Automating Science

How Changing Technologies are Changing Job Requirements

Would you hire someone with out-of-date skills? You might be doing just that. We are in the midst of a transformation of the way laboratory work is done. Those changes are not going to be incremental; they will require a major change in the skills and experience people need in order to be effective. This article takes a look at what working in an automated laboratory will be like and what the needed skills are based on the experiences of those working in facilities dependent upon successful implementation of automation technologies (which include laboratory informatics).

Automation has been shown to be an effective tool in improving productivity and reducing the cost of labor-intensive work. This has been demonstrated not only in manufacturing environments but also in clinical laboratory systems and contract testing labs where laboratory automation has been effectively implemented.¹

During e-mail and telephone conversations with people at contract testing labs, made in preparation for this article, the phrase “We couldn’t work profitably without automation” would come up. One lab stated that their prices would be two to three times higher without automation. The driving factor in clinical chemistry automation has been to improve productivity and reduce costs (Mt. Sinai Medical Center reported a ninefold increase in sample throughput with a fivefold per-test cost reduction.)² These are the same issues that drive any organization.

Properly done, laboratory automation is a useful tool in both research and testing environments. In fact, most laboratories would find it difficult to function without automation and computer systems. While some measurements are difficult to make in an automated environment (the effect of lotions on skin, for example), most laboratory instrumentation work is done by equipment with embedded computers in addition to data acquisition/reduction/reporting and management functions.

How will lab work change?

Automation has already had an impact on laboratory work. Those working in chromatography, for example, no longer measure peak areas or heights by hand, constructing calibration curves, or evaluate results manually. That work is done by computer systems, with the results shown on printouts or transferred to other systems. In many cases, sample preparation work is done by robotic systems that relieve people from labor-intensive efforts. An autosampler was released last year that carries out sample preparation functions within a small footprint. For the most part, however, these systems represent incremental changes to lab work, with people still providing linkage—the integration— between one automated task and another. Ideally, in a fully automated laboratory workflow, the system would carry out all steps of the analysis from the initial sample to the final result. This is not fantasy. This type of work is going on in clinical chemistry and contract testing laboratories today. In clinical applications, the systems are the result of decades of work in standardization and client vendor cooperation. We see the results of standardization in life science workflows that use micro-titer plates with standardized plate formats, allowing vendors to create a range of equipment with different capabilities (stackers, washers, readers, etc.) that can be put together into functioning systems.

The increased use of automated systems does raise one concern: trusting the automated equipment too much—the development of “push-button science.” We cannot let the complexity of systems intimidate us into not asking questions about how they work and produce results. As you will see below, laboratory professionals need to be educated so that they can challenge vendors and ensure that they are using the right products for their work. The fact that a vendor has produced a software package doesn’t mean it is right for your application. End users need to understand how it works, how it is converting instrument output into results, and whether it fully fits their needs.

One of the major changes in lab work is that people will be less involved in the execution of tasks and put more effort into managing systems. However, that is not the only change. The introduction of automated systems changes the way one thinks about laboratory work. For example:

• The development of testing methods has to be done with an eye toward their ability to work in an automated environment. This means one needs to think about how things are being done and whether or not they lend themselves to implementation and routine use with automated equipment. Techniques that are easily automated will be preferred.
• People will be concerned about both the science and the implementation of automated systems (e.g., whether they are functioning properly and tracking down problems). The work will be more along the lines of managing a production environment than typical hands-on task execution. This will require training that is more sophisticated, since they will need to understand how things work in addition to the science. As Diana Mass³ of Associated Laboratory Consultants put it, “What I have observed is that automation has replaced some of the routine repetitive steps in performing analysis. However, the individual has to be even more knowledgeable to be able to troubleshoot sophisticated instrumentation. Even if the equipment is simple to operate, the person has to know how to evaluate quality control results and have a quality assurance system in place to ensure quality test information.”
• More time can be spent evaluating results. The shift from hands-on task execution to knowledge-based work has been one of the promises of automation.

Changing skill sets

Questions about skills needed for work in an automated facility were posted last summer (2010) on LinkedIn discussion groups. The following are examples of some of the responses from clinical, pathology, pharmaceutical, and biotechnology professionals:

• “The ability to troubleshoot/think and validation experience are now the two skills that I look for most when hiring an experienced chemist.”
• “For today’s modern lab practices you should equally know all about your lab automation and their troubleshooting. Knowing in detail the software program of the instrument is a must. Good knowledge of the technical aspects will always put you in a better position.”
• “Risk management techniques, best laboratory practices and [being] software literate are skills that are required in a clinical laboratory at every level (technicians through managers) plus good old-fashioned ingenuity and critical thinking.”
• “Though automated systems have replaced manual work, I feel, the key personnel should have a basic understanding of the process. Software skills are essential and also the interpretation of the output data. I am writing this with my experience in R&D biotechnology. The technicians should also have basic knowledge on troubleshooting. We generate so much information through the automated systems that interpretation of important information becomes very critical and the onus lies with the supervisors.”

The following quote from Martha Casassa, Laboratory Director, Braintree Rehabilitation Hospital, (Braintree, MA) gives a contrast between clinical and non-clinical environments:

“Having a background both clinical (as a medical technologist) and non-clinical (chemistry major and managing a non-clinical research lab), I can attest to the training/education being different. I was much more prepared coming through the clinical experience to handle automation and computers and the subsequent troubleshooting and repair necessary as well as the maintenance and upkeep of the systems. During my non-clinical training the emphasis was not so much on theory as practical application in manual methods. I learned assays on some automated equipment, but that education was more to obtain an end product than to really understand the system and how it produced that product. On the clinical side I learned not only how to get the end product, but [also] the way it was produced, so I could identify issues sooner, produce quality results, and more effectively troubleshoot. I do not believe a specific degree level or amount of education is the key to success with automation in a lab setting. It is the type of training and the curriculum’s approach that [are] critical.”

We need to think differently about the education of laboratory managers and lab personnel. Business as usual is not going to cut it. We did a review of 88 job descriptions posted last summer for open positions, looking to see what backgrounds people were asking for. The bulk of the positions were in life sciences and evenly split between management and non-management positions. We were looking for keywords that pertained to automation and a regulatory environment. The chart above shows the results:

The list on the left contains the keywords that occurred (with one exception) versus the percentage of the number of job descriptions in which they occurred. The exception was “instrument automation,” which was an aggregate of occurrences (interestingly, some noted specific vendor products).

For an industry that is moving toward the increasing use of automation and informatics technologies, these are low numbers. We have moved past the point where “general computer,” word processing, and spreadsheet skills are a reasonable basis for competence in a modern laboratory.

Personnel requirements

Laboratory managers need to be conversant with automation and informatics technologies and the planning needed to design effective programs for the use of the technologies available. Beyond that, they need to understand their future needs, and how those needs match up against current product capabilities, well enough to advise vendors on how product characteristics and functionality have to be changed. This last point is essential to forging effective vendor-customer relationships. Vendors have their views of what the market needs based on what they see, but having customers who can articulate technology needs and integration requirements will advance the field much more quickly.

That point was noted in a recent meeting in which representatives from Velquest, LabWare, Accelrys and Rescentris discussed “Future Directions in Lab Informatics.” They agreed that more input and involvement from customers was needed in their planning and the shaping of the evolution of informatics systems.⁴ Successful vendor-customer relationships are going to be critical as laboratory work moves from manual effort to full automation.

Laboratory automation engineers are needed to assist in the planning, design, implementation, and support of laboratory automation systems. This is something the ILA has been working on for some time.

Laboratory personnel need to be trained in the products and technologies as well. We need to move beyond the ability to use preplanned settings for equipment, to having them understand what the equipment does “under the hood” so that problems can be traced or avoided. They need to understand how things work—not at the programming level, but how data is acquired from instruments, how peaks are detected, how baselines are drawn, and how setup parameters can affect those functions.

The move toward automated laboratories is more than purchasing equipment. It is the effective use of products and technologies by people educated to work competently in that environment.

References:

1. Total Laboratory Automation, Michael Bissell, Laboratory Medicine Grand Rounds, downloaded from iTunes U – ResearchChannel.

2. Sarkozi, Simson, Ramanathan, “The Effects of Total Laboratory Automation on the Management of a Clinical Chemistry Laboratory. Retrospective Analysis of 36 Years.” Clinica Chimica Acta, Volume 329, Issues 1-2, March 2003, pgs 89-94.

3. Diana Mass was formerly Professor and Director (Retired), Clinical Laboratory Sciences Program, Arizona State University.

4. “Future Direction in Laboratory Informatics,” Joint LIMS/LI & ALMA Meeting, Sept. 27, 2010, held at Millipore Corp., Bedford, MA.