“The good and bad of lab automation….”

A little while ago I said I would summarise the potential quality benefits of “Total Lab Automation” (TLA) systems (e.g. Kiestra, Biomeurieux, Copan Wasp). You might be involved in a business case for one for your own laboratory. Even if you are a trainee, this is the sort of stuff I would be asking in an exam question, not the nuances of the CAMP test.

I have summarised what I think are the main advantages below. Please feel free to add any extra you can think of in the comments section.

  • Plate Tracking: Each plate has a  comprehensive electronic audit trail attached to it, including when it was inoculated, incubated, imaged and read, and by whom.
  • Less menial tasks: Gets rid of finding appropriate plates for each sample, carrying the plates to and from the incubator, “putting up” of specimens, and other repetitive, manual tasks.
  • Better plate spreading. Automated spreading performed by machine will almost always succeed in better use of the whole agar plate and improved isolation of single colonies. It will also be a standardised procedure.
  • Less plate contamination: As the plate has less manual handling and less time spent with its lid off, the risk of plate contamination is much reduced, which is very important for those “sterile site” specimens.
  • Storage of digital images of plates. Plates eventually deteriorate, images do not, and images can be stored to be viewed again at any satge in the future depending on how long you want to store them for.
  • Standardised incubation times. No matter when the plate was inoculated, the system will image the plate after a pre-set incubation time, and thus allow plate reading. This in turn will allow reduction in turnaround times for specimens. The old concept of Day 1, Day 2 etc plate reading should disappear and be replaced by 1st reading, 2nd reading….
  • Less time out of the incubator. the plate goes straight into the incubator when it is inoculated, and essentailly stays in the incubator whilst it is being examined. No hanging around and very little downtime.
  • Remote plate reading: The system should allow you to view the plate images from anywhere, including home.
  • Plate checking: As the person reading the digital image is usually different from the person doing any further work-up, it allows the plate to be checked in case something was missed initially.
  • Plate interpretation: This is still in the developmental stage but software is now available allowing rapid detection of plates with no growth, and plates with colonies of a particular colour. Further development in such software will eventually lead to massive gains in efficiency.

I am sure there are others I have not thought about off the top of my head, but it’s a start that you can add to.

Sounds like a no-brainer. However one must balance all these advantages against two main disadvantages. These disadvantages essentially apply to all forms of automation.

  • Redundancy of staff: Whether staff are made redundant or not due to the implementation of such a system, the fact is that this type of automation will get through more specimens with less staff required. Some (managers) might see this as an advantage, but from a people point of view it is a big downside.
  • What happens if it breaks down? Because of the above, and because it is a complex operation, the consequences can be severe if the system goes down for any length of time….


8 thoughts on ““The good and bad of lab automation….”

  1. We have one of these automation systems… So just a couple of points…. The Inoculum size it uses means often no single colonies but confluent growth on the agar… They are designed to use expensive chrome agars, plate checking by the second reader yes, but only the plates sent for follow up work, so first reader has to recognise significant isolate, if not then not checked by other reader, follow up person may be a tech not a scientist, ….. Inoculation steps quicker, timed incubations and tracking all good… Still many improvements to be made but these are happening fast.

    1. Many thanks for your comments Wendall. Our system arrives second half of 2015 so will be interesting to see how the reality compares with the theory!

  2. Automation will be the key however to ensure quality processing of samples and consistency in interpretation of cultures and sensitivity testing…. And should be embraced…

  3. What about flat digital image vs holding plates in your hand and using your natural stereoscopic vision? And losing the dimension of smell?

    1. Hi Anthony. Don’t think we are really supposed to be smelling plates these days (although we all do!). Agree though with stereoscopic vision. As far as I am aware the current systems only take an aeriel photograph so you are definitely going to lose something here. Newer developments will include photographing the plates from different angles although this may be a few years away. And there is always the option to call the plates for manual review, although I suspect users will want to make this the exception rather than the rule. Many thanks for your comments, much appreciated.

  4. Automation should be looked at as if it will provide a benefit. Will it decrease TAT, will it enable a better flow within the laboratory, if the answer is no you are turning a manual mess into an automated mess. Also what can be automated, does it readily interface with your LIS. There is much more that the glossy brochures, the hype by the company trying to sell you the product. It may be worth automatic some of the process but not all. None of the systems are totally automated they still involve picking the plates out and grabbing stuff for maldi or biochemical ID. I draw you to an article on James Westgards
    site on when not to automate. I am also wary of the belief in these systems and the liability of management with the amount of sunk cost in such a project so that they will not let a bad project fail if it is a bad project. And oversell a few points to give the illusion of effecieny.

    1. Yes, automation is not perfect by any means, nor is the bacteriology culture process completely automated yet as you point out. It is very important to set up your systems around any new automation, as opposed to trying to fit automation into your old systems. Such an approach is doomed to failure. Before considering something like Kiestra, a long period of pre-evaluation is required to ensure it is a good fit for your laboratory. Such an evaluation can take 2-3 years. However if you ask most staff that work in Kiestra labs whether they would go back to the older more manual systems, I am fairly sure the majority would say no.

Leave a Reply

Your email address will not be published. Required fields are marked *