An Examination of Social Determinants of Health and their Association with Preventable and Urgent Readmissions in Pediatric Acute Care Cardiology

Abstract

Hospital readmissions are costly, often preventable, and associated with both medical and socioeconomic factors. The primary objective of this study was to evaluate the relationship between social determinants of health (SDOH) and readmission to a pediatric acute care cardiology unit. Secondary objectives included assessing the incidence of preventable or urgent readmissions and suggesting interventions to address common causes of these readmissions. This single-center, retrospective analysis reviewed readmissions to the acute care cardiology unit at Cincinnati Children’s Hospital within 7 days of discharge from 2019 to 2022. A preventability score and urgency metrics were assigned to each readmission, and multivariable logistic regression was used to obtain odds ratios and 95% confidence intervals according to individual and community-level measures of SDOH. Out of 265 readmission encounters, 27 (10%) were preventable and 74 (28%) were urgent. Readmission length of stay was significantly longer for preventable or urgent readmissions. Birth weight, gestational age, English-speaking status, race/ethnicity, sex, insurance category, deprivation index, and single ventricle status were not significantly associated with preventable or urgent readmissions. The most common reason identified for preventable readmission was insufficient discharge education, and the most common reason identified for urgent readmission was advancement of chronic disease. Measured SDOH were not associated with preventable or urgent readmissions, which may be partially due to our institution’s multidisciplinary approach to address SDOH. However, as insufficient discharge education was a common reason for preventable readmission, clearly more work is needed to delineate what specific strategies are most effective for mitigating adverse SDOH.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00246-025-03879-5.

Introduction

Pediatric hospital readmissions are common and quite costly to the healthcare system, estimated to cost roughly $1.5 billion annually [1]. Though readmission rates are decreasing, approximately 5–7% of pediatric patients are readmitted within 30 days of index admission [1]. Readmissions within pediatric cardiology are even more common, with 30-day readmission rates ranging from 11 to 20% in the literature [2–6]. Evaluating preventability of readmissions is challenging [7, 8], but one study in general pediatrics indicates up to 30% of admissions may be preventable [9]. Both medical and socioeconomic factors are associated with readmission after congenital heart surgery [6, 10].

Many studies have established the relationship between social determinants of health (SDOH) and clinical outcomes in pediatric cardiology. Poverty, housing instability, parental educational attainment, minority status, food insecurity, and transportation barriers have been associated with higher mortality, less frequent prenatal diagnosis of CHD, higher incidence and prevalence of CHD, adverse post-surgical outcomes, impaired neurodevelopmental outcomes, and poorer quality of life [11–14]. Further studies have noted the association between community-level characteristics and clinical outcomes, finding that community deprivation metrics are associated with mortality, length of hospitalization, and cost, even after controlling for individual characteristics [15–19].

While a clear relationship exists between SDOH and clinical outcomes in pediatric cardiology, including mortality, prevalence of congenital heart disease, and post-surgical outcomes, the impact of SDOH on readmissions remains less explored. Existing studies primarily focus on medical factors associated with readmission after congenital heart surgery, with some acknowledging a link between ethnicity, socioeconomic status, and insurance status and readmission rates [3, 4, 6, 10, 20–26]. However, these investigations often overlook the complex interplay between SDOH and clinical risk factors, such as younger age, lower body weight, feeding difficulties, comorbid conditions, and complex cardiac anatomy, which may synergistically influence readmission outcomes [3, 10, 27, 28]. Although some studies in general pediatrics show reduced readmission rates when SDOH are addressed [24, 25], higher readmission rates persist in children with complex CHD, despite the use of discharge practices similar to those employed in these studies.

The primary objective of this study was to evaluate the relationship between SDOH and readmission within 7 days to an acute care cardiology unit. Secondary objectives included assessing the incidence of preventable or urgent readmissions and identifying common causes of these readmissions, thereby allowing for identification of potential interventions to address these risk factors.

Methods

This single-center, retrospective analysis reviewed pediatric patients readmitted to a cardiac unit at Cincinnati Children’s Hospital Medical Center within 7 days of discharge from the acute care cardiology unit from 2019 to 2022. The acute care cardiology unit is a step-down unit designed to provide care for all patients with cardiac disease who do not require ICU-level care. Patients of any age, with medical and surgical conditions, are managed on this unit by the cardiology team. Readmissions that were planned or expected at the time of discharge such as additional testing or procedures scheduled at a future date were identified and removed from the in-depth analysis of this paper, as they were felt to be in a different classification of readmission type, preventability and urgency. The 7-day timeframe was selected based on our initial unpublished pilot study, which found that the majority (68%) of preventable readmissions occurred within the first 7 days after hospital discharge. While many prior studies have examined 30-day readmission rates, the investigators chose to examine readmissions within 7 days; this narrow window was chosen as it was felt it would have a greater probability of identifying factors which could potentially have been addressed by the admitting care team, thereby allowing for future practice changes to hopefully decrease avoidable readmissions. As part of a standardized institutional quality improvement process, a multidisciplinary team at our institution convenes monthly to review all readmissions that occurred in the preceding month. This ongoing review process, conducted closer to real-time, minimizes the impact of recall bias and provides detailed assessments of each case. This study utilized the results of these standardized, ongoing reviews, which included assigning a preventability score and metrics of urgency at the time of readmission to each case. A previously described 5-point Likert scale preventability score [8] was used to describe how likely a readmission was due to preventable factors, where 1 represented “Readmission not preventable under most circumstances” and 5 represented “Readmission preventable in most circumstances” (Online Resource 1). Readmissions were classified as preventable if the reviewing team rated them as either"more likely preventable"or"preventable in most circumstances."For instance, a potential non-preventable cause of readmission could be the acquisition of a new viral infection, whereas a preventable cause might involve non-adherence to the prescribed medical plan at home. Urgency was defined based on level of intervention needed at the time of readmission. Readmissions were considered urgent if any of the following were required within 24 h of readmission: ICU admission, intubation, inotropic support, unplanned interventional procedure, resuscitation with 40 mL/kg or more of fluids (including blood products), code or rapid response called, or death. Data were collected regarding each patient’s diagnoses and demographics.

A community-level deprivation index developed at our institution was derived according to each patient’s address. This was measured at the zip code level to encapsulate community social and economic context and approximate individual-level SDOH. The zip code for each subject was geocoded to a US Census Bureau defined zip code tabulation area, which are representations of the United States Postal Service zip code service areas. This zip code tabulation area was then used in conjunction with data from the 2018 US Census’ American Community Survey to generate a deprivation index for each patient [29]. The deprivation index incorporates six variables associated with material deprivation and ranges from zero to one, with one reflecting the greatest level of community material deprivation. These variables include measures of poverty, income, education, assistance, housing, and insurance, and the national median deprivation index is 0.33 [29]. Geocoding and deprivation index derivation were completed at our institution using DeGAUSS.

Statistical analysis

Medians with interquartile ranges (IQR) were summarized for continuous variables and compared between groups using the Kruskal–Wallis test. Frequencies with percentages were summarized for categorical variables and compared between groups using the Chi-square test. Multivariable logistic regression was used to assess factors potentially associated with preventable or urgent readmissions, including race, sex, insurance category, index length of stay, deprivation index, ventricle morphology, and presence of CHD. Multivariable logistic regression was used to obtain odds ratios and 95% confidence intervals for readmission status according to individual and community-level measures of SDOH. A p < 0.05 was considered significant. Statistical analyses were performed using SAS 9.4.

Results

Over the study time period, there were 4100 total index admissions to the acute care cardiology unit. 197 patients were readmitted to a cardiac unit within 7 days of discharge from index admission, and 43 of these patients (21.8%) experienced more than one readmission for a total of 265 readmission encounters. Patient demographics are presented in Table 1.

Table 1.

Patient demographic and clinical characteristics. N = 197 patients

Age, years/months (median; Q1, Q3)	0.75/9 (0.15/2, 10.4/125)
Birth weight, kg (median; Q1, Q3)	3.06 (2.62, 3.44)
Sex, female (n, %)	74 (37.6)
Race/ethnicity (n, %)
Non-hispanic white	151 (76.6)
Non-hispanic black	28 (14.2)
Non-hispanic other	10 (5.1)
Hispanic	8 (4.1)
English speaking (n, %)	191 (97.0)
CHD diagnosis (n, %)
CHD diagnosis	148 (75.1)
No CHD diagnosis	25 (12.7)
Cardiac transplant	24 (12.2)
Single ventricle (n, %)	81 (41.1)
Deprivation Index (median; Q1, Q3)	0.31 (0.25, 0.41)
Insurance type (n, %)
Public	142 (72.1%)
Private	55 (27.9%)

In univariate analysis, chronological age, birth weight, gestational age, English-speaking status, race/ethnicity, sex, insurance category, index admission length of stay, deprivation index, single ventricle status, and presence of CHD diagnosis were not significantly associated with preventable or urgent readmission (p > 0.05, Online Resources 2 and 3). In multivariable analysis, there was no significant association between preventable or urgent readmission and deprivation index, single ventricle status, or presence of CHD diagnosis after controlling for race/ethnicity, sex, and insurance category (p > 0.05). Index admission length of stay was significantly longer for preventable readmissions (10 days versus 6 days) only after controlling for race/ethnicity, sex, and insurance category (OR 1.009, 95% CI 1.001–1.018, p < 0.05). Index admission length of stay was not significantly associated with urgent readmission, including after controlling for race/ethnicity, sex, and insurance category (p > 0.05, Online Resources 4 and 5).

Reasons for readmission were categorized by the review team according to standard definitions adapted for pediatrics from the American Case Management Association Compare Readmissions Benchmarking tool [30] and are shown in Fig. 1, with the most common reason being uncontrolled advancement of chronic disease (felt to be likely related to underlying cardiac physiology). Other medical reasons for readmission included: medical plan of care – treatment or complications, medical plan of care – progression of acute disease (a treatable or temporary condition that progressed further after discharge), unrelated (non-cardiac) readmission, or medication management. Readmissions that were felt to be secondary to failures of the discharge or transition plan were categorized as such. Readmissions felt to be related to patient or caregiver deviation from the care plan as prescribed, for any reason, were categorized as “Patient Activation.”

Fig. 1.

Reasons for readmission by frequency

Twenty-seven readmissions (10%) were categorized as preventable, and 74 (28%) as urgent. Preventable admissions were more likely to be associated with patient activation or discharge/transitional care plan (p < 0.0001). Only 31.1% (n = 23) of urgent readmissions required ICU-level interventions (intubation, inotropic support, unplanned procedure, fluid resuscitation, code or rapid response). The remainder were categorized as urgent only because the patient was admitted to the ICU, which is a common result of the triage process for high-risk patients. There were no deaths within 7 days of index admission discharge, and one death during hospital readmission.

Readmission length of stay was significantly longer for preventable readmissions compared to non-preventable readmissions (9 days versus 4 days, p = 0.003) and occurred closer to initial discharge date (3 days after discharge versus 4 days, p = 0.021). Similarly, urgent readmissions were associated with a longer readmission length of stay (7.5 days versus 4 days, p = 0.002). Readmissions that were classified as both preventable and urgent were significantly longer than those admissions that were neither preventable nor urgent (24.5 days versus 4 days, p = 0.0002) and occurred closer to initial discharge date (2.5 days versus 3.5 days, p = 0.027).

Details of each preventable or urgent readmission were reviewed for identifiable causes of readmission, particularly modifiable reasons. The most common reason identified for preventable readmission was insufficient discharge education, followed by challenges in accurately anticipating disease trajectory, inadequate outpatient follow-up or communication in the first week following discharge, non-adherence to the medical plan, and a presumed hospital-acquired illness (Fig. 2). The most common reason identified for urgent readmission was advancement of chronic disease, followed by complication related to medical care, progression of acute disease, insufficient discharge education, inadequate outpatient follow-up, non-adherence to the medical plan, and challenges in accurately anticipating disease trajectory (Fig. 3). More than one reason was identified for 37% of urgent readmissions and 15% of preventable readmissions, and all identified reasons are included in Figs. 2 and 3.

Fig. 2.

Pareto chart of reasons for preventable readmissions

Fig. 3.

Pareto chart of reasons for urgent readmissions

Details of the six readmissions deemed both preventable and urgent were analyzed further, as these are the cases of primary interest for internal prevention efforts (Fig. 4). Examples of these cases include viral respiratory illnesses with persistent symptoms after discharge, feeding intolerances unresponsive to initial management strategies, vaccine-preventable illnesses arising due to incomplete understanding of discharge education recommendations, and post-operative lesions whose severity was not fully recognized until they became symptomatic after discharge.

Fig. 4.

Preventable and urgent readmissions

Discussion

Although SDOH are known to be associated with poorer outcomes in patients with CHD, we surprisingly did not find a significant association between measures of SDOH and preventable or urgent readmissions to our institution’s acute care cardiology service. This may be attributable to the relative homogeneity of our patient population (predominantly white and English speaking), the relatively high proportion of outside referrals and patients with single ventricle physiology at our institution, and a median deprivation index slightly lower than the national median (0.31 versus 0.33) [29]. However, these findings may also be related to the use of several strategies at our institution to mitigate adverse SDOH, including routine screening, inpatient support, education, and outpatient follow-up. As many of these support practices are similar to those implemented at other large institutions, alone they are unlikely to fully account for the lack of association between SDOH and readmissions in this study. However, we briefly outline several of our institution’s strategies below and discuss potential additional approaches.

Mitigation of SDOH

An inpatient social work assessment is completed for families and identified needs are addressed in the form of meal cards, transportation and hotel vouchers, and community referrals. Care managers are present on daily rounds and function to coordinate discharge needs and outpatient follow-up, including connecting families with our institution’s financial advocates if indicated. A discharge checklist is visible in EPIC to all care providers as a visual reminder of the barriers to discharge and remaining education needs. Vulnerable families who have struggled with education and certain at-risk populations (single ventricle, heart transplant, tracheostomy, home oxygen, or continuous feeds) typically receive higher-level or repeat education. Pre-discharge family education is completed by bedside nursing, specialty nursing, cardiothoracic surgery (if wound care required), and cardiology providers. Multidisciplinary care rounds, attended by social work, care management, cardiology attendings, and frontline providers, are held weekly to address families’ social and emotional concerns and home care needs.

Outpatient primary care providers and cardiologists are notified by phone for notable admissions, and discharge summaries are routed to these providers for all discharges. Upon discharge, families are provided a telephone number to reach a 24/7 on-call physician for urgent questions. Forty-eight to 72 h post-discharge, post-operative and newborn patients receive a discharge phone call from a cardiology nurse to assess their transition to home and any potential concerns.

Additional considerations

One consideration is improving the process of assessing a family’s comprehension and preparedness for discharge. This may include discharge readiness scores, automated scoring systems to identify patients at high risk of readmission, standardized discharge checklists, and “teach back” strategies to identify areas needing clarification [24, 31–33].

It was also observed in our case review that in several cases of preventable readmissions, written discharge instructions lacked complete contact information or contained unclear feeding instructions and troubleshooting tips. Providing discharge instructions both verbally and in writing, and in the family’s preferred language, is essential. For instance, at our institution, a centralized translation service ensures that translated discharge instructions are delivered within 1 to 2 days of submission.

Future considerations include working with families through interviews or focus groups to obtain a more nuanced understanding of factors that contribute to unplanned readmissions. For example, a study involving parents of children with congenital heart disease post-discharge revealed a strong desire for visual aids to facilitate discharge instruction, potentially in the form of a mobile application [34]. While there is clear room for improvement in the discharge process, we recognize that both the clinical team and the family may not fully identify the shortcomings in patient care until the family experiences challenges at home. Consequently, not all readmissions should necessarily be viewed negatively.

Improving Assessment of Diagnosis and Disease Progression

Advancement of chronic disease and challenges in accurately anticipating disease trajectory represent significant secondary factors contributing to preventable and urgent readmissions. Striking a balance between over-testing/over-treatment and under-testing/under-treatment has long been a central objective in clinical practice. This is reflected in initiatives such as the “Choosing Wisely” campaign, which has sought to foster national dialogue aimed at reducing unnecessary medical tests, treatments, and procedures [35, 36].

However, the application of these principles proves particularly challenging in high-risk specialty patient populations, such as children with congenital heart disease. For example, one preventable readmission in this study involved a patient with congenital heart disease who was initially found to have a positive blood culture, which was presumed to be a contaminant, leading to a delay in the administration of antibiotics. Subsequently, the infection was confirmed as endocarditis, necessitating prolonged antibiotic therapy and eventual mitral valve replacement. In retrospect, early empiric antibiotic treatment would likely have been beneficial, as the potential advantages of prompt intervention likely outweighed the risks of unnecessary treatment in this higher-risk patient.

Limitations

Our relatively small sample size from a single center limits the generalizability of our findings, and we recognize that the abundance of resources at our large institution may not be available elsewhere. There also remains the possibility of confounding factors in our investigation despite multivariable analysis, such as the fact that many children with congenital heart disease qualify for public insurance even when family income does not meet the required thresholds. Although we did not find an association between English-speaking status and preventable or urgent readmissions, only 3% of families were non-English speaking, which may not have been a large enough sample size to detect a significant difference. A hypothetical limitation of this study is that it was conducted across the years of the COVID-19 pandemic, and the social and medical factors contributing to readmission during this time may be different than non-pandemic years.

Conclusion

Measured SDOH were not associated with preventable or urgent readmissions to an acute care cardiology unit in children with cardiac disease. This is surprising yet encouraging, as many other studies in the literature have demonstrated an association of adverse SDOH with poor clinical outcomes, including increased readmission rates [23]. We postulate that our findings may be partially due to our institution’s multidisciplinary care teams and current strategies in place to ameliorate adverse SDOH. As the most common reason identified for preventable readmission was insufficient discharge education, however, clearly more work is needed to delineate what strategies are most effective for mitigating adverse SDOH, including more robust discharge education for families of our complex patient population.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CHD

Congenital heart disease

SDOH

Social determinants of health

Author Contributions

K.P. wrote the main manuscript text and prepared all figures. T.S. performed data collection and analysis. All authors reviewed the manuscript and provided pertinent edits.

Data Availability

The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request. Data are located in controlled access data storage at Cincinnati Children’s Hospital Medical Center.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical Approval

This study was reviewed by the Cincinnati Children’s Institutional Review Board and was considered exempt.