A machine-learning approach for decision support and risk stratification of pediatric perioperative patients based on the APRICOT dataset.
anesthesia
artificial intelligence
machine learning
pediatrics
preoperative care
Journal
Paediatric anaesthesia
ISSN: 1460-9592
Titre abrégé: Paediatr Anaesth
Pays: France
ID NLM: 9206575
Informations de publication
Date de publication:
09 2023
09 2023
Historique:
revised:
28
04
2023
received:
24
08
2022
accepted:
09
05
2023
medline:
3
8
2023
pubmed:
22
5
2023
entrez:
22
5
2023
Statut:
ppublish
Résumé
Pediatric anesthesia has evolved to a high level of patient safety, yet a small chance remains for serious perioperative complications, even in those traditionally considered at low risk. In practice, prediction of at-risk patients currently relies on the American Society of Anesthesiologists Physical Status (ASA-PS) score, despite reported inconsistencies with this method. The goal of this study was to develop predictive models that can classify children as low risk for anesthesia at the time of surgical booking and after anesthetic assessment on the procedure day. Our dataset was derived from APRICOT, a prospective observational cohort study conducted by 261 European institutions in 2014 and 2015. We included only the first procedure, ASA-PS classification I to III, and perioperative adverse events not classified as drug errors, reducing the total number of records to 30 325 with an adverse event rate of 4.43%. From this dataset, a stratified train:test split of 70:30 was used to develop predictive machine learning algorithms that could identify children in ASA-PS class I to III at low risk for severe perioperative critical events that included respiratory, cardiac, allergic, and neurological complications. Our selected models achieved accuracies of >0.9, areas under the receiver operating curve of 0.6-0.7, and negative predictive values >95%. Gradient boosting models were the best performing for both the booking phase and the day-of-surgery phase. This work demonstrates that prediction of patients at low risk of critical PAEs can be made on an individual, rather than population-based, level by using machine learning. Our approach yielded two models that accommodate wide clinical variability and, with further development, are potentially generalizable to many surgical centers.
Sections du résumé
BACKGROUND
Pediatric anesthesia has evolved to a high level of patient safety, yet a small chance remains for serious perioperative complications, even in those traditionally considered at low risk. In practice, prediction of at-risk patients currently relies on the American Society of Anesthesiologists Physical Status (ASA-PS) score, despite reported inconsistencies with this method.
AIMS
The goal of this study was to develop predictive models that can classify children as low risk for anesthesia at the time of surgical booking and after anesthetic assessment on the procedure day.
METHODS
Our dataset was derived from APRICOT, a prospective observational cohort study conducted by 261 European institutions in 2014 and 2015. We included only the first procedure, ASA-PS classification I to III, and perioperative adverse events not classified as drug errors, reducing the total number of records to 30 325 with an adverse event rate of 4.43%. From this dataset, a stratified train:test split of 70:30 was used to develop predictive machine learning algorithms that could identify children in ASA-PS class I to III at low risk for severe perioperative critical events that included respiratory, cardiac, allergic, and neurological complications.
RESULTS
Our selected models achieved accuracies of >0.9, areas under the receiver operating curve of 0.6-0.7, and negative predictive values >95%. Gradient boosting models were the best performing for both the booking phase and the day-of-surgery phase.
CONCLUSIONS
This work demonstrates that prediction of patients at low risk of critical PAEs can be made on an individual, rather than population-based, level by using machine learning. Our approach yielded two models that accommodate wide clinical variability and, with further development, are potentially generalizable to many surgical centers.
Substances chimiques
4-azidosalicylic acid-phosphatidylserine
124155-78-8
Types de publication
Observational Study
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
710-719Commentaires et corrections
Type : CommentIn
Informations de copyright
© 2023 John Wiley & Sons Ltd.
Références
Kurth CD, Tyler D, Heitmiller E, Tosone SR, Martin L, Deshpande JK. National pediatric anesthesia safety quality improvement program in the United States. Anesth Analg. 2014;119(1):112-121.
Owens WD. American Society of Anesthesiologists Physical Status Classification System is not a risk classification system. Anesthesiology. 2001;94(2):378.
Hartley B, Powell S, Bew S. Safe delivery of paediatric ENT surgery in the UK: a national strategy 2019. Accessed February 14, 2023. https://www.entuk.org/news_and_events/news/77/safe_delivery_of_paediatric_ent_surgery_in_the_uk_a_national_strategy/
Whippey A, Kostandoff G, Ma HK, Cheng J, Thabane L, Paul J. Predictors of unanticipated admission following ambulatory surgery in the pediatric population: a retrospective case-control study. Paediatr Anaesth. 2016;26(8):831-837.
Junger A, Klasen J, Benson M, et al. Factors determining length of stay of surgical day-case patients. Eur J Anaesthesiol. 2001;18(5):314-321.
Green Z, Woodman N, McLernon DJ, Engelhardt T. Incidence of paediatric unplanned day-case admissions in the UK and Ireland: a prospective multicentre observational study. Br J Anaesth. 2020;124(4):463-472.
Rabbitts JA, Groenewald CB, Moriarty JP, Flick R. Epidemiology of ambulatory anesthesia for children in the United States: 2006 and 1996. Anesth Analg. 2010;111(4):1011-1015.
Tait AR, Voepel-Lewis T, Christensen R, O'Brien LM. The STBUR questionnaire for predicting perioperative respiratory adverse events in children at risk for sleep-disordered breathing. Paediatr Anaesth. 2013;23(6):510-516.
Lee LK, Bernardo MKL, Grogan TR, Elashoff DA, Ren WHP. Perioperative respiratory adverse event risk assessment in children with upper respiratory tract infection: validation of the COLDS score. Paediatr Anaesth. 2018;28(11):1007-1014.
Nasr VG, DiNardo JA, Faraoni D. Development of a pediatric risk assessment score to predict perioperative mortality in children undergoing noncardiac surgery. Anesth Analg. 2017;124(5):1514-1519.
Valencia E, Staffa SJ, Faraoni D, DiNardo JA, Nasr VG. Prospective external validation of the pediatric risk assessment score in predicting perioperative mortality in children undergoing noncardiac surgery. Anesth Analg. 2019;129(4):1014-1020.
Habre W, Disma N, Virag K, et al. Incidence of severe critical events in paediatric anaesthesia (APRICOT): a prospective multicentre observational study in 261 hospitals in Europe. Lancet Respir Med. 2017;5(5):412-425.
Lonsdale H, Jalali A, Ahumada L, Matava C. Machine learning and artificial intelligence in pediatric research: current state, future prospects, and examples in perioperative and critical care. J Pediatr. 2020;221:S3-S10.
Lonsdale H, Gray GM, Ahumada LM, Yates HM, Varughese A, Rehman MA. The perioperative human digital twin. Anesth Analg. 2022;134(4):885-892.
Collins GS, Reitsma JB, Altman DG, Moons KG. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. J Clin Epidemiol. 2015;68(2):134-143.
Pedregosa F, Varoquaux G, Gramfort A, et al. Scikit-learn: machine learning in python. J Mach Learn Res. 2011;12:2825-2830.
Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: synthetic minority over-sampling technique. J Artif Intell Res. 2002;16:321-357.
White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30(4):377-399.
Pierpont GL, Parenti CM. Physician risk assessment and APACHE scores in cardiac care units. Clin Cardiol. 1999;22(5):366-368.
Allaudeen N, Schnipper JL, Orav EJ, Wachter RM, Vidyarthi AR. Inability of providers to predict unplanned readmissions. J Gen Intern Med. 2011;26(7):771-776.
Saklad M. Grading of patients for surgical procedures. Anesthesiology. 1941;2(3):281-284.
Haynes SR, Lawler PG. An assessment of the consistency of ASA physical status classification allocation. Anaesthesia. 1995;50(3):195-199.
Jacqueline R, Malviya S, Burke C, Reynolds PI. An assessment of interrater reliability of the ASA physical status classification in pediatric surgical patients. Pediatr Anesth. 2006;16(9):928-931.
Burgoyne LL, Smeltzer MP, Pereiras LA, Norris AL, De Armendi AJ. How well do pediatric anesthesiologists agree when assigning ASA physical status classifications to their patients? Pediatr Anesth. 2007;17(10):956-962.
Tollinche LE, Yang G, Tan K-S, Borchardt R. Interrater variability in ASA physical status assignment: an analysis in the pediatric cancer setting. J Anesth. 2018;32(2):211-218.
Nasr VG, Staffa SJ, Zurakowski D, DiNardo JA, Faraoni D. Pediatric risk stratification is improved by integrating both patient comorbidities and intrinsic surgical risk. Anesthesiology. 2019;130(6):971-980.
Aplin S, Baines D, De Lima J. Use of the ASA physical status grading system in pediatric practice. Pediatr Anesth. 2007;17(3):216-222.
Malviya S, Voepel-Lewis T, Chiravuri SD, et al. Does an objective system-based approach improve assessment of perioperative risk in children? A preliminary evaluation of the ‘NARCO’. Br J Anaesth. 2011;106(3):352-358.
Faraoni D, Vo D, Nasr VG, DiNardo JA. Development and validation of a risk stratification score for children with congenital heart disease undergoing noncardiac surgery. Anesth Analg. 2016;123(4):824-830.
Tariq S, Syed M, Martin T, Zhang X, Schmitz M. Rates of perioperative respiratory adverse events among Caucasian and African American children undergoing general anesthesia. Anesth Analg. 2018;127(1):181-187.
Burton ZA, Lewis R, Bennett T, et al. Prevalence of PErioperAtive CHildhood obesitY in children undergoing general anaesthesia in the UK: a prospective, multicentre, observational cohort study. Br J Anaesth. 2021;127(6):953-961.
Hofer IS, Burns M, Kendale S, Wanderer JP. Realistically integrating machine learning into clinical practice: a road map of opportunities, challenges, and a potential future. Anesth Analg. 2020;130(5):1115-1118.
Artificial intelligence and machine learning in software as a medical device action plan. Silverspring: U.S. Food and Drug Administration. 2019. Accessed February 14, 2023. https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device