Automation, IoT and the future of smarter research environments
How smarter labs are benefiting from new technologyAdd bookmark
Technology and automation can be used to drive quality and efficiency in manufacturing processes, improving drug development through data analysis, process monitoring, and continuous feedback.
Big data, paperless processes, automation and cloud technologies are driving the laboratory informatics industry forward, creating smarter research environments which can optimise and ensure the quality and integrity of data.
Through the use of automation, and smart technology, R&D experts can alleviate the burden of many of the time and labour intensive tasks such as manual data entry and retrieval. Their time can then be used to focus on activities that leverage their expertise.
Ahead of their participation in the SmartLab Forum 2018, Pharma IQ spoke with Benjamin Schulz, Project Lead , Fraunhofer Institute for Manufacturing Engineering and Automation IPA and Jonas Angstenberger, Head of Process Automation & Data Science, AbbVie in order to hear their thoughts on how automation and connected devices can assist with ensuring data integrity and shape the future of the R&D environment.
Benefits of automating pharmaceutical research environments
Smarter research environments assist with the repeatability of studies by reducing the human factor and human error in the process and tests. they can also dramatically increase the quality of products through the standardisation of systems.
Schulz explains that “the accuracy of automated dispensers is improving, and this reduces the cost of kits and material in the laboratory”. He also commented that the volumes needed for research can be reduced with the increased use of automation.
However, Angstenberger notes that “automated environments demand a different set of skills”, something which is critical to consider during implementation and has cost implications in terms retraining or hiring new people.
He notes that “there are several papers that state organisations will need a year or even longer to train people in such a scenario”.
New tools and processes are enabling smart, decentralised production, with integrated IT systems and IoT connected R&D environments become increasingly flexible and highly integrated. This new wave of technological advances will drive forward the next phase of pharmaceutical manufacturing, enabling greater visibility of operations and allowing for agility in processes, bringing connectivity of equipment, people, services, and supply chains.
With the growth of automation, there is more functionality to integrate these systems.
“I think IoT devices are very rare in the lab at the moment” Angstenberger explains.
“We are looking at small automation islands such as liquid handling systems”. The implementation of these systems is also heavily reliant on a strong IT infrastructure, with increased numbers of businesses moving towards automation, the IT environment is changing. It is also critical that senior management are aware of the specific IT environment that is needed to handle automated systems, including the use of private cloud-base software.
Cost vs Benefits
The definition of IoT can be fluid and therefore establishing the cost benefits of acquiring such devices is highly dependent on what an organisations considers to be an IoT enabled device. Irrespective of this, there are a number of benefits of introducing these devices into the research environment and once an organisation has invested in automation, it is generally much safer and more agile in the research process.
Angstenberger explains that a key benefit of connected devices is that “there are cost implications if there are a higher number of errors when doing things manually, and lower costs if you automate.”
“The whole environment will need to adapt and choose different analytical methods based on results, and this is a huge challenge,” explains Angstenberger. Although this will incur initial investment costs, the alternative is to carry out the processes manually which will likely mean lower throughput and a lower quality with exponential long-term costs.
Using automation to ensure data integrity
In the past, manual data handling meant copying the raw data files together into one merged file, and “inevitably this manual process incurred a number of mistakes,” says Angstenberger, adding that in the beginning, “these mistakes are not always recognised and it is nearly impossible check all these processes manually”.
Documentation processes which would usually take two weeks when carried out manually can now be completed in ten to fifteen minutes, demonstrating the obvious time benefits of automation.
Schulz explains that “data is automatically collected and directly imported into a server where you can watch and analyse it, and therefore reduce the chance of errors compared to using paper to report data”.
However, he also notes that that “one remaining problem with automated data handling is that it's missing standardisation in the data format”, meaning that while it is far more efficient to automatically collect the data and import it into the server, there is still a challenge with metadata as the format is not standardised.
Schulz echoes Angstenberger’s statements, adding that “if the data is just collected and nobody works with it, then the benefits of automation are limited”.
Security risks with IoT devices
Security can be an issue with connected IoT devices, particularly when it comes to storing data on external platforms, “most often IoT is associated with importing the data into a worldwide specific cloud, and that’s not necessarily secure,” says Schulz.
The security of external storage is highly dependant on the cloud infrastructure. Angstenberger notes that “in general, I believe it’s not particularly well accepted to have data stored outside”.
An increasing a number of companies are utilising in-house cloud storage software which can reduce security risks, although it is important to note that even with in-house storage, every desktop must be secure in order to ensure the security of the entire network.
Integrating Automated Systems
“In a perfect world everything would be standardised and the risk of siloed systems would be minimal,” says Schulz.
It is important to establish integration standards in the laboratory, not only because it is something which is highly valued by the customer, but also so that standardisation is ensured between devices and there is standardisation throughout the data format. However, to ensure that everything is working within an organisation’s automation line, a separate network is required, and therefore in this sense companies are isolated and without outside contact, making integration more challenging.
Angstenberger explains that communication paths blocked with firewall systems can prevent fully integrated automation, “you run into issues with very old-fashioned drivers for outdated environments that have strange communication paths which aren’t particularly functional”. He explained that at AbbVie they understood the need to be separate but on the other hand there needed to be a connection to the outside world through a firewall. After intensive discussions about the IT infrastructure they are now looking into a special infrastructure that supports automated devices.
Future digital transformation
Looking forwards, Schulz believes that “digital transformation will assist people in the lab during the manual steps, tracking the manual steps executed by a person, and automatically generating the protocols”.
It is likely that automated assisting systems will become more commonplace, as R&D environments require increased levels of flexibility which automation alone cannot fulfil.
“Moving forwards having a strong network is a basic necessity, if you don’t have a strong network you will not be able to organise your automation workflow” says Angstenberger.
All information on IoT devices is communicated through the network, so if the network fails or one element breaks down, this can have repercussions for the rest of the laboratory. Schulz expressed optimism about future reliance on network strength, commenting “hopefully fail safe systems are something which will be coming soon”.
In more technologically advanced industries such as the automotive industry, automated systems implement an automated decision based on predictive analytics algorithms.
In the pharmaceutical industry this is still a long way off, but would be a “huge step forwards because you could automate operational decisions in an R&D lab” says Angstenberger.
This is something that could be introduced in a standardised fashion in an automated scenario but it is not likely to be introduced into labs within the next few years.
Automation and IoT systems can potentially transform processes within pharmaceutical manufacturing facilities, helping to realise major performance improvement. Companies that take the initiative early stand to gain the biggest competitive advantage, ensuring that they can operate with greater agility, cost-efficiency and compliance.
The SmartLab Forum 2018 will focus on the future of IOT and SmartLab Technologies as well as overcoming LIMS interoperability challenges and data transformation.