative likelihood ratios were observed for the inferences of the group with DT. Therefore, it is concluded that the tool validated in this study contributed to inferences of higher quality, thereby increasing the accuracy of diagnostic determinations. Considering the high sensitivity rate and high rates of false-positive found in this study, the DT has typical characteristics of a validated diagnostic screening tool.