Discussion

We describe pediatric surgeries and SSI rates for 6 common surgical procedures in a large cohort of US hospitals from 2015 to 2018. SSI rates for colon, small bowel, and spinal-fusion surgeries were relatively stable or showed a modest increase in relationship to surgical volumes, which differs from the expected higher volume–better outcomes relationship. ^{Reference Calderwood, Kleinman, Huang, Murphy, Yokoe and Platt4} In contrast, appendix and ventricular-shunt surgeries appeared to have lower SSI rates with higher surgical volumes. Our findings are consistent with a prior systematic review that demonstrated an association between higher volume and improved outcomes for appendicitis and ventricular-shunt surgeries in children, as well as mixed findings for procedures that are less common, more resource intensive or more likely to occur in tertiary-care centers. ^{Reference McAteer, LaRiviere, Drugas, Abdullah, Oldham and Goldin5} Our data reflect the challenges with measuring quality in pediatrics; surgical volume and procedural complexity are often correlated and current metrics do not adequately capture this complexity.

The highest SSI rates were observed following colon surgery, which parallels the extant adult SSI literature on colorectal procedures associated with high SSI rates and substantial variability (3%–30%). ^{Reference Smith, Bohl and McElearney6} Comparable pediatric data using NHSN definitions are scarce. However, our data are similar to those reported by the Pediatric National Surgical Quality Improvement Program (NSQIP-P), which described an overall SSI rate for colorectal procedures of 5.9% with higher rates (11.4%) for total abdominal colectomy in a cohort of 50 hospitals. ^{Reference Feng, Sidhwa, Cameron, Glass and Rangel7} Unlike our experience with other pediatric HAIs in a similar study population, we did not observe a consistent downward trend in pediatric SSI rates over the study period. ^{Reference Hsu, Mathew and Wang8,Reference Patrick, Kawai and Kleinman9} We also did not see improvements in SIRs for any of the procedures (except gallbladder surgery) over the 4-year period of this study (Supplementary Material 3 online). The high SIRs noted in this study may be due to the differences in patient and surgical characteristics in our cohort compared to the national aggregate. Interpreting these findings is difficult, given challenges in appropriate risk adjustment for pediatric SSI models (Supplementary Material 1 online). For example, the current SIR model for ventricular shunt surgery only includes age, and the SIR model for spinal fusion does not include surgical indications such as neuromuscular versus idiopathic scoliosis. ^{Reference Matsumoto, Larson and Warren10} Assessing quality of care in surgical patients in pediatrics and benchmarking may require adjustment for common pediatric comorbidities that have an impact on outcomes.

Our study has several limitations. We did not have granular data on individual risk factors, which may have further informed the differences in rates across varied settings. Participating hospitals tended to be larger, metropolitan, academic institutions and may not fully represent the spectrum of pediatric care in the United States. In addition, accurate comparisons are challenging for pediatric procedures because the risk adjustment models used by NHSN for children are often limited. We did not include superficial SSIs; however, they carry lower morbidity and mortality, and they are currently excluded from the CMS benchmarking for adult SSIs. Also, we did not validate the surveillance approaches used by various healthcare facilities, which may have affected surveillance and reporting.

This study is one of the largest on pediatric SSI rates reported to the NHSN; further research on pediatric SSI measurement and prevention efforts is needed. ^{Reference Schaffzin, Harte and Marquette11} Harmonization of criteria across different entities (NHSN, NSQIP, etc) to allow for reduced reporting burden and capture of key improvement opportunities should also be addressed. Benchmarking targets may be challenging; thus, an initial focus on common procedures such as appendix or colon surgery is warranted. Continued efforts are needed to refine and validate risk adjustment models and to promote actionable improvements in care.