Overview of Large Database Analysis in Renal Transplantation
The discipline of renal transplantation has been fortunate in having one of the largest and most complete databases of any field of medical inquiry. Many analyses have been performed utilizing these databases and a wide array of views regarding the role and limitations of these analyses exists. In this manuscript, we hope to present the merits and limitations of large database analysis in renal transplantation. In addition, we will attempt to go over the major databases' structures and the critical issues of coding, verification and clinical awareness that help maintain the integrity of this type of analysis. Other fundamental issues that will be covered include the relationship of single center and randomized prospective studies with database analyses, statistical considerations, and hidden selection bias inherent in registry analysis. We hope to present both a guide to investigators who are first embarking on large-scale database analysis and also to present a fair view of the utility and limitations of a potentially very useful research tool in renal transplantation.
The Organ Procurement and Transplantation Network (OPTN) is the unified transplant network established by the United States Congress under the National Organ Transplant Act (NOTA) of 1984 to be operated by a private, nonprofit organization under federal contract. The United Network of Organ Sharing (UNOS) is the contracted body under the OPTN that has developed a system for data collection and storage of transplantation information among other tasks. The OPTN database contains a wide array of information about transplant recipients and donor characteristics for organ transplantation. These data constitute a primary source for descriptive summaries and analytical studies concerning organ transplantation. In addition to this primary source of data, the United States Renal Data System (USRDS) and the Scientific Registry of Transplant Recipients (SRTR) work with OPTN/UNOS to validate structure and combine data from other sources obtaining data from registries. The data used by the USRDS (Coordinating Center) originates from CMS, UNOS, the CDC, and the ESRD Networks, as well as from the USRDS special studies. The USRDS provides a comprehensive data source available for research using the accumulation of this information. However, with data deriving from multiple sources, users should be cognizant of the primary sources of specific variable information and the relative reliability of each. Of pertinence to researchers, is that data have been made available with different criteria in different eras, and as such one must be cognizant of the changes. For example, centers and dialysis units were only required to provide data for nonMedicare patients after 1994. Prior to 1988, CMS was the main source of transplant data, after which UNOS was established and provided transplant data in conjunction with CMS. After 1994, transplant files from the UNOS database was the source of transplant events for USRDS files.
The Scientific Registry of Transplant Recipients (SRTR) is a nonprofit making organization administered by the University Renal Research and Education Association (URREA). The SRTR also possesses a vast database of organ transplant information deriving from public reports, and special data requests. One advantage of the SRTR database is that data have been made available for more recent transplant events (relative to the USRDS); however, current standard analysis files do not possess the volume of variable information that is derived from the extensive sources as in the USRDS.
The Collaborative Transplant Study (CTS) is another source of multicenter transplant data. The CTS possesses international transplant information supplied from centers which voluntarily participate. General outcome graphics can be obtained through their web site for multiple organ transplant types and more specific information is available for participants of their study.
In general, to obtain data from these sources, an analysis plan must be completed, a security and confidentiality plan addressed, and depending on the information requested, a variable cost may be involved. Some general transplant data are readily available via the web sites of the organization or through written reports, but more substantial files must be purchased. Research intended for publication purposes must be reviewed by the applicable organization, and investigators must be cognizant of the policies and procedures for the handling and presentation of any research deriving from these sources. In all cases, investigators must administer care regarding the security of such data and limit access to those individuals on agreement forms. It is important to note that the data from each of these sources tend to have unique characteristics; time and effort must be afforded to critically reviewing the unique weaknesses and strengths of each source. Appendix 1 outlines the procedures necessary to obtain files from these agencies.