Background

This Government Department has the largest integrated healthcare network in the United States, contracting 1,255 medical institutions to serve 9 million military veterans every year. It has dedicated Veteran Integrated Service Networks (VISNs) that work with selected hospitals and outpatient clinics to meet veterans’ medical needs in the covered states. VISNs employ medication wholesalers and Pharmacy Benefit Management (PBM) on a contractual basis to supply pharmaceutical products to the designated healthcare institutions and bill the government healthcare department. 

Use Case

This Department works with a medication wholesaler that provides pharmaceutical products to hospitals and outpatient clinics on its behalf to fulfill veterans’ medical needs. The wholesaler sends invoice data in a PDF format to the healthcare facilities, and a dedicated pharmacoeconomist collects them for internal reporting. The invoices generally range between 1,000 to 5,000 lines and contain drug and patient information details for different healthcare treatments based on daily utilization.  

Once procured, the data from these invoices needs to be uploaded onto the Microsoft SQL Server database. The SQL Server database supplies data for reports that tracks medication usage, assess cost trends, and predicts inventory shortages. The reports also enable the pharmacoeconomists to plan the VISN’s monthly budget and identify opportunities for potential cost savings based on the supply and demand trends of various pharmaceutical products.  

The VISN also uses the reports to identify fraud and make sure that only eligible patients are utilizing the service. 

For the VISN, gathering insights from invoice PDFs was the main challenge because the SQL Server Integration Service (SSIS) doesn’t support importing data directly from this file format into the SQL Server database. This meant the pharmacoeconomists had to manually extract and aggregate data from more than 1,000 transactions every day . It prevented the team from having an in-depth view of the medication wholesaler’s purchasing patterns and performing accurate cost analysis to utilize the monthly budget better. Similarly, PBMs sent monthly invoices consisting of prescription and billing information in a spreadsheet., making it challenging to aggregate usage data for medication.  

The VISN sought a solution that can parse the incoming invoices, whether in PDFs or spreadsheets, and convert them into a format that SSIS can read and query. In addition, the solution needed to be able to build a repeatable process so that the business users didn’t have to extract data manually from the invoices.  

“For the most part, we couldn’t really use the data before because it was in a format we couldn’t use. For the wholesaler invoices, we just look at it manually. We looked them over individually and used the best guess on what to buy or identify any trend.” 

Enter Astera ReportMiner

After reviewing multiple products in the market and trying Altair Monarch, the team chose Astera ReportMiner to build their extraction templates.  

ReportMiner is a complete data extraction solution that is built to handle a range of unstructured data formats including PDFs, PDF forms, TXT, PRN, RTF, DOC, DOCX, XLS, and XLSX. It combines rule-based data extraction with an enterprise-grade ETL engine to help businesses free data from unstructured files. With Report Miner, VISN’s process became  as easy as copy-pasting the invoice PDFs into the server folder and receiving the extracted data in an Excel file after half a minute.  

ReportMiner helped VISN easily and automatically parse through incoming invoices and without any manual extraction. With high-quality invoice data available, the VISN’s pharmacoeconomists now get deeper insights into the medication requirements and utilization trends at a granular level. Furthermore, they can better react to changing market dimensions, like medicine prices, availability of higher quality brands, and better deals on bulk purchases, providing more opportunities for higher cost savings. 

“Before ReportMiner, it used to take anywhere between 30 minutes to a couple of hours to process a single invoice. Now, it takes less than a minute.”