We would like to thank Dr. Gutgesell for reading our manuscript and providing his insightful comments. The primary objectives of our study were to characterize outcomes and resource utilization of all infants born with hypoplastic left heart syndrome (HLHS) in the Intermountain West. There were significant differences between our study and administrative database studies. Although our study was from a single center, it was truly population based and captured all infants born with HLHS in the state of Utah for the following unique reasons.

First, unlike other tertiary care pediatric cardiology centers, Primary Children’s Medical Center is the only children’s hospital with a congenital heart surgery program in the entire state of Utah. Primary Children’s Medical Center provides medical care not only to all children born with HLHS in the state of Utah but also to children born with HLHS in many surrounding intermountain states, including Idaho, Wyoming, Montana, Nevada, and Colorado.

Second, to ensure that we captured every child born with HLHS in the state of Utah, the study cohort was identified using multiple databases, including the Utah Birth Defect Network (UBDN), the Enterprise Data Warehouse (EDW) of Intermountain Healthcare, the University of Utah Fetal Heart Center database, and the University of Utah Pediatric Cardiothoracic Surgery database.

Third, the EDW is the electronic administrative data set of Intermountain Healthcare and contains data for every inpatient and outpatient encounter, including hospitalizations, procedures, and charges from the 23 hospitals in the Intermountain Healthcare System, including Primary Children’s Medical Center. Intermountain Healthcare is world renowned for its quality of health care, huge database of patient records, and efficient management. Utah children with HLHS receive all of their inpatient care and outpatient cardiology visits at 1 of the Intermountain Healthcare facilities, so their data are available in the EDW. The completeness and granularity of data obtained from the EDW database are unmatched by any other database.

Fourth, the UBDN is an active surveillance statewide birth defects registry that collects information on children from 20 weeks’ gestation to 2 years of age. All Utah hospitals and birthing centers are required to report a specific set of information to the UBDN when an infant is born with a birth defect. Each reported case is then actively confirmed and collected by a UBDN staff member. This method of data collection with checks and balances ensures capture of not only all live-born infants with HLHS but also those who are stillborn, electively terminated, or spontaneously aborted.

Last, the unique religious and sociocultural makeup of the Utah population leads to a population that tends to stay in the area. Few patients are lost to follow-up, unlike in other database populations. Of the 245 subjects included in this study, complete follow-up was available for 240 (98%).

The strengths and weaknesses of administrative databases are well known. Administrative health care retrospective databases continue to be an important data source for outcome research, however, because they provide large numbers from many institutions, as Dr. Gutgesell notes. Despite this, there are multiple well-known shortcomings and methodologic challenges unique to these data sources that introduce bias, including the following: (1) Identification of the study cohort is often problematic, leading to issues of reliability and validity. International Classification of Diseases or Current Procedural Terminology codes for identifying patients and procedures change with changes in reporting and coding practices over time. Quality assurance, quality checks, and data cleaning of databases often lack standardization of charges across different centers. Even the best administrative databases rarely capture the entire study population, particularly over time, (2) Linkages among data sources and/or different care sites may be problematic because of differences in coding and reporting across sources, (3) The information in most administrative database is not sufficiently granular to study clinically meaningful differences in the important outcome. Data fields that are not required for reimbursement may be particularly unreliable, and (4) Follow-up is often short and incomplete.

These problem areas introduce bias and render administrative database studies subject to error.

The inclusion of resource utilization associated with cardiac transplantation was beyond the scope of our study.

In conclusion, although from a single tertiary care pediatric cardiology center, because of the unique nature of our study population and data collection processes, our study provides comprehensive population-based data regarding resource utilization and outcomes of all infants with HLHS in Utah and overcomes many limitations of registry- and administrative-based studies.