Relationship Between Social Determinants of Health and Domains of Care Addressed During Pediatric Palliative Care Visits for Children with Cancer

Abstract

Highlights: What are the main findings? Although disparities exist in pediatric palliative care (PPC) by social determinants of health, domains of care addressed in PPC visits were similar across patient race/ethnicity, social deprivation index score, primary language, and patient/clinician race concordance. Some differences in subdomains emerged, with symptoms (e.g. nausea/vomiting) more likely to be addressed in Non-Hispanic White visits than other races/ethnicities. What are the implication of the main findings? Disparities in PPC and end-of-life (EOL) outcomes likely emerge from a complex interplay of variables beyond visit content, including health literacy, patient preference, historical context, and systemic factors. Clinicians can attempt to mitigate disparities by ensuring proactive symptom management and goal concordant care, and researchers can further investigate how PPC visit content influences patient outcomes. Introduction: Pediatric palliative care (PPC) improves symptom management and end-of-life (EOL) outcomes. Disparities exist in access to PPC and EOL care related to social determinants of health. Less is known regarding how the content of PPC visits varies by sociodemographic factors like race/ethnicity, socioeconomic status, and language. Methods: This retrospective cohort study included patients 0–27 years old with cancer receiving PPC between 2017 and 2022. After each PPC visit, the documenting clinician selected the domains of care addressed during the visit (Goals of Care, Symptom Management, and Care Coordination with respective subdomains). Differences in frequency of subdomains discussed were compared across patient race/ethnicity, social deprivation index (SDI) score, language, and concordance with clinician race/ethnicity. Chi-square or Fisher’s exact test assessed differences in proportions of visits with each subdomain discussed, and Kruskal–Wallis tests assessed differences in the frequency of total subdomains discussed. Results: Among 467 patients, there were 7548 PPC visits. Most patients were non-Hispanic (n = 384, 82.2%), English-speaking (n = 425, 91.0%), and identified as White (n = 270, 57.8%) or Black (n = 166, 35.5%). A median of 8 (IQR 7, 11) subdomains were addressed per visit. One more subdomain was addressed in non-Hispanic White visits (9) compared to all other races/ethnicities (8, p < 0.001). Certain topics, like symptoms (e.g., nausea/vomiting), were more likely to be addressed in visits with White and Hispanic/Latino patients. One more subdomain was addressed in the intermediate disadvantage group (9, IQR 7, 11) compared to high and low disadvantage (8, IQR 7, 11) (p = 0.092). Both English- and non-English-speaking visits addressed a median of 8 subdomains (p < 0.001). One more subdomain was addressed in patient/clinician race-discordant (9, IQR 7, 11) than race-concordant encounters (8, IQR 7, 10) (p < 0.001). Conclusions: While EOL outcomes often differ for groups of different races, ethnicities, social deprivation indices, and languages, the frequency of subdomains discussed during PPC visits was fairly similar across groups. Disparities in PPC and EOL outcomes likely emerge from a complex interplay of variables beyond visit content, including health literacy, patient preference, historical context, and systemic factors.