Understanding the burden of pediatric gastrointestinal diseases – does a look from the perspective of the payers help?

Gastrointestinal (GI) diseases affect millions of US citizens yearly; recent studies estimate 4.6 million hospitalizations and an estimated $142 billion per year in both direct and indirect health costs.¹ Studies on the national burden of gastrointestinal diseases, however, have not been specific to the pediatric population. Because the most common GI diseases in hospitalized children (acute diarrhea, appendicitis, general abdominal pain, esophageal disorders) are different from those of adults (gallbladder disease, diverticular disease, pancreatitis) ^2–3 the cost burden GI diseases also is distinct in children. As healthcare spending in the United States continues to be source of national debate, it is imperative to better understand the burden of diseases specific to children to identify priorities in delivery of care, research and policy. A growing body of literature has contributed to defining this burden of GI diseases through the use of hospital administrative billing databases.

The Kids’ Inpatient Database (KID), one such database, is released every 3 years by the the Healthcare Cost and Utilization Project (HCUP), which itself is a federal-industry partnership lead by the Agency for Healthcare Research and Quality. ⁴ KID is one of the largest, publically available, all pediatric inpatient databases in the country, and includes a representative sample of > 3 million all payer discharges per year in 44 states. ⁵ Variables include patient demographics, primary and secondary diagnoses, procedures, measures of severity, and billed charges accrued through the hospitalization, as well as hospital characteristics (eg size, teaching status, census region). Diagnoses and procedures codes are assigned for billing purposed by trained medical coders at the hospitals using the International Classification of Disease, Ninth & Tenth Revision, Clinical Modification (ICD-9-CM & ICD-10-CM) codes.

The use of hospital administrative databases to define the inpatient burden of pediatric GI disease has been performed in a limited fashion to date. For example, Guthery et al. first utilized the KID data collected in 1997 to report that GI disorders, including appendicitis, intestinal infections, abdominal pain, and congenital GI anomalies, were a leading cause of hospitalization in children. ³. Like all secondary datasets, the KID has strengths and weaknesses. One advantage is its sheer size: rare diseases are present in sample sizes unavailable via the usual single-institution retrospective review. In addition, the KID contains data from a majority of medical centers across the country, yielding data that are thus nationally representative of the patients and care settings at large. The availability of large databases specific to pediatric patients can help address issues that arise from single center studies, where variations in length of stay, cost, and other outcomes centers may be present as the result of differences in geographic location, community vs academic institutions, and patient complexity. ⁶ There are, however, limitations with the use of these administrative databases. They rely heavily on ICD-9 and ICD-10 coding, which can lead to mischaracterization of patient populations, depending on whether a unique code is or is not available for a diagnosis or procedure. Furthermore, they lack detailed clinical data, which can be helpful to assess disease severity and other important factors. Finally, they are often cross-sectional and/or deidentified at the patient level, thereby precluding longitudinal follow-up of patients.

Acute pancreatitis is a GI disease in children with significant morbidity; however, its burden on the healthcare system is not well understood. The incidence of pediatric acute pancreatitis is estimated to be 3.6–13.2 per 100,000 in a small number of single institution studies. ⁷ Some studies have shown an increasing incidence of primary pediatric acute pancreatitis between 2000 to 2009. ⁸ Acute pancreatitis leads to hospitalization in many cases, and ~25% of children diagnosed with primary acute pancreatitis go on to develop a severe course, increasing the burden of disease. ^8–9 With an increasing number of children being diagnosed and hospitalized for acute pancreatitis, the growing burden on the healthcare system warrants further investigation.

In this edition of The Journal, using the KID database, Abu-El-Haija et al report on the growing burden of acute pancreatitis in children across the US in 2006, 2009, and 2012 and describe factors that impact healthcare use and disease outcomes in this population. ¹⁰ Across the 3 surveyed years, 28,288 total discharges with a primary diagnosis of acute pancreatitis were identified, and the number of discharges increased with age (<5 years old:5%; 5–14 years old: 29%, >14 years old:67%), with a predominance of females in all age groups. Gallstones, hypertriglyceridemia, obesity, and alcohol use were more common in older children ( 28.2% in adolescents vs 4% in the young aged cohort). The population <5 years of age had the greatest cost and length of stay per admission. A multivariate logistic regression model showed that chronic pancreatitis led to increased admissions for acute pancreatitis over time.

In North America, congenital anomalies are ranked first in the top 25 global causes of death in the Global Burden of Disease, Injuries, and Risk Factors study. ¹¹ Anorectal malformations (ARMS eg imperforate anus with and without fistula, various GI/genitourinary fistulae, and other conditions) are a type of congenital anomaly that has not been extensively studied. Single center studies indicate that associated urogenital, cardiac, renal, tracheoesophageal, limb, and spinal anomalies (the so called VACTERL association) are present in up to 60% of patients with ARMs. ^12–13 These patients require multiple extensive surgical interventions and suffer from long term problems, including fecal and urinary incontinence, chronic constipation, and sexual dysfunction. ^14–16

Also in this volume of The Journal, Kovacic et al. report the prevalence of congenital anomalies associated with ARM and hypothesized that health care use for ARMs increased in patients with multiple associated anomalies. ¹⁷ The KID was analyzed for the years 2006, 2009, and 2012 to identify a large national cohort of patients with ARMs. Of the 2,396 children with ARMs in the study, 80% of children had other congenital anomalies, including urogenital anomalies (30%), cardiac anomalies (21%), renal anomalies (15%), genetic disorders (14%), and spinal anomalies (6%). Further analysis showed a significant association in the number of associated congenital anomalies, length of stay and cost. Patients with isolated ARM had an average length of stay of 4 days and a mean hospital charge of $35,029. In contrast, patients with 4 or more VACTERL associations had a 10-fold increase in their lengths of stay and hospital charges.

These 2 studies have demonstrated how the use of a large, nationally representative, hospital billing database can contribute to our understanding of the burden of GI diseases. Although several limitations exist in using these databases, studies such as these document the financial and health burden of pediatric GI disorders. More such studies are needed to guide future resource allocation, design more detailed areas of investigation, and inform policy and health care decisions.