We found that increased community-level material deprivation, a measure of the SDOH, was associated with a longer length of stay following congenital heart surgery across a wide variety of cardiac surgeries, but not with mortality or readmission at our center. While prior works have demonstrated increased LOS for certain high-risk patient groups or specific surgeries, we found that even in patients undergoing less complex (i.e., lower STAT category) cardiac surgeries, increased community-level deprivation was associated with longer LOS.

The association of SDOH with LOS has been mixed across recent studies, such as those by Bucholz and Spigel. Bucholz et al. demonstrated that a marker of socioeconomic status was not associated with LOS following the Norwood procedure (a STAT 5 procedure) [20]; however, for a similar patient population (patients post Norwood procedure), Spigel et al. demonstrated a significant increase in LOS [21]. While there is variability in the relationship between SDOH and various clinical outcomes, a growing body of literature is demonstrating a relationship between SDOH and increased LOS, including in patients with surgical CHD across multiple STAT categories [22]. The specific underlying factors driving this relationship are not yet clear. Longer LOS is expected with more complex (i.e., higher STAT category) operations with anticipated longer recovery times and often more involved discharge coordination needs. Patients with more straightforward procedures (i.e., STAT 1 operations) typically recover more quickly and have simpler medication regimens and follow-up needs at discharge. Families with greater social needs may be able to navigate a shorter and less complex admission associated with lower risk cardiac surgeries, but the prolonged strain of an admission for higher risk surgical procedures may be overly taxing to their resources. Patients who have higher risk procedures or longer lengths of stay may have additional non-cardiology services that are needed after discharge such as physical, occupational, or speech therapy to build or regain skills that were delayed during the prolonged admission. In these patients, there is a greater requirement of families to navigate maintaining life’s demands during their child’s admission, understanding their child’s more complex medical requirements, medication regimens, and more intense outpatient care needs.

Patient discharges are delayed not only due to medical requirements, but also due to family obligations outside the hospital. These may include caregiver limitations to be present for discharge education due to their work schedules, other childcare considerations, or transportation issues. Additionally, certain patients may remain inpatient longer while strategies to minimize the risk of readmission and plans to ensure that the family will be able to follow-up for outpatient post-operative care are put in place. This often entails working with a multidisciplinary team of social workers, dieticians, pharmacists, and inpatient and outpatient nurses and cardiologists to address barriers to optimal discharge readiness, such as appropriate post-discharge supplies and equipment, or transportation to the outpatient follow-up appointment. Patients in higher STAT categories typically have more complex care needs, requiring additional coordination and investment to ensure a safe discharge and minimize the likelihood of readmission. Navigating this sometimes-taxing process may be intensified for those caregivers who face these challenges amidst negative SDOH, likely with fewer resources to absorb the significant financial and temporal investment required to care for post-operative children with complex CHD. At our institution, significant effort goes into ensuring all post-operative patients and families are medically and socially ready for discharge, even if it results in longer LOS; this may contribute to our finding that community-level deprivation was not associated with readmission in our cohort.

Prolonged LOS following CHD surgery is associated with worse patient outcomes and incurs high personal and financial stress to the family. Longer LOS after congenital heart surgery has been shown to be associated with increased morbidity, such as increased risk of hospital acquired infections [23, 24] and worse neurodevelopmental outcomes [25]. Because of the association of shorter LOS with improved patient outcomes, some advocate to include LOS as one of the quality metrics for congenital heart surgery [26]. However, factors that affect a metric like LOS are numerous and complex, with growing evidence that SDOH impact LOS independently of medical and surgical complexity. Hence, work to identify metrics and strategies that could lead to better outcomes for patients with CHD must be careful not to create unintended disparities. By accounting for SDOH and being mindful of inequity gaps, improvement interventions are more likely to improve outcomes, such as LOS, without unintentionally widening outcome differences. Further work is needed to investigate what specific components of SDOH and health-related social needs drive longer LOS and may serve as targets for future interventions to equitably decrease LOS and improve outcomes. Potential areas for improvement could include an automatic visit with a social worker for all pre-surgical patients (rather than only those who self-refer or are referred by their cardiologist) to do a detailed needs assessment and ensure common needs, such as lodging and transportation, have been addressed beyond simply getting to the hospital the day of surgery and home the day of discharge. Other programs for patients with a prenatal diagnosis could focus on maternal nutrition to positively impact both maternal and fetal outcomes. Interventions to address specific SDOH components will also need to be flexible to adapt to individual patient and family needs.

The mortality rate in our cohort was similar to the reported Society of Thoracic Surgeons mortality for the individual STAT categories during the time period of the study [4]. Prior works have shown an association between mortality and SDOH [20, 27]. Anderson et al. found that children from lower-income neighborhoods undergoing cardiac surgery had increased odds of mortality [14]. Bucholz et al. and de Loizaga et al. used census-based neighborhood scores or indices to examine neonates with single ventricle heart disease, and both demonstrated an association of SDOH markers and mortality. In our study, the DI approached but did not meet a significant association with mortality. Our results may have been influenced by sample size as mortality is a rare event, which may have made it difficult to gather sufficient patients to detect an effect in a single-center study. Unlike our study, Anderson, de Loizaga, and Bucholz were database studies with larger cohorts.

Our study has several limitations. While our single-center cohort was not small, larger multi-center or registry studies will be needed to further analyze the relationship between mortality and SDOH. There are current efforts among national registries such as the Pediatric Acute Care Cardiology Collaborative (PAC3) to examine these outcomes across multiple institutions and will hopefully provide insights into these associations [28, 29]. As with mortality, the factors that drive LOS are complex. We chose to model a linear relationship between community deprivation and LOS, but clinically there are multiple factors that influence a patient’s discharge date, and the relationship is likely more complex. We attempted to account for some of these considerations by adjusting for factors known to affect LOS, such as birth weight, age at admission, and the presence of a syndrome or chromosomal anomaly, but given the myriad of factors than can affect LOS, the adjustment is likely incomplete. Our LOS metric was captured in such a way that it did not differentiate pre- and post-operative length of stay. The LOS for patients who do not undergo surgery on the day of admission, such as neonates with prenatally diagnosed congenital heart disease, will thus be longer. We attempted to control factors that could significantly prolong the pre-operative length of stay as discussed above, but further work will be needed to investigate the impacts of socioeconomic factors on post-operative length of stay. Furthermore, the DI is generated using a patient’s permanent address and does not account for alternative discharge destinations such as to long-term care facilities or short-term bridges to home near the hospital, like the Ronald McDonald House. While these discharge plans may offer closer access to medical resources following discharge and help mitigate certain health-related social needs (such as transportation barriers), this benefit may be outweighed by effect on other needs (continued financial burden and isolation from social supports). Future work with larger data sets, analysis of the myriad of components that make up socioeconomic deprivation, and examining the impact of different discharge practices should be considered to provide greater clarity on this complex topic. Additionally, the DI is a community-level measure and therefore may not reflect the individual’s risk. Moreover, although census tracts are more specific to assess community-level deprivation, we were limited to the more heterogeneous zip codes or ZCTAs, potentially further biasing our results toward the null. Finally, as with any retrospective study there is the potential for missing data and selection bias.