Methods:
The Kids’ Inpatient Database (KID) for the years 2000, 2003, 2006, 2009,
2012 and 2016 was used to identify all ICD related hospitalizations. The
KID is part of the Healthcare Cost and Utilization Project (HCUP),
managed by the Agency for Healthcare Research and Quality (AHRQ) and is
the only all-payer inpatient care database for pediatric admissions
(defined as age at discharge ≤20 years) in the United States that. It
represents 2-3 million discharges per year from public hospitals,
specialty hospitals and academic medical centers. Weighted totals were
used for analysis. Specific details regarding data in the KID databases
are provided in the supplemental section.
For the purposes of this study, ICD admissions were included, defined as
patients with existing ICDs at the time of admission. Admissions related
to new ICD implantation based on ICD-9 and ICD-10 procedure codes
(listed in supplemental information) during the admission were excluded.
Procedural codes for replacement of leads or generators without new
implant ICD codes were included in ICD admissions. Diagnostic codes were
used to categorize admissions into one of four mutually exclusive
primary diagnostic categories of underlying disease which were defined
as primary 1) CHD (e.g. Tetralogy of Fallot) 2) cardiomyopathy (CM, e.g.
hypertrophic CM), 3) arrhythmia disorder (e.g. Long QT syndrome) and 4)
Other (e.g. muscular dystrophy or patients in whom diagnostic category
could not be determined based on ICD codes available). Coding was
hierarchical, with CHD before cardiomyopathy and cardiomyopathy before
arrhythmia. As an example, a discharge with Tetralogy of Fallot and
cardiomyopathy diagnostic codes was considered primary congenital heart
disease. Minor defects such as atrial septal defects (ASD), patent
ductus arteriosus (PDA) and patent foramen ovale (PFO) were not included
in the CHD category. Secondary cardiomyopathy due to nutritional
deficiencies (e.g beri beri) and alcoholic cardiomyopathy were not
included in the cardiomyopathy category.
The first aim of the study was to evaluate trends over time in pediatric
admissions with ICDs. Rates of hospitalization were calculated by using
included hospitalizations < 21 years of age as the
numerator and total hospitalizations < 21 years as the
denominator, and were calculated by year. The second aim was to
determine rates of in-hospital death among admissions with ICDs. All
deaths during ICD hospitalizations were identified and compared to total
ICD hospitalizations by year. Discharges resulting in hospice care were
excluded from this analysis. Our third aim was to evaluate factors
associated with in-hospital death in pediatric admissions with ICDs.
Data collection included patient demographics and hospital
characteristics [hospital region (Northeast, Midwest, South, West) and
hospital type (rural, urban non-teaching, and urban teaching)].
Patient characteristics included age, sex, race/ethnicity (non-Hispanic
White, non-Hispanic Black, Hispanic, or other which included
Asian/Pacific Islander, American Indian, and other), and primary payor
(government included Medicare and Medicaid, private insurance company
including health maintenance organization, and self-pay, no charge, or
other), admission season (fall-September, October, November; winter –
December, January, February; spring – March, April, May; Summer –
June, July, August). Specific details regarding categorization by
diagnosis are provided in the supplemental data. Data regarding other
admission factors including whether the admission was elective, whether
there was a code for heart failure, cardiac surgery or procedure, ICD
complication (consisting of diagnostic codes for mechanical
complications, ICD infection, pneumothorax, hemothorax or procedural
codes of ICD-revisions) or cardiac arrest were collected. Specific ICD
codes used are provided in the supplemental section.