One of the most difficult challenges clinical laboratories face is implementing new laboratory automation systems. This occurs infrequently—just once every 5 to 15 years—but once started requires extensive advance planning and ample flexibility and patience. The University of Chicago Medical Center (UCMC) clinical chemistry laboratory recently went through this process and learned several lessons that may be of value to other labs.
Our previous automation system—implemented over 17 years—included pre-analytical instruments that centrifuge, aliquot, and sort specimens, as well as ion-selective electrode, general chemistry, and immunoassay instruments. The new system replaced the analytical instruments and added a new post-analytical refrigerated automated specimen storage and retrieval system. The new instruments—by the same manufacturer as our old system—provided higher throughput, consolidated multiple instruments, and enabled us to add the post-analytical system without increasing the footprint of the overall system, which handles about 5 million tests annually.
One of the most difficult challenges we faced was that the new system needed to be installed in the existing space as the old system without any disruption to testing. We accomplished this in multiple phases that required decontaminating, uninstalling, and removing different sections of the old system and then installing and validating the new instruments.
During this phase, we had to perform all our testing on the old system’s remaining instruments. This necessitated major workflow changes, including more hands-on specimen processing and transporting because instruments might or might not have been connected by the track system. Our lab staff really had to rise to this challenge because even with a less efficient process, our turnaround time goals remained unchanged.
We also ran into some unplanned construction work. Sometimes structural, electrical, or water damage doesn’t become apparent until the older instruments have been removed. Depending on the severity of the damage, this may cause minor or significant delays for the entire project. We also had to deal with added safety hazards when removing old instruments and installing new ones as this exposed electrical and water lines in an area that had heavy foot traffic.
The IT Compatibility Challenge
Another major challenge we faced involved computer network and information technology (IT) hardware and software security issues. While every lab likely will have some hiccups in this realm during an automation installation, each one may have a different experience. This is because each institution has different IT infrastructure and support, and software and hardware security protocols. Manufacturers also offer different software and/or hardware security solutions as part of their systems.
Problems arise particularly when the manufacturer’s security solution is not compatible with the institution’s network, software, and hardware security protocols. That’s why it is essential for the right IT, software, and hardware groups from both manufacturer and institution to meet early on and discuss these potential issues in detail.
We also faced a big challenge in completing our validation experiments and related data analysis for the many assays that run on the new system. In just 4 weeks, we installed, validated, analyzed, and reviewed 150 assays. As demanding as it was, we had to pass this critical juncture in order for the installation project to proceed.
Validation studies on some assays may be especially challenging because there may be a limited number of samples or routine samples may not cover a broad range of analyte concentrations. In anticipation of this, almost 6 months before installation began we identified and started biobanking unique specimens (very high or low analyte concentration and/or low testing volume).
We also made other important assay validation decisions while we were in the initial planning phase. We considered two key questions: 1) Should an assay be validated on the instrument assigned to run that test or should it be validated on all instruments capable of running that test? 2) Should an assay be validated on the primary and backup instruments assigned to run the test? In making these decisions, labs need to balance the available resources, costs and time required to complete and analyze data. At UCMC, we validated each assay on both the primary and backup instruments assigned to run each test.
Maintaining effective communication throughout the entire project was another challenge. The overall process spanned more than a year, and during this time there were several key personnel changes, including the manufacturer’s project manager (twice), and UCMC’s lab manager (once). Even though we documented our discussions from every meeting, some key details inevitably slipped through the cracks.
New Instruments, New Questions
We were the first organization in the Midwest to install the manufacturer’s post-analytical refrigerated automated specimen storage and retrieval system, and our early adopter status for this new technology put everyone involved on a learning curve. It was not uncommon to hear the field service or application specialists mention, “I have never encountered that before” or “That’s the first time I was asked that question.”
Our challenges didn’t stop with our new system’s installation. We had to make several crucial adjustments to optimize the system. Most notably, when we were designing the system, our manufacturer used data such as the test menu and testing volume to determine which instruments and how many of each we needed, and which assays would be assigned to each instrument. However, we discovered that the algorithm they used to determine these outputs didn’t account for certain variables important to our lab’s daily operations, which increased instrument downtime. The solution was to reassign some assays to different instruments and revalidate these assays on the new instruments.
Implementing and validating a new laboratory automation system is a very long and challenging but ultimately rewarding project. Client labs and manufacturers will learn important lessons all along the way, and labs definitely should give manufacturers constructive feedback so they can refine their procedures for other clients.
Detailed planning, clear communications, and a flexible response to unavoidable glitches are key to keeping the implementation on track.
Edward Ki Yun Leung, PhD, DABCC, FACB, is assistant professor of pathology and assistant director of clinical chemistry, director of phlebotomy and pre-analytical services, and director of point-of-care testing at University of Chicago Medicine. +Email: [email protected]