In total laboratory automation (TLA)10, preanalytic, analytic, and postanalytic phases of laboratory testing may be combined into an integrated system such that specimens are processed, tested, and even stored with minimal user intervention. Given the pressures of an ongoing workforce shortage of laboratory professionals, laboratory automation offers an attractive, albeit expensive, solution that laboratories are increasingly considering in planning for future growth and work flow requirements. In an ideal system, TLA handles routine, repetitive steps—leveraging the quality and efficiency obtainable in the manufacturing industry and freeing operators to focus on specialized testing that benefits from their unique training and expertise. A variety of laboratory automation solutions have been available globally for decades, with technologies that have advanced based on engineering innovation and practical trial and error. To address the difficulties and benefits involved in implementing and sustaining TLA in a clinical laboratory setting, we invited a group of 5 experts to share their perspectives on laboratory automation and provide real-world advice based on their experiences with TLA at their respective facilities.
In what ways has automation impacted your clinical laboratory?
Giuseppe Lippi: Automation has completely revolutionized the organization of my laboratory. The main advantages that we have observed are improved standardization, more simplified and efficient approaches for managing work flow, improvement in the performance of complex tests, better high-volume test management, shorter turnaround time (TAT) by elimination of some manually intensive preanalytical steps, provision of valuable walk-away time, and reduction of personnel costs (i.e., laboratory technicians and subsidiary staff), along with notable reductions in errors and biological risks attributable to manual handling of specimens. Another important advantage is represented by more efficient management of reruns and reflex testing, which can now be performed automatically with little user action needed.
Carey-Ann Burnham: Historically, the clinical microbiology laboratory has lagged behind other areas of the clinical laboratory with regard to automation. For many years, microbiology automation was focused primarily on continuously monitored blood culture systems and automated susceptibility testing. Recently, TLA systems for microbiology have emerged. These automate inoculation and streaking of culture plates, movement of culture plates to incubators, and these incubators are coupled with high-resolution cameras to image the culture plates. This is a dramatic change for clinical microbiology laboratories and has forced changes in work flow. For example, certain specimen types that were historically batch processed can now be analyzed in a continuous work flow.
To maximize the benefits of this technology, cultures should be read on all shifts, whereas historically most microbiology laboratories have read cultures on only the first shift. With this automation, we have also moved away from a model where one individual “owns” each culture, and different technologists are responsible for various aspects of the culture workup. The laboratory is becoming accustomed to reviewing culture plates on computer monitors and has learned the TLA software—both of which have a learning curve. However, this is potentially safer, as it reduces the chance of exposure to organisms that could be infectious to laboratory workers.
Finally, we have observed that with immediate and consistent incubation, we are able to detect microbes faster than with conventional microbiology, and we are seeing enhanced recovery of fastidious microbes, such as Neisseria gonorrhoeae, in urine specimens. Automation has had a major impact on our laboratory, although in many ways we are just beginning to understand that impact and realizing the full potential of the automated system.
Charles Hawker: The principal impact was enabling our laboratory to handle an expected, highly significant increase in volume. Since automation implementation was first being evaluated and planned (1996), our daily volumes have grown at least 10-fold. Although this anticipated growth was the main reason for implementing automation, we clearly expected and achieved significant improvements in TAT, quality, and productivity. Today, automation is a way of life in our laboratory. Our automation engineers have completed numerous projects for automated equipment requested by various laboratory sections, in addition to building most of the new robotic devices in use on our system. Most importantly, employees recognize the value of automation and are not concerned about “being replaced by robots.”
Robin Felder: Automation is an integral part of my clinical laboratory as well as my basic biomedical research laboratories. We find that the principal benefit is the reduction in errors and the improved precision of our analytical data with robotics. We also find that any progress with implementing a Six Sigma quality control program can be “locked down” with automation. This obviates the backsliding on quality when new untrained laboratorians are hired.
Octavia Peck Palmer: With over 4 million billable tests performed annually, automation has had a positive impact on our laboratory. Since 2009 (when we first implemented TLA), automation has allowed us the opportunity to consolidate testing across our health system, increase in-house test menu offerings, enhance our preanalytical quality check of specimens (hemolysis, lipemia, and icterus), decrease test result TAT, optimize add-on testing (specimen retrieval and reanalysis is more efficient), streamline result reporting, and increase the efficiency of organizing specimens that are shared between different laboratories in various locations. Importantly, in many cases there has been a reduction in the volume of specimens that our clinical laboratory staff manually handle, which translates to a reduction in biohazard exposures and risk of injury.
Implementing an automation solution can be a challenge, particularly when working within a limited or current space. What practical advice would you share in helping to identify and prevent potential problems?
Charles Hawker: Although implementing automation in a new open space is ideal, this rarely is an option. Most laboratories must install a new or upgraded system in their existing space while maintaining their service obligations. Fortunately, most automation vendors have considerable experience doing this and can offer competent assistance in developing a plan. If new analyzers are being acquired along with the automation, one approach might be to install and validate the analyzers separately before installing the automation. This allows the laboratory to complete the transition to new analyzers at a pace that ensures service is maintained. Then the automation can be installed, tested, and connected to the analyzers.
Giuseppe Lippi: Space availability is indeed one of the most important, limiting factors in developing a model of laboratory automation. Ideally, the best possible scenario entails the so-called “open space,” where laboratory assembly lines can be developed without physical constraints and can be customized to more suitably support the local organization. The most relevant challenge here is that many hospital laboratories, including mine, were designed and built when automation was still in its infancy (i.e., nearly 30–40 years ago). Therefore, developing a lean approach within an unsuitable environment is always a challenge. We have overcome some of the problems by using a model of “flexible” automation, designed around the available spaces and, whenever possible, by readapting the local environment and removing physical barriers. Additional critical issues are the risk of overcrowding personnel at workstations and the excessive noise and heat generated by many analyzers located in the same space. Therefore, a preliminary analysis of workforce location, waste disposal, heat generation, and soundproofing is mandatory before establishing a definitive layout.
Octavia Peck Palmer: It is important to remodel the clinical laboratory space with both your current needs and potential future needs in mind. Trying to make an automation solution “fit” in a space that is not appropriate will lead to issues in work flow. Issues include, but are not limited to, inappropriately increased indoor temperature because of poor ventilation, inability of the floor to handle the weight of the instrumentation, and inefficiency because of the crammed space. Have the full commitment of the hospital to ensure the space needed is realistic and can be achieved.
Carey-Ann Burnham: TLA systems for microbiology are very large, and it may be a challenge to fit these systems into an existing microbiology laboratory space. Our laboratory was fortunate because we were moving to a new space at the same time we were transitioning to automation. There are several requirements to support a TLA system that can be easy to overlook, such as special electrical requirements (and electrical backup), building requirements to support the weight of the system, air handling and compressed gas, and appropriate server and networking requirements. The laboratory should start working with facilities management and information technology teams as early as possible in the process of implementing automation.
For microbiology specifically, the laboratory will need to streamline culture collection devices (such as liquid swab collection systems and urine collection devices) so that the containers will be compatible with the automation system. Ideally this change would occur well in advance of the automation transition. Finally, the laboratory will need to identify if the culture media they have historically worked with will be compatible with the TLA, or if another brand or even custom media will be required.
Perhaps the biggest fear in automation—particularly TLA—is the risk of prolonged downtime with automation failure. How have you approached this concern? What can laboratories do to address single point-of-failure risks?
Octavia Peck Palmer: Once you bring in automation you will have downtimes, and you must have the support of the laboratory automation vendor to ensure these downtime(s) are limited. Ongoing dialogue between the clinical laboratory and the company is imperative to troubleshoot issues. There are several factors one should consider that impact downtimes.
1. Customer service: It is important that the automation system vendor you are considering has a positive track record in customer support. Specifically, review the downtime data, which you can obtain both from the company and your colleagues, and gauge the expertise of the technical support field engineers. Also, you should work with the company to set up an onsite area where parts can be stored to reduce waiting time if a key part must be shipped from a long distance. Automation systems that provide real-time remote diagnostic checks are also imperative to reduce the duration of downtime or ideally prevent it from occurring.
2. Appropriate instruments for your laboratory's test volumes: Visit laboratories using the system and see it in action. Assess how automation would work for your clinical laboratory's test volume and have a real-time assessment with the company. If test volumes warrant, have more than one instrument type to ensure testing is always available. Electrical, water supply needs, and temperature conditions within your laboratory space, may also impact the automation system's efficiency.
3. Clinical laboratory staff training: The clinical laboratory staff must have adequate time to train on the system before going live. Shortchanging the staff on training can lead to downstream issues and potential downtimes. Do not downsize before bringing in automation—the clinical staff must be confident in running the system and maximizing the features.
Carey-Ann Burnham: Downtime is indeed something that we worry about. For microbiology TLA, an entire system failure is uncommon (thankfully), but downtime may require manual plating or manual reading of cultures. We do have redundancy within the system (e.g., more than one incubator for each incubation condition) as one risk mitigation approach. Downtime procedures and ongoing training in manual microbiology processes are still required. In addition, a major change for our laboratory is that we now have a full-time engineer in our laboratory and a supply cabinet that contains almost every part for the TLA system. Thus, the engineer can respond quickly when issues come up so that we can get back up and running. Even with all of this in place, downtime does occur from time to time. This can be incredibly frustrating and time-consuming for laboratory staff and can result in delays in culture TAT.
Robin Felder: It is important that laboratories have an engineer on staff who is trained to service the robotic system. The vendors are more than happy to train engineers because it reduces their burden of needing to support a 24/7 service contract.
Giuseppe Lippi: System failure is indeed the most important risk in laboratory automation, especially for those laboratories that have remodeled their organization around a particular TLA. When we reorganized our laboratory around a model of flexible automation, the identification of one or more emergency plans became a leading issue. This has been achieved by implementing analyzers that can also work as stand-alone instrumentation, despite normally being linked to the automation lines. Although the number of personnel required has been reduced in this model (by approximately 20%), the workforce has been trained to enable adequate responses to emergency situations, such as when manual loading of the specimens is needed in the event of system failure.
How should the practice of priority (i.e., STAT) testing be considered and/or incorporated in automated solutions? Have you seen efficiency in automated processes such that STAT testing is no longer necessary?
Robin Felder: We have virtually eliminated STAT testing using automation. For sites that frequently order STATs that are legitimate for critical patient care, we set up satellite laboratories to take care of their needs.
Giuseppe Lippi: The management of urgent testing depends on many variables, such as the type of healthcare facility served by the laboratory (e.g., tertiary care, acute or general hospitals, presence of short-stay and emergency units, presence of pediatric wards), as well as by the type and volume of testing. In the case of TLA, for which most tests are performed within a “core laboratory,” the TAT of urgent testing becomes critical. Therefore, an accurate analysis of work flow is also critical to define whether (a) both routine and STAT samples can be analyzed simultaneously without priority flagging, (b) STAT samples should be prioritized for testing, or (c) certain analyzers should be dedicated only to STAT testing. Although the first solution seems the most practical, the high volume of samples in my laboratory has persuaded us to use priority flags for urgent samples, which allows them to enter the automation line using a more rapid route for loading and analysis.
Octavia Peck Palmer: The clinical laboratory's current STAT testing should be considered when reviewing the various automation solutions. The automation solution should have the capability to perform STAT testing with a dedicated point of entry. Our institution has not eliminated STAT testing, as we process a unique mix of specimen types. We monitor the TAT times between STAT and non-STAT testing, and for some assays (analysis times do vary among assays) we have found no significant differences.
Charles Hawker: There is evidence in the published literature that well-designed automation can enable laboratories to deliver results within the time frame defined as STAT by an institution. Most automation systems can load STAT testing onto the automation in a designated rack or lane that allows the system to prioritize the handling of those specimens. However, certain factors may come into play, making generalization difficult. For example, some institutions have preferred manual centrifugation in lieu of centrifuges on their automation system.
Carey-Ann Burnham: In the world of culture-based microbiology, where most specimens must incubate 12 to 24 h before being evaluated, “STAT” does not really apply. However, historically the laboratory has placed priority on plating and incubating some specimen types over others (e.g., tissues before urine specimens). With TLA, our focus is shifting to more of a “first in, first out” model of loading specimens for inoculation. For reading plates, the system software can be set up so that highest priority specimens are read and reported in advance of lower priority specimens. However, with the enhanced rate of culture growth in the automated incubators, as well as time saved by eliminating tasks such as organizing culture plates, it is possible to complete microbiology cultures faster than using conventional methods.
An automated clinical laboratory bears many similarities to a factory—one that is “manufacturing” test results from an input of clinical specimens that are processed by instrumentation. How can manufacturing improvement techniques such as lean and Six Sigma be used to improve quality in automated laboratories? How should laboratories address clinically unique specimens (e.g., pediatric and nonstandard tubes) when processes are otherwise designed for standardization?
Charles Hawker: The clinical laboratory is really no different than a productive manufacturing plant or service operation. There are inputs (specimens) and outputs (test results), and the objective is to report the test results as quickly as possible, with the highest possible quality and minimal expense. By eliminating waste (lean) and maximizing quality and efficiency (Six Sigma), clinical laboratories can clearly improve quality. There is ample published evidence supporting this. Going further, laboratories should continually use these techniques and seek ways to improve quality. In fact, I would suggest that no laboratory should implement automation without first undergoing a review of their work flow and using these techniques to improve their processes.
Unfortunately, there are not good answers to using nonstandard containers. Automation systems are designed for tubes that generally range from 11 to 12 mm up to 15 to 16 mm in diameter and from 75 to 110 mm in length. Tubes not within those ranges probably cannot be transported and handled by the various robotic systems on the automation. Moreover, if tubes do vary in size within those ranges, they may pose challenges for bar code readers or other devices on the automation system. It is important that an automation system be validated for the range of tubes that it will be handling. And, to prevent surprises, the laboratory should determine from the automation vendor before signing a contract whether the system can handle the expected tube sizes. Pediatric specimens can be transported using tube extenders, but analyzer probes may have difficulty aspirating specimens from these containers without special programming and engineering. This capability has been described in a 2005 report in Clinical Chemistry by Demir et al.
Octavia Peck Palmer: Studies have demonstrated that lean and Six Sigma can positively impact the clinical laboratory's work flow and efficiency. Determine the top reasons your laboratory wants to implement automation, and then perform a process review (walk through). Identify areas where you can begin to review process problems and the desired outcome you would like to achieve. For example, the processing area may be located on the other side of the room away from key instrumentation. Reorganization of the laboratory to decrease the space (steps or time) one takes to place the specimen on the automation will improve efficiency. Continuous monitoring of your work flow once automation has been implemented will continue to improve the laboratory's efficiency.
If the laboratory does not have an automation platform that provides pediatric adapters or is unable to utilize polypropylene tube extenders, there are commercially available pediatric tubes that have an evacuated space for lower volume collection but are the standard size to fit on the automation.
Robin Felder: One should not automate a laboratory that has not successfully implemented a lean and Six Sigma program. Automation will greatly assist with maintaining the quality associated with these programs. There are always workarounds for nonstandard specimens, e.g., pediatric specimens can be inserted into adult-size tubes and the automation programmed to deal with the change in sample height.
Automation can be expensive, although there may be savings associated with labor. This is particularly significant given the workforce shortage seen across the clinical laboratory community. There are relatively few published studies, however, that investigate the financial impact of TLA. How should healthcare facilities and clinical laboratories approach cost justification when considering automation? Have previous investments in automation lived up to financial and quality expectations?
Carey-Ann Burnham: Cost justification is always difficult. Although there are aspects of laboratory testing that are easy to quantify, other potential benefits of automation, including reduced repetitive-motion injuries, improved reproducibility, improved TAT, and reduced errors, can be more challenging to assign a dollar value to. Many health systems are consolidating microbiology services, and justification of TLA in this setting is typically focused on handling additional specimens without the need to expand staffing. Cost justification models can consider throughput, staffing, supply costs, and TAT. In addition to considering the actual cost of the equipment, the cost justification must include ongoing maintenance costs, which may be a significant percentage of the overall purchase price on an annual basis. To date, there are no published data demonstrating that automation in clinical microbiology improves patient outcomes. Evaluation of the impact of automation on outcomes is needed, and this information may inform cost justification models in the future.
Robin Felder: The challenge with cost justifying automation is that the accounting systems do not put a value on reduced errors, improved overall quality, and reduced TAT. An overarching data gathering for an entire medical system would have to be conducted to look at the effect of a more efficient laboratory on patient discharges and physician and nursing efficiency.
Octavia Peck Palmer: Cost justification should be approached by identifying the key factors that are hindering efficient work flow in the laboratory. It is important to assess the nonfinancial benefits of utilizing automation: reduction in staff stress and risk of exposure/injury, increased confidence and pride in streamlining testing in a timely manner, and allowing more time to devote to troubleshooting critical cases.
Giuseppe Lippi: The widespread incorporation of laboratory automation was mainly driven by the concept derived from industry that robots decrease manufacturing costs and can perform tasks at a pace no human could possibly manage. The main issue to be understood here is that the financial benefit from laboratory automation can be realized only after the recovery of the capital investment in the project. There are several published studies that have demonstrated that a suitable model of automation can have a positive impact on costs.
Charles Hawker: Generally, automation has lived up to expectations whether financial or quality. If that were not the case, the thousands of systems in laboratories around the world would not have been installed. There have been a few examples of laboratories in which the automation has been a poor investment. Typically, these have been instances of purchasing the automation in anticipation of a significant gain in volume that did not materialize. If laboratories base their cost justification on realistic workloads and labor savings and use lean and Six Sigma quality improvement processes, they should be able to justify the investment and, more importantly, achieve the predicted savings.
While the diagnostic industry is seeing remarkable advances in point-of-care and microfluidic technologies, core laboratory automation seems to keep getting bigger and remains “macrofluidic” in design. Are current automation solutions simply lagging behind in technological advancement, or are there advantages and economies of scale in conventional technology that benefits current design? Which potentially disruptive technologies could impact our current model of clinical laboratory automation?
Robin Felder: Current automation systems need to be redesigned for speed in routine testing as a first step. Next is the move toward miniaturization to reduce specimen volumes. Essentially, we should convert automation to handle only pediatric-size specimens. The most difficult part of designing an ideal point-of-care alternative is the difficulty with the human interface. Fortunately, we are going to experience a major disruption in conventional laboratory automation as we move toward personalized medicine. Wearable and insertable sensor technologies that will measure a broad menu of analytes and human physiological parameters are currently going through Food and Drug Administration submissions.
Charles Hawker: Core laboratories are all about throughput, efficiency, quality, and TAT. Current automation solutions clearly deliver on those objectives. To some extent, however, the current macrofluidic designs are driven more by the analyzers that laboratories are using, as these laboratories still use conventional tube sizes and larger sample volumes. The automation is supporting the current analyzers. The automation also prepares labeled aliquots for other ordered tests that may be performed manually in the local laboratory or sent out to a reference laboratory. The bundling of all these functions into today's automation may make it more difficult to adopt a revolutionary change. Microfluidics will probably find different niches in the laboratory industry other than the core laboratory, at least for some time.
Carey-Ann Burnham: I think that microbiology testing in the future will look very different than the microbiology testing of today. With advances in technology and infrastructure, as well as the growing emphasis on antimicrobial stewardship and avoiding unnecessary antimicrobial use, I think that we will have 2 types of testing: “culture-independent, near patient testing” paired with “central laboratory testing” for culture-based work. Automation will continue to evolve for these 2 types of laboratory testing. The cost associated with these 2 types of testing will be very different, with the central laboratory testing having a lower cost per specimen.
Sequencing and molecular testing have many advantages, but ongoing phenotype monitoring is necessary to inform new molecular assays. In addition, not all types of antimicrobial resistance have a straightforward mechanism that can be measured using molecular diagnostics. I do think that the future state will involve more miniaturized testing, and measuring the host response to infection (rather than just focusing on the microbe). That said, I do not think that TLA used in microbiology today will become “outdated” during its expected life span; culture-based testing will be an important component of the clinical microbiology laboratory for the foreseeable future.
Giuseppe Lippi: The process of miniaturization inherently increases costs. So, the critical questions are as follows: What are the extra costs we are willing to pay for miniaturization? What are the real benefits of microfluidic technology in routine medical laboratories? Can smaller instrumentation achieve the same degree of analytical quality as conventional laboratory analyzers? These challenging issues can be summarized in a simpler question: Do we really need microfluidic technologies in the routine medical laboratory? At this time, my answer is no.
Footnotes
↵10 Nonstandard abbreviations:
- TLA,
- total laboratory automation;
- TAT,
- turnaround time.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: C.D. Hawker, ARUP Laboratories.
Consultant or Advisory Role: C.-A.D. Burnham, Monsanto, Thermo Fisher.
Stock Ownership: None declared.
Honoraria: C.-A.D. Burnham, BD, Accelerate Diagnostics.
Research Funding: None declared.
Expert Testimony: None declared.
Patents: None declared.
- Received for publication July 12, 2017.
- Accepted for publication August 8, 2017.
- © 2017 American Association for Clinical Chemistry
In this issue
