Introduction

Reparative and palliative congenital heart surgery during infancy is associated with high resource utilization at the time of surgery and well beyond hospital discharge. Resource utilization among patients with congenital heart disease (CHD) is high, even in simple or uncomplicated cases [1,2,3,4,5]. Specifically, in a population-based case–control study in the United Kingdom, patients with CHD consulted primary care more frequently, were issued more prescriptions, and were referred to a specialist more often than their matched controls [6].

Extracardiac co-morbid conditions (ECC) in children who undergo cardiac surgery can contribute to notable deterioration in health status measures [7] and likely concomitant greater resource utilization. A ten-year prospective study at Free University of Brussels found that one-fifth of children with CHD had significant ECC. Eleven percent of affected children had genetic or syndromic conditions. Most of ECC included neurologic, pulmonologic, orthopedic, nephro-urologic, gastroenterologic, and endocrinologic diseases [8].

We sought to first describe the frequency and types of chronic ECC in infants who undergo congenital heart surgery during the first year of life and its association with subsequent resource utilization, and then model the relation between the number of chronic ECC and health care utilization as a strategy for identifying patients who might most benefit from cardiac complex care programs similar to those shown to improve healthcare and decrease costs for children with medical complexity in general.

Methods

Children’s Hospital of Wisconsin (CHW) Institutional Review Board approved this retrospective chart review of infants less than 1 year of age who required surgery for a congenital heart defect from July 2006 through June 2011. Thirty patients who survived the index surgery hospitalization but died before the end of the two-year follow-up study period were excluded from the analysis. Additional exclusion criteria included isolated patent ductus arteriosus and non-structural cardiac conditions. The cohort included 876 unique patients.

Data were obtained from various sources. The medical record was used to obtain patient characteristics and demographics, as well as cardiac diagnoses, the index procedure, and associated STAT score. The STAT score is a tool designed by the Society of Thoracic Surgery and the European Association of Cardiothoracic Surgery to analyze the risk of mortality associated with congenital heart surgery procedures (STS-EACTS Mortality Risk Category) [9]. For each patient, all chronic ECC diagnosed prior to surgery and within first 2 years of index procedure were identified by manual chart review. The ECC were categorized by organ system as previously described [10]. In addition, they were divided into acquired or congenital conditions. Tertiary center resource utilization including hospital admissions, inpatient hospital days, emergency room visits, short stay hospitalizations, and total charges for two years after index procedure hospitalization was acquired from the administrative data. Demographic data including commercial versus governmental insurance and zip code data were also obtained from administrative data as a surrogate for socioeconomic status. Median income according to zip code was obtained from US Census data. Income data were not adjusted for inflation.

The characteristics of patients with ECC were compared to those without ECC using a Mann–Whitney U test. Linear regression analysis assessed the association between the number of complex, co-morbid conditions and unplanned readmissions, inpatient hospital days, and cost after adjusting for surgical complexity (STAT score). The associated predicted probabilities were plotted for each of the three outcome variables. All tests were conducted as two-tailed tests with p < 0.05 considered statistically significant. All analyses were conducted using Stata 12.1.

Results

Cohort Description and Extracardiac Co-morbid Conditions (ECC)

During the study period, 876 unique infants had surgery during the first year of life. Fifty-five percent (481/876) of the cohort had at least one chronic ECC. There was no gender difference (57 vs. 53% male, p = 0.3) and the age at index procedure was similar for those with and without ECC (62 vs. 54 days, p = 0.3). The ECC group had a significantly higher median STAT score of three compared to two in those without ECC (p < 0.001); however, there was variation of STAT scores across all numbers of ECC (Fig. 1).

Fig. 1
figure 1

Distribution of STAT scores across number of extracardiac conditions (ECC)

Full size image

Of the 481 patients with chronic ECC, 196 (41%) patients had one, 183 (38%) had two to three, and 102 (21%) had four or more ECC. The most commonly observed ECC included gastrointestinal, respiratory, and genetic diseases (Table 1). More specific diagnoses within the top three involved systems included need for enteral tube feeds, malrotation, vocal cord dysfunction, airway malacia, Trisomy 21, and DiGeorge Syndrome. Of the 481 patients with ECC, 197 (41%) had both acquired and congenital ECC, 101 (21%) had only acquired ECC, and 183 (38%) had only congenital ECC.

Table 1 Extracardiac conditions (ECC) in infants with 1 or more ECC
Full size table

Resource Utilization

Resource utilization was higher in patients with ECC and increased with the number of ECC (Table 2; Fig. 2). Specifically, median days hospitalized throughout the two years after the index procedure increased from zero for those with zero and one ECC to 27 for those with four or more ECC. Likewise, median charges significantly increased to greater than half a million dollars for those with four or more ECC.

Table 2 Median tertiary care resource utilization beyond discharge after index procedure (range in brackets)
Full size table
Fig. 2
figure 2

Median unplanned hospital days and median charges following discharge from index procedure for number of extracardiac co-morbid conditions (ECC). Unplanned hospital days (bars) and charges (solid line) increased with increasing number of ECC. * Denotes p < 0.05 for all comparisons

Full size image

When adjusted for STAT score, a direct correlation was observed between the total number of ECC and the predicted median number of unplanned hospital readmissions, predicted inpatient hospital days, and predicted cumulative charges minus index procedure (Fig. 3).

Fig. 3
figure 3

Predicted median admits charges and length of stay by total number of extracardiac co-morbid conditions adjusted for STAT complexity score

Full size image

Disparities by Socioeconomic Status

The presence of ECC was not significantly related to income bracket or insurance type. However, resource utilization was impacted by these factors. Median household income for the zip code of residence less than $30,000 was significantly associated with a higher number emergency room visits (p < 0.001), unplanned hospital days (p = 0.02), and total charges (p = 0.04) excluding index procedure cost when compared to all other income categories (Fig. 4). The median number of clinic visits was significantly greater for those with commercial insurance than government insurance (10 vs. 7, p = 0.03), while those with governmental insurance had more total hospital days (5 vs. 2, p = 0.02).

Fig. 4
figure 4

The relationship between median household income by zip code and median number of emergency visits, unplanned hospital days, and total costs beyond index procedure. Median household income less than $30 K was significantly associated with a higher number emergency room visits, unplanned hospital days, and total charges

Full size image

Discussion

The goal of this study was to describe the prevalence of extracardiac co-morbid conditions (ECC) in the congenital heart disease (CHD) population and relate this to resource utilization after discharge from the index cardiac procedure. More than half of infants in this cohort had at least one chronic condition in addition to their congenital heart defect. This is markedly increased from Greenwood’s report in 1975 that demonstrated a prevalence of ECCs in approximately 25% of infants with CHD; 40% had one ECC and another 40% had two conditions. In the earlier report, the more commonly involved systems were musculoskeletal, nervous, and renal-urinary systems [11], whereas we found the most common conditions to include morbidities with the gastrointestinal and respiratory systems, and include genetic syndromes. Reasons for differences in observed extracardiac morbidity are likely multifactorial, and may include improvements in operative and perioperative care that are protective to the brain and kidneys, improvement in nutritional support and respiratory support technology developed for home care, and improved diagnostic capability as observed with genetic testing.

Heightened attention to resource utilization associated with congenital heart surgery particularly with the more complex lesions over the past two decades have predominately focused on the surgical procedure and hospitalization. [1,2,3,4,5]. Our study reports median hospital charges at the time of the index procedure exceeding charges reported in 2005 by at least $20,000. [2] In Simeone’s report of inpatient costs associated with congenital heart disease, 70% of costs are attributed to infants less than 1 year of age and average approximately $50 million dollars annually within the state of Arkansas alone [1]. What is unclear in Simeone’s study as well as other reports on resource utilization is whether a significant portion of costs is associated with the presence of concomitant ECC. Ungerleider et al., however, reported high resource utilization was most common in patients less than 1 year of age, those who underwent more complex surgical procedures and in patients who were premature or had non-cardiac anomalies [3]. Connor et al. also supported the finding that more complex procedures as defined by RACHS-1 risk score was indeed associated with higher resource utilization at surgery [2].

Our study is the first to report unplanned hospital days and charges after discharge from the index procedure, and the relationship between utilization and ECC. Specific emphasis on unplanned hospital days stems from Cohen’s report of more than 15,000 hospitalized children of medical complexity. In his report, 27% of health spending was associated with readmissions and the readmission rate increased from 13% for those with one complex condition to 24% for those with multiple complex conditions. Similar to inpatient reports of costs associated with congenital heart surgery, higher resource utilization beyond discharge from the index procedure was associated with greater surgical complexity as defined by STAT category. Interestingly, the STAT score did not necessarily predict the presence or absence of ECC. When adjusting for STAT score, the linear regression models for unplanned readmissions, hospital days beyond discharge from the index procedure, and total charges demonstrated a direct and positive predictive correlation for a given number of ECC, thus emphasizing the importance of not solely relying on diagnosis or surgical complexity to identify patients who have complex extracardiac care needs and who are likely to consume a greater proportion of resources for ongoing care beyond cardiac surgery. The most striking finding in this report is that the 20% of the infants in this cohort with at least 4 extracardiac conditions had median charges of $529,000 over the two years following discharge from their index procedure. From our data, we are unable to determine the proportion of resource utilization attributed to care for the underlying cardiac defect alone versus care for the ECC.

It has been previously shown that cardiac patients with Medicaid or ‘other’ insurance have increased utilization [2], which was also the case for this cohort. Although ECC were not associated with insurance status or median household income, there was a difference in resource utilization, which is suggestive of the impact of socioeconomic status and health disparities in healthcare access for these patients. Lower income families were more likely to utilize emergency rooms and have unplanned hospital visits, while those with commercial insurance had a greater number of outpatient clinic visits to coordinate their care. Although healthcare reform has increased access to health insurance for a significant number of children, it often has limited access to specialty care providers [12]. This is further support for the role of increased care coordination in patients with CHD, which could help families overcome these disparities through improved support and access to subspecialty providers and resources.

Children with special health care needs (CSHCN) are those who have or are at increased risk for chronic physical, developmental, behavioral, or emotional conditions and require health services beyond what is generally expected for children [13]. Over 10 million children in the United States meet this definition [14]. Children with medical complexity (CMC) are a subset of CSHCN with multiple chronic conditions and account for the top 1% of pediatric healthcare spending [15]. Care coordination through complex care programs has been shown to decrease healthcare costs while improving quality of care for CMC and their families [10, 16]. Specifically, a complex care program enhances the inpatient care and transition to a medical home by coordinating medical and non-medical services for children with medical complexity and multiple chronic conditions [10]. Although not specific to cardiac disease alone, enrollment in the program has led to decrease hospitalizations, length of stay, and tertiary care charges as well as increase outpatient clinic use for children with medical complexity [14]. Children with CHD and ECC are a subgroup of CMC and therefore might benefit from similar care coordination. These data could serve as a model when considering structured enrollment of patients with multiple comorbidities into a cardiac complex care program. In infants admitted for surgical intervention of their congenital heart defect, attention to ECC may be delayed until full operative recovery thus delaying diagnosis, management, and care coordination, and ultimately contributing to longer hospital stays and greater costs.

Conclusion

The results of this study indicate that both STAT score and an increasing number of chronic extracardiac co-morbid conditions are associated with greater resource utilization following the index cardiac surgical procedure. Furthermore, these data may assist in determining which children might most benefit from a cardiac complex care program that partners families and providers to improve health and decrease healthcare costs. Future studies are needed to evaluate the impact of a cardiac complex care program on health, healthcare, and costs as well as family satisfaction and quality of life.