51 - The de-regionalization of neonatal surgical services across multiple US states

Publication Number: 3048.51

Scott A. Lorch, The Children's Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA, United States; Molly Passarella, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Ciaran Phibbs, Stanford University School of Medicine, Stanford, CA, United States

Background: There is growing proliferation of low-volume, high-level neonatal intensive care units in the US. There are no studies of where surgical services for newborn patients are being performed during the same time period.

Objective: To determine the number of hospitals and characteristics of hospitals performing surgery on newborn infants in 6 US states (CA, FL, MI, OR, PA, SC) from 1995-2021, and the association between condition and year with centralization of surgical services.

Design/Methods: We constructed a retrospective cohort of all births in 6 US states from 1995-2021 by linking infant birth and death certificates to maternal and infant hospital discharge records (N=15,913,012). ICD-9 and ICD-10 procedure codes identified infants who received surgery during their initial birth hospitalization excluding inguinal hernia repair (N=97,310). For each hospital-year, we calculated annual surgical volume for any surgical procedure; pediatric general surgical procedure; and organ specific surgery. Linear regression models determined the association of year, state, and hospital characteristics with the percentage of hospitals performing surgery and the centralization of surgical services as measured by the Herfindahl index (HHI), where values > 0.15 are considered centralized.

Results: Of the 7,293 hospital-years identified in the cohort, 4,129 (56.2%) performed at least 1 surgery during the calendar year. Among these hospitals, 2,533 (61.3%) performed < 10 surgeries of any kind on newborns, making up 7.1% of all surgical cases. Hospitals with lower surgical volumes were more likely to have lower numbers of births and be of lower NICU level (Table 1). The percentage of hospitals performing surgery and of low volume surgical hospitals remained statistically unchanged over time (p=0.76 and 0.22, respectively, Figure 1A) but differed by state (P < 0.001, Figure 1B). Surgical services overall and by condition were de-centralized during the study period, with HHI of 0.05-0.07 for all surgical procedures and general surgical procedures (Figure 2A), with only omphalocele from 1995-2005 and CDH after 2013 approaching the 0.15 cutoff for centralized services in condition-specific analyses (Figure 2B). State was significantly associated with variation in HHI, with adjusted values ranging from 0.03 to 0.29.

Conclusion(s): A large percentage of newborn surgical-providing hospitals perform < 10 newborn surgeries annually. Surgical care remains generally but the location and centralization of care varies widely by state, suggesting the need to understand mechanisms to optimize risk-appropriate access to surgical care.

Patient and hospital characteristics of newborn surgical performing centers
All differences statistically significant at the P<0.05 level between surgical volume groups.

Distribution of hospital surgical volumes over time (Panel A) and by state (Panel B)
Slide1.jpegPercentage of hospitals within each surgical volume group shown as a stacked bar: light blue (1-5 surgeries annually), red (6-10), green (11-25), yellow (26-50), purple (51+). Height of the bar reflects the total percentage of hospitals providing any surgical services. Year was not associated with a change in the percentage of hospitals providing surgery (p=0.76) and the percentage of low volume surgical hospitals < 10/year (p=0.24), but state was associated with both (P <0.001).

Measures of centralized care for newborn surgeries over time, all conditions (panel A) and by surgical condition (panel B)
Slide1.jpegGraphs present the Henfindahl index for all conditions (Panel A - any surgical procedures (blue) and pediatric general surgical conditions (red)) and condition specific (Panel B). The Herfindahl index (HHI) is a measure of the concentration, or centralization, of a given service. It is the sum of the squares of each hospital's market share for a given year. Values < 0.15 suggest a de-centralized, or de-regionalized market; 0.15-0.25 suggest a centralized market; and > 0.25 suggest a monopoly. Neither condition or year was associated with a change in HHI, but state remained associated with variation in HHI after adjusting for year, condition, and hospital characteristics.