RESUMO
BACKGROUND: In order to improve the consistency of anaesthetic risk scoring, we have developed an automated method for the calculation of ASA (cASA) scores using decision logic programming. We investigated whether ASA scoring by anaesthetic caregivers could be matched or closely approximated by a cASA. METHODS: We used a web-based preoperative assessment system to present a structured questionnaire comprising 22 questions. These were designed to score and identify conditions that are known, from the literature and expert opinion, to be risk factors. The answers from 14,349 cases were processed using decision logic to provide a variety of risk scores including a computed overall anaesthetic risk (cASA), which was then compared with the ASA score estimated by anaesthesia caregivers (eASA). RESULTS: We found a close agreement between the two measures in almost all cases. In 159 cases (1.1%), there was an underestimation of cASA, in comparison with the eASA, which appeared to be a result predominantly of incorrect or incomplete answers, or an overestimation of the ASA score by the human classifier (43%). CONCLUSION: We showed that ASA scores estimated by a heterogeneous group of anaesthesia caregivers (anaesthetists, anaesthesia trainees, and physician assistants) could be mimicked by the cASA computed by our preoperative assessment system.