The implementation of a replacement Washington state Disease Reporting System (WDRS) presents an opportunity to reevaluate businesses processes and practices around health information and sharing. In doing so, biomedical informatics methods and tools can be used to streamline business flows in order to decrease workload, prevent duplicate data entry, and automate repetitive tasks.
METHODS:
WDRS implementation provided many opportunities to interact with subject matter experts (SMEs) concerning ongoing business processes for capturing and sharing health information as well as harmonization of existing collected information. Particular business processes were chosen for potential informatics solutions through specific criteria: 1) Does an already developed, extensible solution already exist, 2) Can the solution work within our IT infrastructure 3) What amount of workload can the solution save. Two processes were noted which currently required double data entry: investigation outbreak information collected on paper forms and then reentered onto DOH spreadsheets for analysis and the manual entry of lab information into our current system from faxed lab reports. As WDRS accepts roster imports, the end output of any solution would require table formatted data. A third process involved the harmonization of existing clinical and risk-factor data so that information can be shared within WDRS to streamline reporting and analytical functions (for example, case-control studies). This involved manual matching of similar terms that referred to the same clinical features/risk-factors across disease conditions.
RESULTS:
Among tested and implemented solutions were the use of adobe pdf forms, created easily from paper forms, which could contain information and output them into a table format. The forms could also contain logic (such as creating exposure date ranges from symptom onset date) written easily in Java. For data harmonization, exact matches were automatically linked in a table format using the R SQL join and similar matches (referring to the same concept but using different terminology) were found using the Unified Medical Language System (UMLS) Mesathesaurus. A local version of google’s optical character recognition software (Tesseract) read results obtained from paper lab reports in R and the data were then output into table rows for upload.
CONCLUSIONS:
The application of these informatics methods and tools show how useful, extensible solutions are readily available, only needing recognition of potential problems and solutions to be implemented. Having regular intervals of business process assessment may help shift the burden of public health work away from monotonous, work-intensive tasks and more towards supporting public health intervention and investigation work.