By Renee DiIulio
This COULTER LH 1500 series automation system is designed to manage the entire hematology testing process, from preanalytical sorting to postanalytical storage.
Tight budgets, staffing shortages, higher volumes, and inadequate reimbursements have become epidemic, followed by directives to Do more with less. Automation is curing some of these ills, as new systems are designed to reduce turnaround times, free staff, increase safety, decrease errors, and improve patient care. And though the initial investment may be high, it can be recouped within 2 years as a result of higher revenues and lower expenses, making it accessible to even midsize labs.
Automation hardware and software eliminate manual steps throughout the entire laboratory workflow: preanalytic processes, such as specimen preparation and sorting; testing procedures; and postanalytics, including verification and results delivery. But laboratories are advised to analyze their own processes before automating anything. Understanding where the bottlenecks really lie is crucial to developing a system and process that can increase the value of the lab and its results.
Sam Terese, BA, BS, MA
Sam Terese, BA, BS, MA, director of laboratory services at Elmhurst Memorial Healthcare (Elmhurst, Ill), has the numbers to back this up. His lab installed a track-based automation system and data manager from Beckman Coulter Inc (Fullerton, Calif) to help process its roughly one million annual billables. This translates to two million tests, says Terese.
The Elmhurst total automation project paid back its roughly $2 million cost, which included approximately $500,000 for lab remodeling, in just over 1 year with labor and nonlabor (primarily reagent) savings. We were surprised at how quickly the system paid for itself. We projected a 2-year payback, but the investment was paid back in just over 1 year, says Terese. Without automation, our budget was projected to be $4 million more than where it is now, he adds.
A portion of these savings is attributed to a smaller staff. We dropped 20 [full-time employees (FTEs)], mostly through attrition, says Terese, who suggests that labs adding automation can begin to leave positions open before the new system goes live.
These dollar figures can help win over a facilitys CFO. Other metrics will appeal to CEOs, who are typically concerned with customer satisfaction, both physicians and patients; quality-assurance staff; and board members. The equipment itself is impressive. We brought our board members down to see it, and it sent a powerful message regarding future automation, says Terese.
Prior to its total automation, the Elmhurst lab had a 15-step process and a 53-minute cycle time; postautomation, the process was reduced to 4 steps and a 10-minute cycle time. Physician satisfaction rose from 72% to 82%. Most of our physicians order tests during their morning rounds and want the results by 8 am. Preautomation, we hit this mark 55 percent of the time. Postautomation, we delivered 96 percent of the time, says Terese. Other improvements included:
The labs target of a 45-minute turnaround time for tests related to patients with chest pains in the emergency department (ED) was met 72% of the time pre-tomation, and now is met more than 90%.
Delivery of routine ED test results in 35 minutes rose from 58% pre-automation to 94% post.
Instead of verifying 100% of the results produced by the lab, technologists now manually verify about 20%; the data-management system autoverifies 80%.
With these capacity and productivity gains, the lab was also able to increase its reference volume by 15% without adding staff, generating new revenues for the facility.
These results are not unique. In a presentation given by Kenneth E. Blick, PhD, ABCC, NACB, director of clinical chemistry for the Oklahoma University Medical Center (OUMC of Oklahoma City, Okla), and Lorne Holland, MD, MS, resident in clinical pathology at OUMC, at the American College of Healthcare Executives (ACHE) 2005 Conference on Health Care Management, they shared that the OUMC brought in an additional $80,000 of referred testing in just 1 month with no added labor expense.
Evaluation Before Automation
Ron Berman, worldwide director of product management, automation, and information systems with Beckman Coulter, notes that automation typically brings improvements in three areas: economics, patient safety, and worker safety.
The gains in productivity allow more work to be done with fewer people and can result in cost savings or new revenues. The faster and more consistent turnaround times eliminate delays in patient diagnosis and treatment, while the reduction in error from the elimination of hands-on tasks produces greater confidence in the results. The use of data-management software and laboratory information systems (LIS) also reduces errors introduced with manual verification and transcription; and it ensures that the results are delivered to the right person in the proper format. Employees benefit from reduced exposure to biohazardous materials as well as decreased risk of repetitive-motion disorders, such as carpal tunnel syndrome.
Labs may want to experience all of these benefits, but should automate according to their specific goals. These differ with every company, says Dave Liptrot, marketing manager, Labtronics (Ontario, Quebec, Canada).
Tereses goals in choosing total automation were not primarily financial but rather were focused on increasing volume and improving service and quality. If you are just doing it to eliminate staff, you wont get much buy-in. We wanted consistency, particularly in delivering results to the ER, he says. He communicated a plan, with targets and deadlines, to staff.
Berman suggests that looking at a labs business objectives before automating is a good idea. It may not be just a question of automation, but what to automate, he says.
Many labs are already working with analyzers. A lot of chemistry and hematology is already automated. These are some of the most common tests, and these instruments have been the most widely adopted, says Jay Abrajano, LIS product manager, CCA (Calabasas, Calif). He cites liver panels and CBCs as examples. Microbiology and blood banks also use automation. Abrajano notes that to a lesser extent, genomic and proteomic testing has been automated.
Total lab automation involves a tracking mechanism that moves samples around the lab, handling the specimen preparation and delivering the samples to connected analyzers. This includes sorting, labeling, and aliquoting. Tracking systems are more than 50 percent faster in the preanalytic process, and 80 to 90 percent of the samples can go through without tech interference, says Kristen Hoefer, senior marketing manager, automation, informatics, and consulting, Abbott Laboratories (Abbott Park, Ill).
LIS and data management systems help with the postanalytic stage of specimen processing and result delivery. Informatics allows the collection of data at a single storage point from where it can be delivered anywhere. Multiple tests can be collated into a single report, says Henry Oglesby, vice president of operations for Orchard Software Corporation (Carmel, Ind). Decision-making software can then verify these results, freeing up staff to handle the 20% or so that require additional attention.
In deciding which processes to automate first, its important to understand where the greatest challenges lie. Automating the wrong processes can result in high investment with little return, says Hoefer. She believes that for many labs, bottlenecks occur in the preanalytic and postanalytic stages. It may be that a lab needs to re-engineer its processes even before adding technology. The track is often seen as the most valuable piece of automation, but this is not necessarily so. Just adding postanalytic software can increase the volume a lab is able to handle, she says.
Terese notes that Elmhurst considered front-end pieces and work cells for the lab but decided that to achieve the hoped-for benefits, total automation would work best for them. It was in connecting the front end to the analyzer where we really saw a reduction in steps, he says.
Research Before Buying
Terese worked with his vendor to make sure that everyone understood the process and goals. Error-risk steps were flagged during the workflow evaluation. Vendor representatives came in at 5 oclock in the morning because that was our busy time, he says, noting that vendor flexibility is key to developing a customized solution.
Terese also evaluated vendors and products by considering reputation, quality, and price. What experience does the vendor have in the market? What do its customers say about their experiences? How easy is the system to use? What level of troubleshooting is required to maintain the system? The most critical factor was whether they understood our process and whether it was worthy of automation, he says.
Facilities considering automation will also want to look at the scalability of systems. Will the lab want to add storage components; and if so, are they available? Can the system expand into other areas? Can it switch samples to analyzer-compatible racks? Are components, such as racks, interchangeable? How will it interface with current systems? asks Abrajano.
Any lab looking at middleware needs to make sure early in the process that the new system can communicate with the LIS. While communication parameters may come close, there are often additional fields or requirements that go beyond what is expected of an analyzer and could cause a delay in implementation or an increase in development costs on both sides, warns Oglesby. Facilities should ask their LIS vendor if they have interfaced with the systems they are considering before selecting a product.
The lab definitely wants to avoid manual entry, so you need to think about integration, says Liptrot. Manual data transcription incurs an error rate of 3% each time the data is transcribed. If not electronic, results will need to be printed or transcribed and manually delivered.
Eliminating manual tasks increases the walkaway capability of staff, meaning that fewer people are needed to do the same work. This is a big bonus for an industry that is suffering a severe labor shortage.
The average age of a technician these days is 38. More than 85 percent are close to retirement age. And over the last 20 years, training programs have decreased by 65 to 70 percent, says Abrajano, citing statistics from a recent presentation he delivered. At this rate, we are not going to be able to replace people exiting the lab field, so the current shortage can be expected to get worse. Automation is a solution to address this issue, he says.
Training staff on automation is not much different than training them on an analyzer, notes Terese. Initially, it looks complicated and people get nervous, but after a month or so, they love it, he says.
The Elmhurst lab has one key operator, or keeper of the line, who completed additional training. Berman notes that typically Beckman Coulter asks two people to complete a 1-week program, which is followed by another week training people in the lab
Liptrot, whose company offers automated data management, says there is not a tremendous learning curve with these solutions. Much of it runs in the background, and when it does require interaction, the electronic form mimics the paper, he says.
Training follows the systems setup, the length of which will depend on any necessary renovations. The space needed is typically no more than what the general chemistry and hematology labs usually occupy, says Terese. Elmhurst had the space, but, as an older facility, it was divided into four rooms. The area needed to be renovated to create an adequate footprint. Now complete, the location of the inlet is four steps from the pneumatic tube station.
The process took 8 months. We signed the agreement at the end of March 2002 and went online in November, says Terese, who adds the project could have been completed in October, but they waited an extra week to allow the keeper of the line to return from his honeymoon.
Automation can move samples around the lab and reduce the potential for error or biohazardous exposure.
The combination of hardware and software providing Elmhurst with its total automation is key to its success. Everyone thinks of automation as instruments, but its important to look beyond instrumentshow are the results significant? asks Brian Keefe, MT, ASCP, director of marketing for the clinical products division of Psyche Systems. LIS, he says, has historically assisted in the manual parts of turning out results. LIS and data-management modules can perform autoverification based on a customized set of rules, and this decision-making capability is expanding into pathology. Web-hosted systems will make even the process of owning an information system automated, says Keefe.
But physical lab processes will also continue to be automated. Hoefer suggests that radio frequency identification (RFID) is the hot new technology for tracking and routing samples through the lab. RFID is like a bar code, but it holds more information. Because its based on radio frequency, samples can be found anywhere, even in storage, she says.
Keefe predicts that esoteric tests, such as those for allergies or hepatitis, will become automated. They dont have a lot of volume, but they tend to be expensive and labor intensive, he says.
Automation allows smaller labs to compete. More labs across the spectrum of size are looking at automation, from the 250-bed community facility to the large teaching hospitals. Everyone is looking at automation, says Beckman Coulters Berman. And with the benefits to be reaped, why shouldnt they?
Renee DiIulio is a contributing writer for Clinical Lab Products.