When we first held a Regenstrief Conference on large databases in 1989, health services research based on databases was relatively young. We were still identifying methodologic problems and trying to demonstrate the legitimacy of database research. There were plenty of skeptics who thought that database research should not be performed in most situations. Since then, huge databases have proliferated. The Medicare claims database has increased in great detail on some subsets of enrollees, large pharmaceutical benefits managers typically store billions of prescriptions in their databases each year, and managed care organizations collect several types of clinical and accounting data. Some of the databases have been linked to provide richer and more useful sets of information on individuals. The use of these databases for administrative and research purposes has become more creative. Policies are bing made by both the government and the private sector based on these analyses. Few would question the use of databases anymore. Yet many limitations and potential pitfalls remain. The promise of large databases can only be fulfilled if we use them judiciously.
The Sixth Regenstrief Conference highlighted several success stories of appropriate uses of databases. However, caution was a theme that ran through many of the presentations by experienced users of various databases. Also presented were some recently developed methods that could enhance the likelihood of drawing correct conclusions from database analysis. Experts spoke on uses of various databases, measurement of outcomes, process and cost of care, establishing linkages among databases, biostatistical issues in database research, report cards and profiling practices, and the future of databases and research.
After the overview by Iezzoni, this supplement to Annals of Internal Medicine organizes contributions into four areas: databases, research, quality measurement and improvement, and statistical methods. The section on databases includes an example of studies based on linked databases as well as major issues concerning linkage, such as standardization of codes and the question of privacy. The section on research presents the strengths and weaknesses of research through use of various databases, and one article presents a method on how to decide when a randomized trial is cost-effective for a managed care organization. The articles on quality measurement and improvement discuss difficulties in measuring health outcomes and promising improvements in new measures. The statistical methods section presents several methods that could improve the validity of findings from observational databases. Each fo these four sections is summarized by a concluding piece that highlights the articles contained in the section. The closing article by Davidoff discusses the future of databases in the next millennium.
Siu L. Hui, PhD
Regenstrief Institute for Health Care; Indiana University Medical Center; Indianapolis, IN 46202
Top
Author & Article Info
Article
