In clinical chemistry laboratories, hundreds of specimens in uniform tubes march like lemmings through a modular pre-analytic processor and onto automated analyzers. Quantitative results for 20 or more analytes are simultaneously reported in a matter of minutes, all without the touch of a human hand.
This type of sample-to-answer automation, commonly referred to as total laboratory automation (TLA), reduces turnaround time and increases the precision of results by eliminating such sources of human error as manual specimen handling, labeling, pipetting, and measurement of analytes using separate assays. Implementation of TLA in chemistry has been associated with a reduction in laboratory expenses, fewer errors, and improved patient outcomes.
Despite these advantages, workflow in the microbiology laboratory has remained a largely manual process. An astounding 32 percent of technologists’ time in a microbiology lab is spent on manual pre-analytical processing and plating of specimens, while another 10 percent is spent physically transporting specimens within the laboratory. By comparison, the technically demanding analytical examination and workup of cultures accounts for only 25-35 percent of technologists’ time.
Such a large investment of personnel resources for each reported result can easily outweigh the cost of reagents. Additionally, reliance on manual specimen processing and result entry can lead to mislabeling, specimen mix-ups, and transcription errors. Indeed, 87-93 percent of laboratory errors occur in pre- and post-analytical phases of testing.
Another challenge of manual culture analysis is the dependence on microbial growth, which typically requires 12-72 hours of incubation. In order to fit into a standard “work shift” schedule, most laboratories opt to read batches of plates once daily (typically on first shift) rather than at optimal incubation times. This artificial read time can lead to delays in result reporting.
Barriers to automation in microbiology
With all the benefits of TLA, why has the microbiology laboratory been so slow to adopt an automated approach? The answer to this question can be distilled down to two words: diversity and complexity.
An array of liquid (sterile and non-sterile body fluids), semi-solid (stool), and solid (tissue and bone) primary specimens, in addition to the ubiquitous swab specimens, are submitted for culture. Each specimen type is received in a different container and requires different pre-analytical processing and inoculation protocols. This diversity does not lend itself to traditional pre-analytic automation platforms. Analytic phase workup of cultures requires skilled technologists to recognize potential pathogens and determine what is significant in the context of each specimen. Subsequent identification of pathogens has traditionally involved multiple phenotypic and biochemical test reactions. This level of analytic and interpretive complexity is not easily amenable to current automated analysis methods.
Tools that overcome automation barriers
Two systems designed specifically for the microbiology laboratory, developed by BD Kiestra and Copan Diagnostics, respectively, overcome some of the existing barriers to automation. Both systems are true TLA, encompassing pre-analytic, analytic, and post-analytic phases of testing; however, each system takes a different approach to TLA.
The WASPLab (Copan Diagnostics, Murrieta, CA) consists of a multifunctional, self-contained “core” (the WASP) that is capable of automating all front-end processing tasks for liquid-based specimens. This design is ideal for retrofitting existing laboratories that have limited space. Key components of the WASP include a universal de-capping mechanism to accommodate most specimen containers, an onboard bar code printer for automated labeling of culture media, and two independent robots that vortex and/or centrifuge specimens prior to plating. Inoculated plates are automatically transported to “smart incubators,” where cultures are imaged using highresolution digital cameras at user-defined intervals. The overarching approach of WASPLab is to minimize human interaction with specimens, thereby maximizing workflow with limited staffing. This can be especially valuable during off-shifts and on weekends, when many labs are short-staffed but still receive a large volume of specimens. One limitation to WASPLab TLA is that liquid specimens are required. This limitation can be addressed by converting to liquid-based swab collection systems, such as ESwab, which are available for routine aerobic and anaerobic culture, stool, and respiratory specimens. Using this approach, up to 90 percent of specimens can be received in liquid form. However, the cost of conversion to ESwab and impact on other laboratory testing (e.g., molecular tests) must be considered.
The BD Kiestra TLA system is composed of individual modules dedicated to specific tasks such as media storage, specimen bar coding, inoculation, and plate reading/analysis. The various modules are connected by a conveyor belt system that transports cultures throughout the TLA system. This modular approach enables scalability in order to meet the specific needs of each laboratory, but also requires a larger physical space. Rather than completely removing human interaction, BD Kiestra TLA aims to integrate automated and manual processes. This integration enables repetitive tasks such as specimen labeling, plating, and streaking to be mechanized while maintaining technologist involvement in processing valuable specimens such as CSF or nonliquid specimens such as tissue or bone. Post-incubation cultures are automatically retrieved and transported to integrated technologist workbenches for manual workup of clinically significant organisms.
Given the differences in TLA approaches, a thorough assessment of each laboratory’s needs is critical when selecting a system. Considerations should include physical space, available staffing, total and peak specimen volume, and specimen collection devices currently in use.
What’s to be gained from TLA?
The immediate impact of either TLA system is seen in pre-analytical processing and plating of specimens. Implementation of TLA can reduce the need for fulltime equivalents (FTEs) dedicated to pre-analytical processing by 75 percent, enabling those FTEs to complete more sophisticated analytic tasks such as plate reading. Additionally, automated plate streaking using either steel loop (WASP) or rolling bead (BD Kiestra) method provides more reproducible quantitative colony counts and generates 5-16 percent more cultures with well-isolated colonies than does manual inoculation and streaking. This aids in culture interpretation and reduces by up to tenfold the number of specimens requiring subculture for downstream identification and susceptibility tests.
Central to both TLA systems is the incorporation of culture imaging within smart incubators and on-screen plate reading. With these features, cultures can remain at optimal growth conditions instead of being removed daily from incubators for manual examination. When combined with high-resolution stereoscopic imaging, the incubation time necessary to observe colony growth is reduced by 30-50 percent, and the recovery of fastidious organisms is up to 370 percent higher than with manual plate reading. In one study, early detection of colonies along with the integration of MALDI-ToF enabled identification of blood culture isolates following an 8-13 hour incubation period, resulting in adjustments to antibiotic therapy for 12 percent of patients.
Published yearly financial benefits of TLA implementation for a typical microbiology laboratory include elimination of two FTEs (est. $105,000), reduction of repetitive motion injury claims (est. $14,000), and reduced waste due to single-use pipettes and loops (est. $4,000).5 Similar hard data regarding the impact of TLA on clinical outcomes are currently lacking; however, reduced time to result and improved accuracy of other recent advancements in clinical microbiology such as MALDI-ToF MS and rapid identification systems for positive blood cultures have led to significant benefit to the patient, including reduced length of hospital stay and an overall decrease in mortality.
Overcoming on-screen plate reading obstacles
Two initial obstacles are often encountered by laboratories shifting to on-screen plate reading. The first is a shift in workflow from standard daily plate reading to a read-when-ready approach. Plate images are automatically sent to an electronic reading queue at designated times for review. Assigning a technologist to a reading station dedicated to initial reading of a triage of all cultures ensures real-time culture analysis. This way, positive cultures can be categorized and culture plates recalled to a bench for physical workup.
The second obstacle is adjusting to the appearance of colonies, hemolysis, and other physical characteristics of bacteria on a screen. While cultures are imaged using multiple backgrounds and lighting, there is often an adjustment period before technologists are comfortable with the appearance of these characteristics on the screen. This issue is easily addressed by implementing a period of familiarity training prior to the transition to on-screen culture analysis.
It is important to recognize that the benefits of TLA, including reduced turnaround time and concomitant benefits to patient care, are realized only if trained technologists are available to read and react to on-screen images in real time. This requires 24/7 staffing sufficient to manage downstream analysis and reporting of positive cultures, which remains a challenge for many laboratories. Automated colony recognition software, commonly referred to as artificial intelligence (AI), is beginning to take hold in microbiology TLA. These software applications compare culture images at time zero and defined incubation periods to identify bacterial growth and enumerate colonies. Current AI applications effectively triage cultures into those with no growth and those with user-defined growth, e.g., <10 cfu, 10-100 cfu, and >100 cfu. Following a validation study, negative cultures may be automatically resulted and discarded without human intervention, which improves efficiency and time to result. The use of chromogenic media for MRSA and VRE expands automated plate reading to enable recognition of positive cultures based on the presence of characteristically colored colonies. In the future, AI may be further expanded to provide quantitative results for multiple pathogens in a single culture based on recognition of colony characteristics such as color, hemolysis, and growth rate.
The microbiology lab requires TLA solutions as unique as the specimens and cultures that are routinely encountered. These systems are now available, and data are beginning to emerge regarding the advantages to workflow, cost, and time to result. Some growing pains can be expected in the adoption of and adaptation to TLA; however, the future appears to be bright and the possibilities wide for laboratories that embrace the new technology and approach to clinical microbiology.