P11

REDUCING ON THE DAY CANCELLATIONS DUE TO RESPIRATORY ILLNESS FOR ELECTIVE SURGERY: HOW MACHINE LEARNING CAN BE USED TO CHANGE CLINICAL PRACTICE

M. Devlin, Royal Manchester Children's Hospital, UK

Introduction

The advent of data-driven research paradigm has many exciting applications in health care.  In our tertiary paediatric centre a hybrid approach leveraging big data and classical risk prediction modelling had identified risk factors for perioperative respiratory adverse events (PRAE) [1,2].  Globally, respiratory illnesses are the leading cause of on the day cancellations of elective procedures [3,4].  These insights, along with the wider literature were used to drive a quality improvement initiative with the aim to enhance patient safety and maximise theatre efficiency.

Methods

Utilising a data-driven model alongside an extensive literature review, we identified factors that present an elevated risk for PRAE within our local paediatric population.  We developed a 48-hour pre-surgery screening tool aimed at identifying patients for whom surgery should be deferred for optimisation, complemented by evidence-based best practice and a decision aid for anaesthetists to utilise on the day of surgery. This initiative was informed by a comprehensive gap analysis via a survey distributed among all anaesthetists, assessing current practices, decision-making processes, and self-rated confidence levels. This comprised a mix of multiple-choice questions, free-text responses, and self-assessment scores. The implementation of care pathways was introduced at an educational event for the entire department discussing and appraising the anaesthetic management of children at high risk of PRAE.  This was subsequently followed by a second survey to evaluate the intervention's effectiveness.  Strategies adopted included pre-procedural screening to identify high-risk patients for pre-optimisation and rehabilitation, adherence to evidence-based best practice, and extensive education and dissemination within the anaesthetic department.

Results

The initial survey garnered 83% response rate (n=30) across all departmental levels. This survey delineated considerable variability in clinical practice, often diverging from established best practice, and revealed low self-rated confidence among anaesthetists. Following an educational event, which was well-received and saw active participation, a subsequent survey demonstrated a significant shift towards uniformity in patient management.  Anaesthetic management are now more closely aligned to best practice, accompanied by a marked improvement in self-reported confidence levels among the staff.

Discussion

This quality improvement initiative underscores the efficacy of leveraging a data-driven approach and targeted survey feedback to refine and develop clinical guidelines and patient flow pathways. By aligning computational methodologies with practical clinical needs, the project exemplifies the dynamic potential of modern data analysis in enhancing patient care and operational efficiency within a paediatric setting. It further highlights the essential role of continuous education, guideline dissemination, and the adoption of a feedback-driven model in effecting sustainable improvements. The ongoing monitoring of on-the-day cancellations, now systematically tracked, serves as a vital performance metric, indicating the initiative's success in reducing elective surgery cancellations due to intercurrent illnesses.  This ensures high levels of patient safety.

References:

1. Kenth J, Martin G, van Staa T, Abdelrahman M. Incidence and Risk Factors of Perioperative Respiratory Adverse Events in Children Undergoing Elective Surgery Employing Logistic Regression and Machine Learning Cluster Analysis. Anesthesia & Analgesia. 2022 May;134(5S):863–5.
2. Kenth J, van Staa T, Abdelrahman M. A Data-Driven Approach to Ascertain Respiratory Adverse Events in Children, during the Early Anaesthesia Recovery Period. Anesthesia & Analgesia. 2021 Sep;133(3):1244–6.
3. Subramanyam R, Yeramaneni S, Hossain MM, Anneken AM, Varughese AM. Perioperative Respiratory Adverse Events in Pediatric Ambulatory Anesthesia: Development and Validation of a Risk Prediction Tool. Anesth Analg. 2016 May;122(5):1578–85.
4. von Ungern-Sternberg BS, Ramgolam A, Hall GL, Sly PD, Habre W. Peri-operative adverse respiratory events in children. Anaesthesia. 2015 Apr;70(4):440–4.