Social Determinants of Health Are Associated With Length of Stay for Chronically Ventilated Children

We reduced overall and post-intensive care unit (ICU) length of stay (LOS) by implementing a standardized discharge process for children being discharged home with chronic mechanical ventilation via tracheostomy [1]. During the immediate post-implementation period of the standardized discharge process, between 2013 and 2015, overall LOS was 143 days, and post-ICU LOS was 50 days [1]. However, LOS has since increased between 2015 and 2022 as a result of both medical and social factors, particularly among children who have never been home before admission (HBA), with overall and post-ICU LOS, respectively, increased to 198 and 65 days [1]. Therefore, we explored the impact of social determinants of health (SDOH) on LOS among children being discharged home with tracheostomy and mechanical ventilation. Various SDOH indices have been developed to characterize social disparities, each using different indicators and domains. These include the Area Deprivation Index (ADI) [2], Child Opportunity Index (COI) [3], and Social Vulnerability Index (SVI) [4]. Considering the variability in SDOH indices, we aimed to determine whether ADI, SVI, and COI are associated with overall and post-ICU LOS in this high-risk patient population. The Colorado Multiple Institutional Review Board approved the Ventilator Care Program (VCP) Registry at Children's Hospital Colorado for secondary use clinical research (COMIRB #19-0643). We used the VCP Registry for this single-center retrospective analysis. Data were deidentified, and the need for informed consent was waived. Subjects were identified using the registry and were included if discharged for the first time with ventilation via tracheostomy between January 2015 and September 2023. Those who were medically ready for discharge but still admitted on September 30, 2023, were also included (with an underestimated LOS calculated using this date). Subjects were divided into two study groups: children who had never been HBA (the “not HBA” subgroup) and those who had been HBA (the “HBA” subgroup). Individuals with confidential addresses and those who transferred to other medical facilities, rather than being discharged home, were excluded. Demographic information was collected, including gestational age at birth, birth weight, sex, race, ethnicity, and spoken language. Medical history was collected, including underlying diagnoses, age at the time of admission, tracheostomy, and discharge. Information regarding SDOH, such as Department of Human Services involvement, discharge to foster care, insurance coverage, home nursing, and caregiver education, was also obtained. The hospital analytics team provided geocode data based on subjects' addresses at the time of discharge. Three SDOH indices were determined using the geocode data: the ADI, COI 2.0, and SVI. The ADI ranks neighborhoods by socioeconomic disadvantage at the national and state levels, using 17 census-derived variables [2]. The COI 2.0 combines 29 neighborhood indicators with both raw values and z-scores to provide a neighborhood ranking from lowest to highest on a scale of 1 to 100 [3]. The Centers for Disease Control's SVI uses 16 census-derived indicators in four themes (socioeconomic, household, race/ethnicity, housing/transportation). Geocode areas are scored from 0 to 1 (from least to most vulnerable) for each indicator to derive theme-specific and composite scores [4]. Basic descriptive statistics summarized outcomes of interest as frequencies (%), means (standard deviation [SD]), and medians (interquartile range: min–max) when evidence of a skewed distribution was observed via the Shapiro–Wilks test for continuous measures. Univariable linear regression models were used to assess the association between SDOH indices and LOS outcomes in both study groups. Because COI, ADI, and SVI (listed as ranked percentile or “RPL”) are composite measures with arbitrary or percentile-based scales, a 1-unit change is not inherently interpretable. To facilitate interpretability and comparability across predictors, results are reported as standardized regression coefficients (β), representing the estimated change in LOS associated with a 1 SD increase in each SDOH index. Corresponding 95% confidence intervals (CIs) were calculated to assess the precision of these estimates. Two-sided p-values < 0.05 were considered statistically significant. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC). One hundred and ninety-one subjects were discharged after tracheostomy placement with chronic ventilation during the study period. Nineteen were transferred to other facilities and thus excluded. One hundred and seventy-two remained, 120 in the not HBA group, and 52 in the HBA group (Table 1). 140 subjects were discharged in Colorado, predominantly in the Denver metro area, while 16 were discharged out of state. Geocodes were not available for 16 subjects. ADI and SVI data were available for 156 subjects, whereas COI data were only available for 119 subjects. There was no association between overall LOS and SDOH indices in not HBA subjects (Figure 1A). However, in the HBA subgroup, there was a significant association between overall LOS and state ADI as well as SVI theme 1 (socioeconomic status) and theme 3 (race/ethnicity) (Figure 1B). In not HBA subjects, post-ICU LOS was significantly associated with COI raw and z-scores, all SVI themes, the SVI composite score, and the state ADI (Figure 1C). Similarly, in HBA subjects, post-ICU LOS was associated with state ADI, SVI theme 1 and theme 3, and COI raw and z-scores (Figure 1D). This study has shown that a lower COI and higher ADI are associated with increased post-ICU LOS in patients discharged with chronic ventilation via tracheostomy. While post-ICU LOS is associated with all SVI themes and composite score in the not HBA subgroup, only SVI theme 1 (socioeconomic status) and theme 3 (race/ethnicity) were found to have an association in HBA subjects. The association with SDOH indices was reduced for overall LOS, especially in the not HBA subgroup, where no association was found. In the HBA subgroup, the state ADI, SVI theme 1 (socioeconomic status), and theme 3 (race/ethnicity) were associated with increased overall LOS. SDOH, such as the ADI, have been previously found to be associated with increased overall LOS and total admission cost [5]. To our knowledge, the COI and SVI have not been previously used to assess the impact of SDOH on LOS after tracheostomy placement; our study is the first to report all three indices in this population comprehensively. We identified that SDOH indices are more consistently associated with post-ICU LOS than overall LOS. This could be due to the significant association of clinical factors such as age and medical complexity (including seizure disorders and congenital heart disease) with overall LOS [1]. In our institution, medically stable children with a tracheostomy and mechanical ventilation are transferred out of the ICU. Thus, time spent in non-ICU settings may be reflective of the need to address non-medical barriers to a safe discharge home. Despite variations in methodology, each index attempts to identify socioeconomic opportunities [2-4], which may have a significant impact on a family's financial resources, work requirements, and access to paid leave. Caregivers of children with tracheostomies can experience significant socioeconomic stressors, which might impact their child's care [5]. In our experience, competing demands on caregivers to meet their family's financial needs while also being present at the hospital for education often impact LOS. The ability to obtain sufficient caregivers to provide continuous care for the child with chronic ventilation via tracheostomy may also be modulated by SDOH. Families with limited support may be at a disadvantage when attempting to identify additional caregivers. The ADI, COI, and SVI all attempt to capture family composition, including single-parent households [2-4]. However, the dynamic between family composition and the ability to recruit caregivers is likely more complex than captured by SDOH indices, as families often rely on extended family and friends for support. In families with limited social support, in-home private duty nursing may become essential to meet the need for continuous supervision. Nursing shortages, which vary based on discharge location, have led to prolonged LOS in this population [1]. This study presents some limitations, including the use of zip-code-based indices, which may not identify variability within these regions. Furthermore, indices may not always capture disparities that disproportionally impact children who are mechanically ventilated via tracheostomy. This includes the ability to obtain supplies from durable medical equipment suppliers, which may vary between regions and suppliers. Furthermore, the ADI, COI, and SVI do not directly assess local access to primary care, which is essential in supporting medically complex children at home [2-4]. Despite the inherent limitations of geocode-based SDOH indices, they represent a standardized and replicable measurement of risk. Additionally, state-specific considerations remain when discharging children with ventilation via tracheostomy. For example, in Colorado, there are no pediatric long-term or subacute care facilities. We identified an association between post-ICU LOS and adverse SDOH using the ADI, COI, and SVI. This is an important step toward addressing prolonged LOS after tracheostomy placement. Further research is needed to explore the underlying causes of these disparities and to identify specific modifiable risk factors that impact mechanically ventilated children with tracheostomies. Early identification of patients at greater risk for prolonged LOS could facilitate timely intervention and the creation of dedicated resources to address discharge barriers, improving equity and quality of care. Audrey Tilly-Gratton: writing – original draft, formal analysis, writing – review and editing. Jessica A. Dawson: investigation, methodology, project administration, data curation, formal analysis, writing – review and editing. Katelyn G. Enzer: data curation, formal analysis, writing – review and editing. John T. Brinton: methodology, formal analysis, writing – review and editing. Christopher D. Baker: writing – original draft, investigation, methodology, project administration, data curation, formal analysis, supervision, writing – review and editing. All authors approved the final manuscript. The authors received no specific funding for this work. The authors declare no conflicts of interest. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.