BACKGROUND: Upon adoption of the Patient Reporting Investigation Surveillance Manager (PRISM) for STD surveillance and case management in 2010, the Tennessee Department of Health (TDH) made significant changes to the application and databases to meet state-specific programmatic needs. This included a module for mapping ELR to a set of tables based on fifteen data elements common in a lab message. Between 2010 and 2016, TDH identified many areas of improvement to PRISM, including an in-depth analysis of the ELR mapping module. In March 2016, the STD program began a major effort to modify PRISM to take advantage of additional system features and ELR enhancements. One of the main changes was to import ELR from HL7 messages and lab results from an internal patient tracking and billing system (PTBMIS) in a single XML file format. All previous mappings were re-evaluated to address changes in file format and inaccuracies discovered during previous analyses.
METHODS: ELR testing of the PRISM replacement application took place during October 27-December 9, 2016. A test environment was established to validate HL7 and PTBMIS labs. Existing mappings were stored in a table and any newly imported ELR were compared against this table. If the system found a match, it was sent to a Lab table to be processed by STD program staff. If there was no match, the ELR was placed in a separate table to be mapped. This table of “missing mappings” was used to write Structured Query Language (SQL) rules that would automatically create new mappings for incoming ELR.
RESULTS: From October 1-December 2, 2016 68,502 HL7 and PTBMIS labs were imported into the PRISM test environment. After the SQL logic was written to automap these labs, less than 1% remained unmapped. The PRISM replacement application went live on December 12, 2016. Between December 12th and 23rd, 22,225 labs were imported, 712 new ELR mappings were automatically created from the SQL logic, and 22,138 labs (99.6%) were automatically mapped.
CONCLUSIONS: Prior to creation of SQL auto mapping rules, a person manually mapped every new unmapped ELR that was imported into PRISM. This required an excessive amount of time and introduced the possibility for human error. Creating rules to automate this process is a positive step towards standardization of lab results and greatly improves timeliness and data quality since ELR are mapped and sent to a user’s task list immediately rather than waiting in a mapping queue.