A Systematic Mechanism for the Collection and Interpretation of Display Format Pathology Test Results from Australian Primary Care Records
Objectives: To demonstrate a mechanism for access to primary care patient data and the use and systematic interpretation of display format textual pathology test results held in primary care – in this case for Chlamydia surveillance.
Background: In Australia, it is common for pathology test results to be stored electronically in primary care medical record systems. At this time, there are few options for wide-scale access to such data that complies with privacy legislation, that supports record-linkage to other clinical databases and that can properly ensure textual results are correctly interpreted.
Methods: The GRHANITE™ software system (offering generic computer system interfacing, consent management, and privacy-preserving record linkage capabilities) was utilized to ethically extract patient consultation, pathology test requests and associated pathology test results for Chlamydia from Primary Care. In a sample of 10 practices (two different GP computer systems), results from 16 separate laboratories were obtained – all of which report test results differently. A rule-set for the parsing of this data was created and a C# program created to auto-generate rule-testing and rule interpretation SQL code.
Results: Utilising the rule-set interpreter, we were able to systematically verify that: All laboratories supplying data to the practices were included in the analysis, all tests for Chlamydia were identified (often several tests from each laboratory), all specimen types were accounted for and all results were correctly interpreted. In total 7,072 test results were analysed producing 236 distinct rules for the interpretation of the data from 16 laboratories. The rule-set generator created 3,102 lines of SQL code for rule verification and 9,135 lines for data consolidation. Later analysis against laboratory records confirmed the parsing strategy to be accurate in all cases.
Implications: GRHANITE™ has shown itself capable of meeting ethical requirements for data access and extracting data at a patient level (that is record-linkable) from a number of practices across multiple distinct general practice systems. The diversity of recording formats utilised by laboratories makes the electronic interpretation of textual pathology data difficult. The rule-set interpreter mechanism utilized here provides a reliable, extensible solution. We now have a sustainable and fast mechanism for future interpretation of any textual pathology data as held in Australian GP computer systems.
This work is licensed under a Creative Commons Attribution 3.0 License.
--- eJHI - electronic Journal of Health Informatics - ISSN 1446-4381 ---