Predicting Postoperative Physiologic Decline After Surgery

Machine Learning-Based Systems for the Anticipation of Adverse Events After Pediatric Cardiac Surgery

Pediatric congenital heart disease (CHD) patients are at higher risk of postoperative complications and clinical deterioration either due to their underlying pathology or due to the cardiac surgery, contributing significantly to mortality, morbidity, hospital and family costs, and poor quality of life. In current clinical practice, clinical deterioration is detected, in most of the cases, when it has already occurred. Several early warning scores (EWS) have been proposed to assess children at risk of clinical deterioration using vital signs and risk indicators, in order to intervene in a timely manner to reduce the impact of deterioration and risk of death among children. However, EWS are based on measurements performed at a single time point without incorporating trends nor providing information about patient's risk trajectory. Moreover, some of these measurements rely on subjective assessment making them susceptible to different interpretations. All these limitations could explain why the implementation of EWS in high-resource settings failed to show a significant decrease in hospital mortality. By means of machine learning (ML) based algorithms we could integrate heterogeneous and complex data to predict patient's risk of deterioration. In this perspective article, we provide a brief overview of the potential of ML technologies to improve the identification of pediatric CHD patients at high-risk for clinical deterioration after cardiac surgery, and present the CORTEX traffic light, a ML-based predictive system that Sant Joan de Déu Barcelona Children's Hospital is implementing, as an illustration of the application of an ML-based risk stratification system in a relevant hospital setting.

Hospital resource use after hip reconstruction surgery in children with neurological complex chronic conditions

AIM

To assess how co-occurring conditions influence recovery after hip reconstruction surgery in children with neurological complex chronic conditions (CCCs).

METHOD

This was a retrospective analysis of 4058 children age 4 years or older with neurological CCCs who underwent hip reconstructive surgery between 1st January 2015 and 31st December 2018 in 49 children's hospitals. The presence of co-occurring chronic conditions was assessed using the Agency for Healthcare Research Chronic Condition Indicator system. Multivariable, hierarchical regression was used to assess the relationship between co-existing conditions and postoperative hospital length of stay (LOS), cost, and 30-day readmission rate.

RESULTS

The most common co-occurring conditions were digestive (60.1%) and respiratory (37.9%). As the number of co-existing conditions increased from one to four or more, median LOS increased 67% (3d [interquartile range {IQR} 2-4d] to 5d [IQR 3-8d]); median hospital cost increased 41% ($20 248 [IQR $14 921-$27 842] to $28 692 [IQR $19 236-$45 887]); and readmission rates increased 250% (5.5-13.9%), p<0.001 for all. Of all specific co-existing chronic conditions, malnutrition was associated with the greatest increase in postoperative hospital resource use.

INTERPRETATION

Co-occurring conditions, and malnutrition in particular, are a significant risk factor for prolonged, in-hospital recovery after hip reconstruction surgery in children with a neurological CCC. Further investigation is necessary to assess how improved preoperative optimization of multiple co-occurring conditions may improve postoperative outcomes and resource utilization.

WHAT THIS PAPER ADDS

Children with neurological complex chronic conditions (CCCs) often develop hip disorders which require hip reconstruction surgery. Co-occurring conditions are common in children with neuromuscular CCCs. Having four or more chronic conditions was associated with a longer length of stay, increased costs, and higher odds of readmission. Malnutrition was a significant risk factor for prolonged hospitalization after hip reconstruction surgery.

Comprehensive Risk Assessment of Morbidity in Pediatric Patients Undergoing Noncardiac Surgery: An Institutional Experience

BACKGROUND

Utilizing the intrinsic surgical risk (ISR) and the patient's chronic and acute conditions, this study aims to develop and validate a comprehensive predictive model of perioperative morbidity in children undergoing noncardiac surgery.

METHODS

Following institutional review board (IRB) approval at a tertiary care children's hospital, data for all noncardiac surgical encounters for a derivation dataset from July 2017 to December 2018 including 16,724 cases and for a validation dataset from January 2019 to December 2019 including 9043 cases were collected retrospectively. The primary outcome was a composite morbidity score defined by unplanned transfer to an intensive care unit (ICU), acute respiratory failure requiring intubation, postoperative need for noninvasive or invasive positive pressure ventilation, or cardiac arrest. Internal model validation was performed using 1000 bootstrap resamples, and external validation was performed using the 2019 validation cohort.

RESULTS

A total of 1519 surgical cases (9.1%) experienced the defined composite morbidity. Using multivariable logistic regression, the Risk Assessment of Morbidity in Pediatric Surgery (RAMPS) score was developed with very good predictive ability in the derivation cohort (area under the curve [AUC] = 0.805; 95% confidence interval [CI], 0.795-0.816), very good internal validity using 1000 bootstrap resamples (bias-corrected Nagelkerke R = 0.21 and Brier score = 0.07), and good external validity (AUC = 0.783; 95% CI, 0.770-0.797). The included variables are age <5 years, critically ill, chronic condition indicator (CCI) ≥3, significant CCI ≥2, and ISR quartile ≥3. The RAMPS score ranges from 0 to 10, with the risk of composite morbidity ranging from 1.8% to 42.7%.

CONCLUSIONS

The RAMPS score provides the ability to identify a high-risk cohort of pediatric patients using a 5-component tool, and it demonstrated good internal and external validity and generalizability. It also provides an opportunity to improve perioperative planning with the intent of improving both individual-patient outcomes and the appropriate allocation of health care resources.

Association Between Race and Ethnicity with Intraoperative Analgesic Administration and Initial Recovery Room Pain Scores in Pediatric Patients: a Single-Center Study of 21,229 Surgeries

INTRODUCTION

Perioperative pain may have deleterious effects for all patients. We aim to examine disparities in pain management for children in the perioperative period to understand whether any racial and ethnic groups are at increased risk of poor pain control.

METHODS

Medical records from children ≤ 18 years of age who underwent surgery from May 2014 to May 2018 were reviewed. The primary outcome was total intraoperative morphine equivalents. The secondary outcomes were intraoperative non-opioid analgesic administration and first conscious pain score. The exposure was race and ethnicity. The associations of race and ethnicity with outcomes of interest were modeled using linear or logistic regression, adjusted for preselected confounders and covariates. Bonferroni corrections were made for multiple comparisons.

RESULTS

A total of 21,229 anesthetics were included in analyses. In the adjusted analysis, no racial and ethnic group received significantly more or less opioids intraoperatively than non-Hispanic (NH) whites. Asians, Hispanics, and Pacific Islanders were estimated to have significantly lower odds of receiving non-opioid analgesics than NH whites: odds ratio (OR) = 0.83 (95% confidence interval (CI): 0.70, 0.97); OR = 0.84 (95% CI: 0.74, 0.97), and OR = 0.53 (95% CI: 0.33, 0.84) respectively. Asians were estimated to have significantly lower odds of reporting moderate-to-severe pain on awakening than NH whites: OR = 0.80 (95% CI: 0.66, 0.99).

CONCLUSIONS

Although children of all races and ethnicities investigated received similar total intraoperative opioid doses, some were less likely to receive non-opioid analgesics intraoperatively. Asians were less likely to report moderate-severe pain upon awakening. Further investigation may delineate how these differences lead to disparate patient outcomes and are influenced by patient, provider, and system factors.