(P-IT-11) Software Validation: Not A New Process

Our system, comprised of 8 hospitals, was looking to improve operational efficiencies across our facilities. As part of that process, the decision was made to implement Immulink (IM) (Werfen, Norcross, Ga.) in our Transfusion Services. Immulink is a software suite designed specifically for Transfusion Medicine to improve testing efficiency and traceability. Functions occur across three separate ImmuLINK modules: ImmuLINK Manage (IMM), ImmuLINK Archive (IA), and ImmuLINK Panel ID (IPID).

While implementation of equipment is common to our system with readily available examples of validation plans, our challenge with implementation was unfamiliarity and experience with such software validation. The purpose of this abstract is to document our approach to validation of this unique software solution.

Study

Design/Methods:

Collaboration occurred with manufacturer’s representatives to understand capabilities, functionality and requirements of the software.  Once this was understood, risk assessment was analyzed to better guide drafting a validation plan that would be comprehensive enough to address and mitigate risks.  We defined the parts of validation: Installation Qualification (IQ), Operational Qualification (OQ) and Performance Qualification (PQ) following our structure for other validation processes. Utilizing Validation Guides provided by the manufacturer, we determined what PQ test cases would apply to our facilities based on our user-defined requirements. 

Results/Findings:

A validation plan was drafted for implementation of IM. Our facility decided to do a phased implementation of IM: to validate IMM and IA first, followed by validation of IPID. This approach allowed us to compartmentalize the risk assessment – focusing first on the interfacing and data management with later focus on IPID functionality.

Attention was focused on infrastructure, security, and interfacing first. PQ test cases included: order uploads, result downloads, feature functionality, security, and data archive.

We defined our timeline as 60-90 days with 5 hospitals going live first, followed by an additional 3 hospitals. The first five hospitals were already live with blood bank automation and LIS software, whereas the 3 additional hospitals needed to implement or upgrade automation and convert to a new LIS.

Conclusions:

Risk assessment is crucial to drafting an adequately comprehensive validation plan. Vendor resources such as validation guides allowed us to develop a logical sequence of testing scenarios. The scope of testing was defined to allow sufficient challenge to the system and provide assurance to the shareholders that the system would function appropriately. Validation of software is not unlike validation of other equipment in the Transfusion Service and is not a new nor difficult process. Although initially intimidating, use of existing validation principles facilitated the implementation.