Many aspects of modern life are becoming tied to the digital world. You are using your smartphone to read emails, order taxi, check for directions, and order food. But there is a place in this world that remained more or less immune to digitalization – a research laboratory.
Why is the place of the greatest innovations so resistant to digital technology? The answer lies in money – laboratories are bottomless pits. Most of the money is used to fund the direct outcome of the research. Much less money is used to update scientific equipment and to optimize the existing processes. And this is where digitalization can help. Another reason is that any optimization requires an investment of time and effort, which is taken from the research work. So when looking at short-term goals, such investments never seem to pay off.
At BioSistemika, we’ve set up several smart laboratories. A large part of the process are discussions with the stakeholders at all levels. Here, we assembled a list of 6 most common myths we hear from people when we plan the digitalization strategy.
Short on time? Read the content of this article as a slideshow.
Download a slideshow that summarizes the content of this blog article.
1. A smart laboratory is complex
A smart laboratory is an umbrella term for different software and hardware tools to simplify or automate the data management and processes in the lab. The laboratory processes can be digitalized at different levels and with different approaches. That’s why the lab digitalization is very complex.
Digital data management
Digital data management is the first level of digitalization due to lower cost and simpler implementation. The idea is to use software tools to document, store, analyze, share, and manage the experimental data in a digital form. The software solutions used here are Electronic Laboratory Notebook (ELN), Laboratory Information Management System (LIMS), Laboratory Information System (LIS), Laboratory Execution System (LES), Laboratory Data Management System (LDMS), Scientific Data Management Systems (SDMS), and Document Management Systems (DMS).
Integration of data
At this level, the laboratory instruments already report the data and experimental conditions automatically to the central data repository, such as ELN or LIS. The goal is to connect the laboratory devices to the network. You can do this by using APIs or middleware solutions that handle the communication between devices and the central data repository. This level of digitalization has three direct effects:
- Prevents errors during the manual data transfer
- Improves the consistency of the data
- Reduces the time required for manual data transfer and verifications of data transfers
Very few scientists think about the laboratory environment data when performing experiments. ELNs can store laboratory environment data, such as temperature, humidity, air pressure, and light as well as environmental conditions in equipment where this is crucial, such as incubators and refrigerators. These data bring immediate value when optimizing the assays and searching for reasons behind outliers. When the assay gets into routine use, environmental data can be used to troubleshoot the process.
Automation is the next level of digitalization and brings the most value to the lab. In an automated lab, liquid handling systems and robotic arms perform the assays and transfer the containers between different devices. The aim is to remove human interaction in the physical laboratory process. This eliminates human factors and enables a very high level of error control. Some laboratories are automated to the level where no human interaction is needed. Robotic arms take samples from the storage, transfer them to the analytical device, and the scientist can check the data in the ELN. With this, we are entering the IoT world, where laboratory things are seamlessly connected. The data flow from devices to the clouds, where they are automatically processed.
Automation does not need to be complete right from the start. Smart laboratory tools, such as smart pipettes and different aids to perform manual tasks easier can already make a large difference.
Currently, the most frequent use of automation in laboratories is partial automation, where only the most routine processes that do not have a lot of variability between runs are fully automated, as these bring the best return on investment.
Automation does not need to be complete right from the start. Smart laboratory tools, such as smart pipettes and different aids to perform manual tasks easier can already make a large difference. These tools can reduce human errors to some level and produce digital data output that can be integrated into ELNs or LIS systems. Compared to manual processes such automation is still superior in pretty much all aspects.
The levels beyond automation are very open. An example is full business process automation, that includes also non-research processes such as handling of incoming samples and materials, invoicing, shipping, and reporting. This calls for integration with Enterprise Resource Planning systems (ERPs).
Another aspect is automating quality control, decision making, reporting, advanced analytics, and predicting trends, or creating scientific hypotheses. This is an area that calls for the involvement of artificial intelligence (AI). The goal is to reduce the time scientists spend at the bench and instead invest it into areas where they can add greater value.
2. Challenging use for unstructured processes in the research lab
The implementation of smart technologies into the laboratory requires that the processes and the data have some structure. This is a standard for many QA, diagnostic and industrial labs, but traditionally impossible and sometimes even intentionally avoided in research laboratories.
The strategy applied in such a research lab is to digitalize the processes in several steps. For example, first install an ELN, which is designed for research labs’ needs. In reality, an ELN in a research lab can increase the quality of the research work. Research labs typically have a high fluctuation of the members, which need to be trained for the processes in the lab. When many scientists work on a single project they need to sync their techniques and protocols and to avoid duplicated work. The ELN solves this by providing a single repository for experimental protocols, sample archives, patient data, and even analytical procedures. It also allows an overview of the experimental work and tracking of progress.
After implementing an ELN, research lab managers can start automating processes. Each research lab has certain assays that they specialize in. For example, if you do high content screening, you can use a liquid handler to prepare the plates with reagents, seed the cells, fix and stain them, and then use a robotic arm to transfer the plate to the analyzer. If you do this often it is best to automate it. There are also assays that are difficult to automate. In such cases, you can instead use smart pipettes and other aids that help you reduce the errors. Regardless of the approach, it is important to have a clear aim, strategy, and plan for digitalization.
3. It is difficult to show the value
Building a smart lab can be a significant investment for any organization. Many organizations still consider investments based on their return. The organizations aim to have the shortest possible Return of Investment (ROI). But this does not work for the investments in the smart laboratory.
The ROI for smart laboratory technologies cannot be measured in the short-term. The real value of a smart laboratory can be only observed after the complete transition and it needs to be measured at many levels.
Therefore both laboratory and organization management needs to support the digitalization. Or in other words, you need somehow visionary approach to it.
Examples of added value
A decrease in electricity costs is the most immediate ROI. A study by Harvard reported that smart scheduling of the equipment can reduce electricity consumption by 51.6% or even up to 70% for the largest consumers, such as thermocyclers. A thermocycler can use a few kilowatts of power – just by powering down the device after use, we can save a significant amount of money. The next level would be to change the assays to use less time on the equipment, if possible.
Smart scheduling of laboratory devices can save you up to 70% on electricity bills.
A less immediate ROI is time savings. Scientists spend much less time on data input, searching, analyzing, sharing, and presenting. Some assays produce a large amount of data and often the scientists manually type the data to the ELN. Manual data transfer is also a large source of errors. We can avoid these errors by integrating the equipment with the ELN. There is no need to search through a pile of notebooks anymore because the data is searchable from the platform. The largest time saving is when sharing and presenting the data. Scientists can share the data with a few clicks. There is no need to make hard copies or scans of the notebooks and ship them to the recipient. Having all the data at your fingertips also makes it easy to insert them into your presentations.
According to the study done by SciNote, researchers save 9 hours per week by using an ELN, while doing the same amount of work. Scientists can make good use of this extra time. They can ramp up the R&D pipelines and reduce the lab-to-market time – this is an important advantage in a fast-changing market. They can also use the extra time to optimize the existing processes which can also reduce the costs and time.
4. Legal implications and intellectual property management
The pharmaceutical industry has a long history of using software solutions for various reasons. The first adopters were the quality assurance departments, where human error could have significant financial consequences. The digitalization of these processes reduced human errors and therefore became the golden standard in the pharmaceutical industry. The quick adoption by the industry was followed by regulators developing legal frameworks to ensure the software was of the highest quality. The rest of the laboratories followed this digitalization spree, with one single reservation – the patents.
The research-heavy processes remain reserved towards the smart laboratory solutions because they often result in patent filings. The reason for this is patent disputes where the inventors need to prove that they were the first to invent. The easiest way to present the evidence to the court is in the form of bound paper notebooks. Although electronic data records are equally relevant in the court, in practice the bound paper notebooks are still preferred. A good practice is to consult with a legal advisor when implementing an ELN. At least, you should choose an ELN which is safe, backed up, multi-user, and at least provides the audit trail capabilities.
5. Poor user adoption
The implementation of the new technologies into the business processes is accompanied by the resilience of the employees. This is somehow understandable as most people don’t want to change their habits. There are several reasons for that, for example being afraid they might lose their jobs or feel underskilled to use the new technology. All technological disruptions throughout history were accompanied by such resilience, most recently with the digitalization of businesses. But in fact, throughout history, the technology created more jobs than it has destroyed. We’ll see whether this will be the case in the upcoming automation-themed industrial revolution.
Another users’ concern is also the big brother effect. When all the data about processes, including quality controls and results are centrally processed, the superiors have more detailed insight into how the work is done. The data is more transparent, which is good for business but might cause resistance from users. This is particularly evident in research environments, where researchers can omit data from the lab notebook and where superiors have insight into work only at regular meetings.
Overcoming the skepticism
It all starts in the planning phase. The approach is to reach out to the users and hear them. You can organize user working groups and include them in the core planning phase. The members of the user groups often become early adopters and other users respect their opinion. The user groups provide very important input for the design implications. They often become the evangelists of the new technology in the organization.
The bottom line of these approaches is to use positive reinforcement, which works better in the long term.
Before implementing the smart technologies, you should try to empower the users by clearly communicating the reasoning for a change. The goal is to show the users the benefits they’ll get with the new technology or, according to the Technology Acceptance Model, show the usefulness. Sometimes you could use the give-take approach. This means that a user gives up something to get something. A relevant example would be providing the extra IT training for faster adoption of the technology. It also works better if the instructions come from the source of authority. Additionally, you should engage the first adopters and the evangelists within the organization to act as good examples for others. The bottom line of these approaches is to use positive reinforcement, which works better in the long term.
6. Data integrity
When discussing digital record keeping, automation, and management we cannot avoid discussing the security and data integrity. Although we often focus on data security, smart labs will need to expand this to general digital security. When all devices are connected to the internet and can be controlled remotely it is even more important to prevent unauthorized access. An attacker could, for example, remotely destroy the machine which would result in high damage costs.
Let’s focus on the data first. The two measures to ensure data integrity are access control and secure audit trail. Access control limits access to the data to authorized people and prevent unauthorized people from accessing it. Besides, it also enables different roles for different people. For example, most users can only generate new data, some users can review data and only a few can edit the data stored in ELN. This combined with encryption ensures the technical safety of data. The users still need to be aware of social hacking, i.e. tricking the authorized users into providing access to unauthorized users.
Another requirement to maintain data integrity is an audit trail. In practice, an audit trail is a log of all changes in the data, either as a consequence of user actions or automated workflows. The audit trail mechanism stores the old and the new version of the data. This allows full traceability of the changes from the original version of the data.
All these systems ensure a certain level of security and integrity. Regardless, there is still a chance that someone will manipulate the data, be it paper or electronic data. In the future, this could be addressed with the blockchain technology, which is an immutable public record database. The system can automatically convert a change in the data to a hash, which is a unique signature of the data that cannot be converted back to the original data. The system then stores the hash signature to a public or private blockchain without exposing any data to the public. To check if the data integrity was compromised we can simply convert the audited data into a hash signature and compare it to the hash signature from the blockchain. This technology can be used to track data integrity for any kind of electronic data.
Many laboratories raise concerns about how external vendors process the data from connected devices. By using all these digital tools the users will also have to give away some control over their data. This is also one of the reasons why they avoid the digitalization of the labs. The vendors should, therefore, be very clear about how they process the data and consider giving the users options to store data locally.
Above, we discussed the obstacles when implementing smart technologies in the laboratory. As we can see, the digitalization, in fact, increases the value of the laboratory. While it takes a while to change the general perception, there are initiatives to reduce the digitalization gap. In fact, there is even an idea to bring the laboratory infrastructure into the cloud. The scientists could control cloud laboratories over the internet, just like cloud computers. This would revolutionize the biotech businesses in a similar way to what cloud computing did for the digitalization. While cloud labs are still in their early days, you can start digitalizing your lab today. Either alone or by consulting with smart lab experts, you would bring immediate value to your lab processes. But overall, we’re still a long way to reach the level of the digitalization present outside of the laboratories. A long way, paved with many opportunities.
Author: Tilen Kranjc