Abstract
Background: The availability of comparable laboratory (lab) results across sites will expand the abilities of the HMO Research Network (HMORN) to perform collaborative research. The Coordinated Clinical Studies Network (CCSN) funded research to define the process of adding a lab result content area to the virtual data warehouse (VDW). We describe the development efforts and implementation steps required to create and maintain a laboratory result content area that is equivalent in definition across HMORN sites.
Methods: Existing information about lab data and coding systems at HMORN sites was reviewed, and we surveyed sites about current lab data. Sites were specifically asked about the availability of Logical Observation Identifiers Names and Codes (LOINC), which are a set of universal names and ID codes for identifying lab and clinical test results. To enable sites to participate in lab research studies, written instructions for incorporating lab tests, including a data dictionary, were prepared. The accuracy of the instructions was assessed for selected tests: serum creatinine and potassium, international normalization ratio, glycated hemoglobin, and fasting blood glucose. When a lab test is to be added to the laboratory content area, the lead site determines a list of LOINC codes that are interchangeable at a component level for that lab test. Sites that have incorporated LOINC into their lab database use this list to pull data. Each site without LOINC and each site where data are needed for time periods prior to LOINC availability will determine a list of local codes. Data pulled by LOINC and data pulled by local codes are tied together by a common test type.
Results: Most HMORN sites indicated they had the LOINC code set available. Kaiser Permanente Colorado is working with several other HMORN sites on hypertension studies. These sites are beginning to implement this process, utilizing LOINC and local codes from 2002 forward. This will be the first large-scale test of the lab component of the VDW.
Conclusions: Implementing the lab content area requires upfront effort from the lead site for each lab test. Each participating site must also invest time determining its local codes. The aggregate information from all sites participating in a study will enhance the usefulness of the VDW to epidemiological, observational, and interventional research in the HMORN and in other collaborative efforts.
- Received September 11, 2008.
