Predictors of aspiration pneumonia: developing a new matrix for speech and language therapists.

ABSTRACT

INTRODUCTION: The contributing factors of aspiration pneumonia have been well documented. However, there are gaps in the literature regarding identifying the weight associated with each factor and the relationship between factors. METHOD: In this study, 20 potential predictors of aspiration pneumonia (with four additional variables) have been applied to historic Speech and Language Therapy records to greater understand the significance of each contributor of aspiration pneumonia. 152 cases with an oropharyngeal dysphagia, and a Speech and Language Therapy recommendation of eating and drinking with known aspiration and the associated potential risk of developing an aspiration pneumonia, were included in the data. These were inpatients and outpatients, and had various diagnoses but all had had a videofluoroscopy. RESULTS: Logistic regression analysis found seven factors that were individually significant in predicting the development of aspiration pneumonia with 84.93% sensitivity and 91.03% specificity DISCUSSION: Logistic regression and random forest analyses led to the proposal of a new matrix of predictors of aspiration pneumonia with respective scoring weights for individual and cumulative contributors (a direction for future research).