Designing a Decision Support System to Diagnose Neonatal Clinical PICC Infection Using Fuzzy Logic Designing a Decision Support System to Diagnose Neonatal Clinical PICC Infection Using Fuzzy Logic

Abstract

Introduction: Peripherally inserted central catheters (PICC) have entered neonatal intensive care units (NICU) as an instrument to reach blood vessels [1]. Compared with central and peripheral venous catheters, PICCs have considerably reduced the side effects and complications [2-4]. The same instruments can be the cause of catheter-related bloodstream infection (CRBSI)[5]. The purpose of this study is to create a fuzzy expert system for the early diagnosis of catheter infection in newborns. Materials and Methods: Factors effective in infection diagnosis were determined by a questionnaire and based on pediatric subspecialists’ comments. The system was designed bilingually (Persian and English) using C# software and SQL Server database. The output of the system is the percentage of infection risk. The system was assessed by the data of newborns’ files in one of the hospitals in Tehran. Results: Following the assessment, the sensitivity of the system turned out to be 95% and its Specificity and accuracy were 88 and 91 percent, respectively. Conclusion: Non-specificity of clinical signs and laboratory findings of newborns’ blood infection made its diagnosis difficult and uncertain. Using the designed expert system can be effective in the diagnosis of catheter-related blood infection.