Database Research in the PICU: Time to Go Big!*

    loading  Checking for direct PDF access through Ovid

Excerpt

The number of physicians who conduct research (aka physician-scientists) has been steadily declining for at least the last decade (1–3). In part, this is due to today’s hypercompetitive research funding environment (2, 3). Providers are under more and more pressure to produce clinical revenue and obtaining independent research funding remains very challenging. The success rate for National Institutes of Health (NIH) funding has also been steadily declining (1, 2), while at the same time, the average age of NIH grant recipients has been rising. In 2012, the average age of NIH grant awardees was 51 years old, and the average age of first time awardees was 44 (1). Given this environment, why would a young trainee or faculty member choose to focus on research when there are more fruitful areas to devote their nonclinical efforts?
Despite these challenges, there are still those who remain committed to conducting research (1, 2). In 2012, 17% of graduating medical students expressed a significant interest in pursuing a research career (3). These young faculty and trainees are hungry for opportunity and they require projects that can be completed within a relatively short time frame that fits their educational schedules.
One such opportunity is the analysis of preexisting databases to explore research questions. In Pediatric Critical Care, the use of databases can be an especially powerful tool (4). There is significant heterogeneity of diagnoses in a PICU and the relatively small numbers of patients seen at any one particular site make single-center studies challenging and poorly generalizable. While some have criticized database research as “parasitic,” (5) analysis of preexisting databases can be an essential tool for exploring research questions and providing preliminary data for prospective trials.
There are several types of these databases available in Pediatric Critical Care (4). The Collaborative Pediatric Critical Care Research Network (CPCCRN) and Pediatric Emergency Care Applied Research Network (PECARN) are federally funded networks that each have public use datasets available online. There are therapy specific databases such as the Society of Thoracic Surgeons Congenital Heart Surgery Database and the Extracorporeal Membrane Oxygenation (ECMO) Registry of the Extracorporeal Life Support Organization that can used by member centers and are excellent sources of outcomes data for these therapies. The Pediatric Health Information Systems (PHIS) database is a quality improvement database that is a rich source of resource utilization data. And there are pediatric registry databases available such as the Virtual Pediatric Systems (VPS) database (Virtual Pediatric Systems, LLC, Los Angeles, CA) in the United States, the Australia New Zealand Paediatric Intensive Care Registry, and the Paediatric Intensive Care Audit Network in the United Kingdom and Ireland.
Each of these databases has its own limitations in terms of accessibility, cost, and level of clinical detail (4). For example, data from federally funded databases (such as CPCCRN and PECARN) are freely available online, but to use data from the ESLO or VPS registry, a researcher needs to either be employed by a center that subscribes to that database (and has paid staff entering the data) or partner with a provider at a site that is a member. Additionally, each database contains different variables that shape and limit the research questions that can be examined. Some have rich physiologic data (ECMO and VPS registries), while others have detailed information about therapies and procedures received (PHIS database). Potential investigators should carefully review their research questions for feasibility when choosing a database to explore.
In this issue of Pediatric Critical Care Medicine, Dr. Halvorson et al (6) used one of these databases, the VPS database to examine the relationship between obesity and vascular access in the PICU.

Related Topics

    loading  Loading Related Articles