A reduced rule-based expert system by the simplification of logic functions for determining the life expectancy of intensive care patients

Abstract

Purpose - Specialist physicians, specialist nurses and personnel in the intensive care unit (ICU) work to treat patients. Compared with other hospital units, ICUs are full-fledged and provide 24-h services. However, patients are admitted to this unit because of conditions such as severe trauma, organ failure or dysfunction and poisoning. The life expectancy of such patients is considerably low. Design/methodology/approach - This study developed a reduced rule-based expert system to determine life expectancy. The purpose of this study is to determine the life expectancy of patients using ten variables. The rules were determined by evaluating 210 = 1,024 results. These rules were minimized using Boolean rules for simplification. Findings - Conclusions were drawn by evaluating the minimized cases rather than evaluating 1,024 different cases. ICU patients were evaluated using this method, and the mean success rate was 63.75 per cent. Originality/value - No rule-based expert system has been designed to determine the death risks of patients in the ICU.