Empirical antibiotic therapy (ABT) of lower respiratory tract infections (LRTI) in the elderly: Application of artificial neural network (ANN). Preliminary results

摘要: LRTI are among the most common diseases in developed countries, including chronic obstructive pulmonary disease (COPD), one of frequent conditions. Their treatment general practice is often unsuccessful and this increases hospital admissions. We know, bacterial infections elderly show a higher morbidity mortality, either for more severe symptoms, than younger adults, or because causing agent remains unknown. The need quick initiation ABT requires to chose on empirical grounds. To date there no official guidelines COPD exacerbations, but only heterogeneous conflicting recommendations exist. aim our study was identify tool guide choice effective when symptoms acute bacteriological tests cannot be performed. used an ANN 117 patients aged between 55 97 years (mean 81.5 ± 8.7 years) (± S.D.), admitted with diagnosis pneumonia, exacerbation pneumonia respiratory failure. registered at onset some individual variables such as age, sex, risk factors, comorbidity, current drug therapies. Then applied choose 20 versus subjects whose therapy chosen by physicians, comparing these groups therapy's efficacy, mean durations hospitalization (H). In learning phase, could predict resolution index 99.05% time (i.e., 104 times) S.D. = 0.23. After training, during test network predicted 91.67% 11 0.54, thus proving validity relations identified phase. Preliminary results application tool, allowed us greatly reduce duration subsequently H. Based preliminary results, we assume that use can make valuable contribution course lung elderly.