Tag Archives: automation

“The chicken and the egg”

I am currently at the ECCMID conference in Copenhagen. So far, and in my humble opinion it has been better than last year’s ECCMID in Barcelona; better conference facilities, more seating, nicer food, better use of Information Technology etc.. Even though the conference is only in its second day it has already provoked several ideas and concepts for me to work on when I get back to New Zealand.

One continuing issue I have though is the difficulty in differentiating between the sessions sponsored by industry and the independent sessions. It is an important distinction to make, for obvious reasons.

During a couple of the industry sponsored sessions, both involving the promotion of new automated platforms, I have heard the following quote: “The shortage of skilled scientists has prompted the need for automation”or something to that effect.

Hmmm.., I am not buying that one.

Automation in the world of microbiology has several benefits (see this article for more), relating mainly to standardisation of processes and reduction of errors. You can also be sure that automation is cost effective compared to the traditional methods, otherwise it would not happen.

These, along with the technological advances that have made automation feasible, are the principle drivers in this direction.

The lack of skilled staff I suspect is not one of the main drivers, but is however still used as an “excuse” for automation, particularly by the companies manufacturing and promoting these systems. 

What automation certainly does do is reduce the number of skilled staff required to process the samples. It also changes the skill sets needed by both incumbent and trainee scientists. 

So for me, in this conundrum, the automation comes first, and the effects on scientists are the consequence, not the other way round…..



“Filling in the Gaps”

If you are in any way involved with clinical microbiology labs, you will be aware of the tsunami of interest in “complete” bacteriology  culture automation. Examples of such systems are Biomeurieux FMLA, BD Kiestra TLA and Copan Wasp, along with a host of others. One of them will be at a lab near you in the near future, if not already…

A lot of these systems claim “total” lab automation. They certainly represent progress, and there are a lot of quality benefits which I will go into in another post.

So where do the new systems fall short of “complete”. Let’s go through the process..

  • Fetching the correct plates.
  • Barcoding the plates.
  • Uncapping the correct sample.
  • Inoculating the plates.
  • Taking the plates to the appropriate incubator.
  • Incubating the plate for a certain amount of time.
  • Imaging the plates.
  • Interpreting the plates.
  • Picking colonies off the plates for ID and susceptibility.
  • Interfacing results to LIS.
  • Validating the results.

The three areas highlighted in red are where these systems are not quite there yet….

Interpreting the plates: There is rapid development in software to aid scientists in interpreting agar plates. Good progress has been made in reading negative plates, reading chromogenic plates, and reading and interpreting susceptibilities. However there are still a lot of plates for which humans are still better than computers for reading. In 5 years time however, I would expect that the majority of plates will be read automatically.

Picking colonies off the plates for ID and susceptibility. At present these newer systems tend to return appropriate plates from the incubators to benches in order for the follow up work to be done. However colony pickers are being developed which will automatically pick marked colonies off plates. These picked colonies will then be used to automatically inoculate plates /broths for susceptibility testing and also a MALDI-TOF plate for identification. It is difficult to say when this technology will be commercialised, but I would guess within 5-10 years.

Validating the results: As information technology improves and the application of “logic” rules to the results becomes ever easier, I would anticipate that the vast majority of bacteriolgy culture results will eventually be released automatically, without human checking. Again I would anticipate this to happen in the next 5 years.

So there you have it. I think it is a little disingenuous for companies to claim that they already have “complete” bacteriology culture automation. Not yet, but it’s coming. I would predict by 2020 at least some samples will go through the whole bacterial culture process without being seen or touched by humans. It is a scary thought, but there is no point in denying that it is going to happen. Instead, embrace it and involve yourself with it. It is what the biochemists and haematologists have already been through in the past couple of decades….



“The Microbiology Lab in 2050….”

What will you be doing in 2050?

I will be 77. Fingers crossed I will still be alive. If so, I will probably still be working, as by then the retirement age will likely have gone up to 80!

Have you thought about how clinical microbiology laboratories in general might look in 2050?

Here are my thoughts:

  • They will be highly automated. Much more like the biochemistry and haematology departments of today.
  • They will generally be big. Because of the above, laboratory “capacity” will often be far in excess of (local) demand.
  • They will be virtually paperless. Accreditation agencies will insist on electronic audit trails and fully interfaced hardware, and rightly so.
  • Electronic ordering will be the norm. …and with it the ability to reject tests due to insufficient details or because a similar test was done a couple of days ago.
  • For bacteria, culture will still be around, but in a hands off mode. I suspect genital and enteric bacteriology will have gone completely molecular. Testing of general wound swabs, respiratory samples, urines and sterile site samples will probably still be culture based in part, but culture inspection will almost certainly be of a digital image, as part of an automated processing system.
  • Antimicrobial Resistance Testing will remain controversial. Likely to remain a combination of phenotype, genotype, and increasingly proteotype.
  • Maldi-TOF will have evolved. ID of filamentous fungi and mycobacteria will be routine on Malditof. Several different resistance determinants will be identifiable by mass spectrometry. The hardware will be more sophisticated than currently.
  • Sequencing will be common. More common than nowadays, but I suspect their main role will still be in roles outwith routine diagnostic microbiology, eg genetic tumour markers, HLA typing, outbreak typing etc.
  • Microscopes will be very different. Will be completely digital, with viewing and manipulation done on a large screen. Looking “down” a microscope will be a thing of the past.
  • Training will be very different. Training will be more focused on troubleshooting automated systems and quality control processes as opposed to learning about the individual micro-organisms.
  • There will be a lot less microbiologists about. Microbiology scientists and technical staff, as well as Clinical Microbiologists will all be reduced in number due to both the automation and also the cost.

I look forward to checking back on this post in 33 years time to see how accurate these predictions turned out to be…..!