Is power everything? What can we learn from large data sets

Abstract

The research enterprise is both human and financially resource intensive, whereas the funding for research is highly competitive. How then can an investigator find ways to begin and sustain a viable program of research in a focus area of science absent of a significant infusion of funding? One option that many investigators do not initially consider is conducting secondary analysis using large health care data sets. Such large data sets are available from federal and state agencies and associations for nominal fees, if not for free. They include patient-level data or aggregated data for a particular type of patient or payor. These data are used by clinical and social scientists as well as by health services researchers to conduct analyses to inform quality agendas, health policy, and health care coverage determinations. Conducting secondary analysis using these large data sets can be the focus of an investigator’s entire program of research or can lead to intervention or translational research in the clinical setting. Large data sets facilitate replicating generalizable findings and understanding patient outcomes. By far, most large database analysis for researchers involves secondary analysis of existing data sets to explore clinical phenomena and answer new research questions. In part, this is achievable because existing databases offer the investigator rich sources of data to answer these research questions when funding opportunities may be limited or scarce and when primary data collection may not be economically or logistically feasible. Secondary analysis is a legitimate form of scientific inquiry in which nurses should be actively engaged (Castle, 2003). Although large data sets, ranging from administrative data to large-scale surveys, are collected for a specific