Simplifying Universal Pediatric Pre-operative Anesthesia Screening: Analysis of NSQIP Data to Evaluate the NARCO Score as a Tool for Pediatric Anesthesia Risk Stratification.

BACKGROUND

Universal pre-operative anesthesia screening may be resource-prohibitive and burdensome to hospitals and patients. However, patients at greatest anesthetic risk cannot be identified without some basis for risk-stratification.

AIMS: This study investigated if a modified version of the Neurological, Airway, Respiratory, Cardiovascular, and Other (NARCO) pre-surgical pediatric risk score could stratify patients into high- and low-anesthetic-risk groups that would be predictive of perioperative adverse events.

METHODS

A retrospective cohort study was conducted using 132,881 records from the pediatric database of the American College of Surgeons National Surgical Quality Improvement Program (2019 ACS NSQIP-P®). Recorded comorbidities were assigned NARCO categories and presumptive NARCO scores based on non-granular data. Two methodologies were used--one in which non-granular comorbidities were considered high-risk (overestimating the high-risk group) and another in which these were considered low-risk (overestimating the low-risk group). These groups were evaluated by demographic and adverse outcome data using chi-square or Wilcoxon-Mann-Whitney analyses (p <0.05).

RESULTS

Both methodologies accurately reflected overestimation of high- or low-risk patients, when comparing NARCO to ASA scores. In both methodologies, the occurrences of 23 of 24 adverse events were significantly (p < 0.05) associated with the high-risk group.

CONCLUSION

A modified version of the NARCO pre-surgical risk assessment can stratify pediatric patients into high- and low-risk groups based on non-granular patient data. In both methodologies, this stratification identifies patients at highest risk for perioperative adverse events, which is a marker for those who require more intensive pre-surgical screening.

TYPE OF STUDY: Retrospective cohort study using de-identified information from a large.