Archiving Pharmacy Information System & Oncology EMR Data as Indexed PDF Documents

The client is a hospital system that covers the locations of Mississauga, Ontario, and Western Toronto in Canada. With two hospitals and a health center, they offer the full range of acute care hospital services and a variety of community-based, specialized programs.

Key Challenges

The client rolled out Epic EHR, which resulted in the sunsetting of their Pharmacy Information System and Oncology EMR. They decided to archive the data of the last two years from these systems as indexed PDF documents, which would be available for viewing from Epic Hyperspace. Unfortunately, neither of the retiring systems had any ability to generate PDF documents. The key challenges that were involved in this scenario were as follows:

• Understanding the data model of both the Pharmacy Information System and the Oncology EMR
• Identifying how many document types are needed to present the information
• Generating a PDF layout that matches the application’s user interface
• Being less aware of the application
• Keeping up with the aggressive project delivery timelines

Solution Approach

314e Corporation undertook the project to extract patient data from the Pharmacy Information System and the Oncology EMR and generate PDF files to load them into a new EMR. Both the retiring software were high proprietary systems with little or no documentation about their database schemas. 314e’s team wrote custom python code that queried data from the Oracle database of one system and the MS SQL Server database of the other system to generate PDF reports with all relevant data.

This solution was developed using Python programming language utilizing Oracle native driver and SQL Server ODBC driver. We wrote SQL queries to retrieve relevant data from the database and generated high-quality PDF files using technologies like LaTeX. The program used a high degree of separation between PDF generation and data composition.

Business Outcomes

The end solution was optimized thoroughly for performance, such that the need for “catch-up” extracts was obviated. All data were extracted, converted to pdf, and loaded into the new EMR over the cut-over weekend.