Assessing the Utility of a Machine-Learning Model to Assist With the Assignment of the American Society of Anesthesiology Physical Status Classification in Pediatric Patients.

BACKGROUND: The American Society of Anesthesiologists Physical Status Classification System (ASA-PS) is used to classify patients' health before delivering an anesthetic. Assigning an ASA-PS Classification score to pediatric patients can be challenging due to the vast array of chronic conditions present in the pediatric population. The specific aims of this study were to (1) suggest an ASA-PS score for pediatric patients undergoing elective surgical procedures using machine-learning (ML) methods; and (2) assess the impact of presenting the suggested ASA-PS score to clinicians when making their final ASA-PS assignment. The intent was not to create a new ASA-PS score but to use ML methods to generate a suggested score, along with information on how the score was generated (ie, historical information on patient comorbidities) to assist clinicians when assigning their final ASA-PS score. METHODS: A retrospective analysis of 146,784 pediatric surgical encounters from January 1, 2016, to December 31, 2019, using eXtreme Gradient Boosting (XGBoost) methods to predict ASA-PS scores using patients' age, weight, and chronic conditions. SHapley Additive exPlanations (SHAP) were used to assess patient characteristics that contributed most to the predicted ASA-PS scores. The predicted ASA-PS model was presented to a prospective cohort study of 28,677 surgical encounters from December 1, 2021, to October 31, 2022. The predicted ASA-PS score was presented to the anesthesiology provider for review before entering the final ASA-PS score. The study focused on summarizing the available information for the anesthesiologist by using ML methods. The goal was to explore the potential for ML to provide assistance to anesthesiologists by highlighting potential areas of discordance between the variables that generated a given ML prediction and the physician's mental model of the patient's medical comorbidities. RESULTS: For the retrospective analysis, the distribution of predicted ASA-PS scores was 22.7% ASA-PS I, 48.5% II, 23.6% III, 5.1% IV, and 0.04% V. The distribution of clinician-assigned ASA-PS scores was 24.3% for ASA-PS I, 44.5% for ASA-PS II, 24.9% for ASA III, 6.1% for ASA-PS IV, and 0.2% for ASA-V. In the prospective analysis, the final ASA-PS score matched the initial ASA-PS 90.7% of the time and 9.3% were revised after viewing the predicted ASA-PS score. When the initial ASA-PS score and the ML ASA-PS score were discrepant, 19.5% of the cases have a final ASA-PS score which is different from the initial clinician ASA-PS score. The prevalence of multiple chronic conditions increased with ASA-PS score: 34.9% ASA-PS I, 73.2% II, 92.3% III, and 94.4% IV. CONCLUSIONS: ML derivation of predicted pediatric ASA-PS scores was successful, with a strong agreement between predicted and clinician-entered ASA-PS scores. Presentation of predicted ASA-PS scores was associated with revision in final scoring for 1-in-10 pediatric patients.

One Size Does Not Fit All: A Perspective on the American Society of Anesthesiologists Physical Status Classification for Pediatric Patients.

BACKGROUND: The American Society of Anesthesiologists physical status (ASA-PS) classification system is used worldwide to classify patients based on comorbid conditions before general anesthesia. Despite its popularity, the ASA-PS classification system has been shown to have poor interrater reliability due to its subjective definitions, especially when applied to the pediatric population. We hypothesized that the clarification of ASA-PS definitions to better reflect pediatric conditions would improve the accuracy of ASA-PS applied to this population. METHODS: A stratified, randomized sample of 120 pediatric surgical cases was collected from a tertiary-care pediatric hospital. A team of senior anesthesiologists reclassified ASA-PS within this patient sample using the suggested pediatric-specific ASA-PS definitions. Interrater reliability was measured using intraclass correlation (ICC) and Fleiss κ statistic. In addition, a qualitative study component using small focus groups of senior anesthesiologists identified areas of ambiguity within the ASA-PS system. RESULTS: Among the 90 reclassifications within each ASA-PS group, 42.2% (n = 38) of ASA-PS I were upgraded to ASA-PS II, and 36.7% (n = 33) of ASA-PS II were upgraded to ASA-PS III. In addition, 28.9% (n = 26) of ASA-PS III were upgraded to ASA-PS IV, and 24.4% (n = 22) of ASA-PS IV were downgraded to III. ICC across the reclassified ASA-PS categories was 0.77 (95% confidence interval [CI], 0.71-0.83; P < .001) demonstrating strong overall agreement. Fleiss κ statistic was lowest in ASA-PS II and III patients (κ = 0.41 and κ = 0.30, respectively) indicating lower agreement beyond chance within these subgroups. Focus groups revealed common themes such as active sequelae of disease, active versus well-controlled presence of comorbidities, and the possible inclusion of functional limitations as important considerations. CONCLUSIONS: The ASA-PS classification system has several benefits including ease-of-use, simplicity, and flexibility. However, revising the ASA-PS system to provide better guidance for pediatric patients could be valuable. While this study demonstrates good interrater reliability with the included ASA-PS pediatric definitions, further work is needed to clarify accurate assignment of ASA-PS within the midrange of the scale (ASA-PS II and III) and explore its implementation in other institutions.

Anesthesia Exposure in Children: Practitioners Respond to the 2016 FDA Drug Safety Communication.

In December 2016, the US Food and Drug Administration (FDA) issued a drug safety warning stating that 11 commonly used anesthetic and sedative medications had potential neurotoxic effects when used in children under the age of 3 years and in pregnant women during the third trimester. A panel presentation at the sixth biennial Pediatric Anesthesia Neurodevelopmental Assessment (PANDA) symposium addressed the FDA announcement in a session entitled "Anesthesia Exposure in Children During Surgical and Non-Surgical Procedures: How Do We Respond to the 2016 FDA Drug Safety Communication?" Panelists included representatives from pediatric anesthesiology, obstetrics, pediatric surgery, and several pediatric surgical subspecialties. Each panelist was asked to address the following questions: How has the FDA labelling change affected your clinical practice including patient discussions, timing, and frequency of procedures? Has your professional society provided any guidelines for this discussion? Has there been any discussion of this topic at your national meetings? The panelists provided important perspectives specific to each specialty, which generated a lively discussion and a detailed response from the Deputy Director of the Division of Anesthesia and Addiction of the FDA describing the FDA procedures that led to this drug safety warning.

Implementing a Pediatric Perioperative Surgical Home Integrated Care Coordination Pathway for Laryngeal Cleft Repair.

BACKGROUND: The Pediatric Perioperative Surgical Home (PPSH) model is an integrative care model designed to provide better patient care and value by shifting focus from the patient encounter level to the overarching surgical episode of care. So far, no PPSH model has targeted a complex airway disorder. It was hypothesized that the development of a PPSH for laryngeal cleft repair would reduce the high rates of postoperative resource utilization observed in this population. METHODS: Institutional review board approval was obtained for the purpose of data collection and analysis. A multidisciplinary team of anesthesiologists, surgeons, nursing staff, information technology specialists, and finance administrators was gathered during the PPSH development phase. Standardized perioperative (preoperative, intraoperative, and postoperative) protocols were developed, with a focus on preoperative risk stratification. Patients presenting before surgery with ≥1 predefined medical comorbidity were triaged to the intensive care unit (ICU) postoperatively, while patients without severe systemic disease were triaged to a lower-acuity floor for overnight observation. The success of the PPSH protocol was defined by quality outcome and value measurements. RESULTS: The PPSH initiative included 120 patients, and the pre-PPSH period included 115 patients who underwent laryngeal cleft repair before implementation of the new process. Patients in the pre-PPSH period were reviewed and classified as ICU candidates or lower acuity floor candidates had they presented in the post-PPSH period. Among the 79 patients in the pre-PPSH period who were identified as candidates for the lower-acuity floor transfer, 70 patients (89%) were transferred to the ICU (P < .001). Retrospective analysis concluded that 143 ICU bedded days could have been avoided in the pre-PPSH group by using PPSH risk stratification. Surgery duration (P = .034) and hospital length of stay (P = .015) were found to be slightly longer in the group of pre-PPSH observation unit candidates. Rates of 30-day unplanned readmissions to the hospital were not associated with the new PPSH initiative (P = .093). No patients in either group experienced emergent postoperative intubation or other expected complications. Total hospital costs were not lower for PPSH observation unit patients as compared to pre-PPSH observation unit candidates (difference = 8%; 95% confidence interval, -7% to 23%). CONCLUSIONS: A well-defined preoperative screening protocol for patients undergoing laryngeal cleft repair can reduce postoperative ICU utilization without affecting patient safety. Further research is needed to see if these findings are applicable to other complex airway surgeries.

Making the pediatric perioperative surgical home come to life by leveraging existing health information technology.

PURPOSE OF REVIEW: To design a patient data dashboard for the Department of Anesthesiology, Perioperative and Pain Medicine at Boston Children's Hospital that supports care integration across the healthcare system as described by the pediatric perioperative surgical home (PPSH) initiative. RECENT FINDINGS: By using 360 Technology, patient data was automatically pulled from all available Electronic Health Record sources from 2005 to the present. The PPSH dashboard described in this report provides a guide for implementation of PPSH Clinical Care Pathways. The dashboard integrates several databases to allow for visual longitudinal tracking of patient care, outcomes, and cost. The integration of electronic information provided the ability to display, compare, and analyze selected PPSH metrics in real time. By utilizing the PPSH dashboard format the use of an automated, integrated clinical, and financial health data profile for a specific patient population may improve clinicians' ability to have a comprehensive assessment of all care elements. This more global clinical thinking has the potential to produce bottom-up, evidence-based healthcare reform. SUMMARY: The experience with the PPSH dashboard provides solid evidence for the use of integrated Electronic Health Record to improve patient outcomes and decrease cost.

Perioperative Care Coordination Measurement: A Tool to Support Care Integration of Pediatric Surgical Patients.

The relationship of care coordination activities and outcomes to resource utilization and personnel costs has been evaluated for a number of pediatric medical home practices. One of the first tools designed to evaluate the activities and outcomes for pediatric care coordination is the Care Coordination Measurement Tool (CCMT). It has become widely used as an instrument for health care providers in both primary and subspecialty care settings. This tool enables the user to stratify patients based on acuity and complexity while documenting the activities and outcomes of care coordination. We tested the feasibility of adapting the CCMT to a pediatric surgical population at Boston Children's Hospital. The tool was used to assess the preoperative care coordination activities. Care coordination activities were tracked during the interval from the date the patient was scheduled for a surgical or interventional procedure through the day of the procedure. A care coordination encounter was defined as any task, whether face to face or not, supporting the development or implementation of a plan of care. Data were collected to enable analysis of 5675 care coordination encounters supporting the care provided to 3406 individual surgical cases (patients). The outcomes of care coordination, as documented by the preoperative nursing staff, included the elaboration of the care plan through patient-focused communication among specialist, facilities, perioperative team, and primary care physicians in 80.5% of cases. The average time spent on care coordination activities increased incrementally by 30 minutes with each additional care coordination encounter for a surgical case. Surgical cases with 1 care coordination encounter took an average of 35.7 minutes of preoperative care coordination, whereas those with ≥4 care coordination encounters reported an average of 121.6 minutes. We successfully adapted and implemented the CCMT for a pediatric surgical population and measured nonface-to-face, nonbillable encounters performed by perioperative nursing staff. The care coordination activities integrated into the preoperative process include elaboration of care plans and identification and remediation of discrepancies. Capturing the activities and outcomes of care coordination for preoperative care provides a framework for quality improvement and enables documentation of the value of nonface-to-face perioperative nursing encounters that comprise care coordination.

The nature and sources of variability in pediatric surgical case duration.

BACKGROUND: Case time variability confounds surgical scheduling and decreases access to limited operating room resources. Variability arises from many sources and can differ among institutions serving different populations. A rich literature has developed around case time variability in adults, but little in pediatrics. OBJECTIVE: We studied the effect of commonly used patient and procedure factors in driving case time variability in a large, free-standing, academic pediatric hospital. METHODS: We analyzed over 40 000 scheduled surgeries performed over 3 years. Using bootstrapping, we computed descriptive statistics for 249 procedures and reported variability statistics. We then used conditional inference regression trees to identify procedure and patient factors associated with pediatric case time and evaluated their predictive power by comparing prediction errors against current practice. Patient and procedure factors included patient's age and weight, medical status, surgeon identity, and ICU request indicator. RESULTS: Overall variability in pediatric case time, as reflected by standard deviation, was 30% (25.8, 34.7) of the median case time. Relative variability (coefficient of variation), was largest among short cases. For a few procedure types, the regression tree can improve prediction accuracy if extreme behavior cases are preemptively identified. However, for most procedure types, no useful predictive factors were identified and, most notably, surgeon identity was unimportant. CONCLUSIONS: Pediatric case time variability, unlike adult cases, is poorly explained by surgeon effect or other characteristics that are commonly abstracted from electronic records. This largely relates to the 'long-tailed' distribution of pediatric cases and unpredictably long cases. Surgeon-specific scheduling is therefore unnecessary and similar cases may be pooled across surgeons. Future scheduling efforts in pediatrics should focus on prospective identification of patient and procedural specifics that are associated with and predictive of long cases. Until such predictors are identified, daily management of pediatric operating rooms will require compensatory overtime, capacity buffers, schedule flexibility, and cost.

Preoperative electrocardiograms for nonsyndromic children with hand syndactyly.

PURPOSE: To examine the efficacy of preoperative electrocardiogram (EKG) screening for Timothy syndrome, a rare and fatal condition characterized by prolonged QT, in children referred for syndactyly release. METHODS: We reviewed the records of nonsyndromic syndactyly patients seen by a hand surgeon at our institution between 2007 and 2013. All underwent a preoperative screening EKG for Timothy syndrome. We reviewed the medical records for demographics, presentation, EKG results, and operative findings, and calculated median age at the time of EKG and surgery and frequency distributions for sex, side affected, EKG result, and clinical finding. The mean patient charge for EKG and interpretation was calculated. RESULTS: We identified 128 syndactyly patients, 72% of which were boys. Median age at the time of EKG testing and syndactyly release was 1 year. A total of 92% of patients had normal EKG results; one patient exhibited a prolonged QT. Ten patients (8%) had further cardiac evaluation because of the EKG result and were found to be normal on repeat testing. No patient met QT threshold for Timothy syndrome and all patients were cleared for surgery. The minimum patient charge for EKG testing was $183. CONCLUSIONS: To improve patient safety, some have advocated preoperative EKG testing for all children undergoing syndactyly release to rule out Timothy syndrome. Analysis of our experience failed to yield an instance of Timothy syndrome over a 7-year period. Although EKG charges were relatively low, costs resulting from additional testing, cardiology consultation, and provider and parent time should be considered. Our study does not support routine EKG testing for children referred for syndactyly release, and we have abandoned this practice. TYPE OF STUDY/LEVEL OF EVIDENCE: Prognostic IV.

Laryngeal cleft: evaluation and management.

OBJECTIVES: Review the latest diagnostic and treatment modalities for laryngeal and laryngotracheoesophageal clefts as they can be a major cause of respiratory and feeding morbidity in the infant and pediatric population. METHODS: Literature review of published reports. RESULTS: The presentation of laryngeal cleft usually involves respiratory symptoms, such as stridor, chronic cough, aspiration, and recurrent respiratory infections. Clefts of the larynx and trachea/esophagus can occur in isolation, as part of a syndrome (Opitz-Frias, VATER/VACTERL, Pallister Hall, CHARGE), or with other associated malformations (gastrointestinal, genitourinary, cardiac, craniofacial). This publication reviews the presenting signs/symptoms, diagnostic options, prognosis, and treatment considerations based on over a decade of experience of the senior author with laryngeal clefts. CONCLUSIONS: Type I laryngeal clefts can be managed medically or surgically depending on the degree of morbidity. Types II, III, and IV require endoscopic or open surgery to avoid chronic respiratory and feeding complications.

The surgical scorecard: a tool to close the loop on OR utilization.

OBJECTIVE: Operating room (OR) suites are among the highest cost- and highest revenue-generating areas in most hospitals. A scorecard containing utilization and performance metrics for each surgical service and surgeon was designed by the OR leadership with results sent monthly to each surgical chief. Recent trends reveal an increased focus on optimizing utilization of OR resources as part of institutional cost-analysis efforts. Protected block time into which elective surgical and procedural cases can be booked must be used appropriately and booked fully to offset the fixed costs of staffing and running the OR. DESIGN AND SETTING: The intent of the scorecard tool was to provide detailed information on utilization of protected block time for performance-improvement planning. First-case on-time start was also measured and reported so that block time at the start of the day was fully utilized. With the granular information on time-use performance of each surgeon, the surgical chiefs were able to make workflow changes to improve utilization of staffed prime-time block hours. The scorecard tool is used ultimately for communication, not calculation, of utilization metrics. MEASUREMENTS AND CONCLUSIONS: Block-time utilization was measured both before and after the implementation of the scorecard. The analysis of the period before and after implementation of the scorecard revealed an improvement in block-time utilization in all but 1 surgical service.

Laryngeal cleft repair: the anesthetic perspective.

INTRODUCTION: Laryngeal cleft is a rare congenital malformation that is being reported with increasing frequency. Diagnosis requires suspension microlaryngoscopy under general anesthesia during spontaneous respiration. Repair may be attempted by a minimally invasive endoscopic approach or open surgical repair. The authors report on their experience with total intravenous anesthesia (TIVA) and spontaneous ventilation without an endotracheal tube during suspension laryngoscopy and CO2 laser application for this specific surgical procedure. Of particular interest were the rate at which this technique failed and rescue techniques were employed and the ability to predict patients in whom this might occur. METHODS: Between July 2004 and September 2012, 110 endoscopic laryngeal cleft repairs were completed under TIVA with spontaneous ventilation without an endotracheal tube. Anesthetic induction was achieved by inhalation of sevoflurane and oxygen by mask or infusion of propofol at 300 mcg kg(-1) min(-1) and remifentanil at 0.05-1.0 mcg kg(-1)  min. The vocal cords and surgical site were sprayed with up to 2 mg kg(-1) of 4% lidocaine. If the oxygen saturation decreased during the procedure or the patient became apneic, a rescue process utilizing jet ventilation or intermittent intubation was instituted. RESULTS: Ten (9.1%) of the 110 cases required rescue (95% confidence interval [CI]: 5.0-15.8%). The most prevalent comorbidities included reactive airway disease, chronic lung disease, failure to thrive, developmental delay, and an unrelated syndrome. Thirty-nine patients (36%) had reactive airway disease and twelve (11%) had chronic lung disease. Intraoperative complications included six cases requiring a brief, temporary period of intubation (5.5%) and four cases requiring a brief period of jet ventilation (3.6%). CONCLUSION: The technique of TIVA with spontaneous respirations without an endotracheal tube is a safe and effective technique for laryngeal cleft repair. Although the potential for intraoperative adverse events may be high, the actual rate was very low. The need to convert to other techniques is not significant although the children who did require brief periods of jet ventilation or intubation tended to have reactive airway disease or chronic lung disease.

Criteria for assessing operating room utilization in a free-standing children's hospital.

BACKGROUND: The staffed hours of operation in any surgical facility are a valuable institutional resource. The realistic target for the utilization of this resource is dependent on many factors including scheduling, efficiency, and culture of the facility. There is no previously reported measure for the actual utilization of staffed regular operating room (OR) hours in an academic pediatric institution. The leadership of the perioperative services at Children's Hospital Boston (CHB) sought to define the utilization of surgical suite staffed block time hours at that institution and in addition determine whether changes in workflow could increase the measured utilization. METHODS: Operating room efficiency in fiscal year 2009 was measured using two variables: utilization and turnover measured in hours for each month in fiscal year 2009, recorded in hours expressed as ratios (observed/expected) and as differences (observed - expected). A total of 27,851 cases from October 1, 2008, through September 30, 2009, were analyzed. All elective cases were scheduled electronically following institutional guidelines; urgent or emergent procedures were scheduled into vacant time slots on the day of the procedure. Time series analysis based on a generalized autoregressive moving average process was used to compare expected with observed utilization and to evaluate changes in utilization and turnover ratios. RESULTS: Efficiency as measured by capped utilization divided by total available time in the OR averaged 79%. Utilization ratios ranged from a low of 73% in February 2009 to a high of 87% in July 2009. An improvement in on-time first-case starts may have contributed to the increase in the utilization of staffed block time. Turnover time as defined by turnover ratio decreased significantly over time, indicating an improved efficiency in the OR starting in April 2009. CONCLUSIONS: Adhering to the specific guidelines that are followed at CHB, the mean utilization of scheduled block time was 79%. This was achieved by maximizing workflow in the surgical, anesthesia, and nursing disciplines to shorten turnover time, fill gaps in the elective schedule with emergency procedures, and provide staffing to accommodate cases that extend beyond the scheduled staffed time prior to the reporting period. Simulated models from other pediatric institutions suggest that the optimal utilization of designated time periods in a surgical facility may range from 85% to 90%.

Endoscopic repair of laryngeal cleft type I and type II: when and why?

OBJECTIVES/HYPOTHESIS: To evaluate the clinical features of children with type I and type II laryngeal cleft and the role of conservative monitoring versus endoscopic repair in their management. METHODS: Clinical presentation and evaluation; findings at the time of laryngoscopy, bronchoscopy, and esophagoscopy; and efficacy and outcome of conservative monitoring and endoscopic CO(2) laser repair. RESULTS: Eighty-one patients were evaluated for aspiration. Seventy-four patients were diagnosed as having a clinically significant laryngeal cleft. Thirty-two patients (14 males, 18 females) were monitored conservatively. Forty-nine patients (26 males, 23 females) required surgical intervention due to failed medical and feeding therapy of aspiration related to their laryngeal clefts (28 type I, 21 type II). Endoscopic CO(2) laser repair was used in all these patients. CONCLUSIONS: Medical and feeding therapy should be the first modality of treatment in patients with laryngeal cleft type I and type II. Factors supporting surgical repair include: 1) clinically apparent aspiration with feeding, 2) severity of pulmonary status, 3) findings on modified barium swallow and chest x-ray, 4) absence of significant comorbid conditions predisposing to aspiration, 5) findings on upper aerodigestive endoscopy, and 6) poor response to medical management and feeding therapy.