Contributing editor Tanuja Koppal, Ph.D., talks to two scientists—one from a large pharmaceutical company and the other from a large academic institution—about their experiences with transitioning from paper to electronic lab notebooks (ELNs). Cecilia Björkdahl, Ph.D., manages the across-the-board research documentation project commissioned by the Board of Research at Karolinska Institutet Medical University. She discusses the challenges of adopting ELNs in a diverse academic setting and offers some recommendations and improvements that can make the transition easier. Stan Piper, formerly principal scientist specializing in informatics projects in Pfizer’s Pharmaceutical Sciences division, outlines the process for evaluating and implementing the right ELN that meets the needs of both the organization and the individual user. He advises managers not only to consider the present use for ELNs but also to keep an eye out for future applications and growth.
Interview with Cecilia Björkdahl:
Q:Can you share your experiences transitioning from a paper to an electronic lab notebook?
A:For us the ELN has been very easy to use. You just need an hour’s introduction to the system, and you’re up and running. It’s quite simple in a good way, without any kind of sophisticated configuration necessary. The ELN we use is completely off the shelf, unlike our laboratory information management system (LIMS), where we did a lot more customization. I would say it’s worked probably better than the LIMS, to date at least, when it comes to getting new users to use the system quickly. Customization is good, but it’s also extremely time-consuming. And in the end, you don’t always get what you thought you wanted anyway.
Q:Do you find that there are some limitations or disadvantages to using ELNs?
A:I think some of the functionality may be too simple. It’s a balance in terms of making it easy to use and not too complicated to maintain, while supporting a lot of what our researchers do. Some of our users have their protocols in Microsoft Word or Excel, and when they wanted to transfer that information into the ELN system, we ran into some difficulties. You have more options to change the layout and format with Word than with the ELN. But I think the main limitations of ELN are related to size. We have a lot of researchers who handle really large data files, in terabytes and petabytes, because they have a lot of genotype data or large image files. We currently have a size limitation of 15 megabytes per attachment or insertion, although there’s no limitation per person or group. Currently the data is stored locally on servers, and the ELN can connect to that. But I think that’s something that we need to address, how we handle the data, ensuring that we have a good reference system for where that data is kept. We have around 1,000 users at the university, and we have the potential to include probably 4,000 to 5,000 users. In that sense, we still have a long way to go.
Q:Have you run into any kind of problems with integration or security?
A:With regard to security, it has to do with accessibility and making sure that the communication with the ELN system and also to the server is encrypted. We have a very diverse environment, and here academia will differ from a company, where it’s a bit more regulated as to the kind of computers and programs people are using. We have 22 different departments and around 2,000 Ph.D. students and 3,000 researchers, and then we have undergraduates too. People have different types of computers, and they have different versions of the systems on their computers. We have had users contacting us with some problems and issues that have come up in their local environments that we didn’t anticipate before we started, and we think a lot of that has to do with the diversity we have.
Q:As the project manager, what are some of your biggest concerns with using ELNs?
A:I think it’s associated with getting a stable system that we can trust and getting it up and running. Once you have that in place, you can always work from that. Also with ELN implementation, don’t make it into an information technology issue but make it a research issue. Here we have managed to include researchers early in the planning, with me functioning as a go-between for the researchers and the IT staff. So an ELN is not just something that IT decides on, gives you, and forces you to use. The researchers should be the ones driving the process.
Interview with Stan Piper:
Q:What factors did you consider when you started your search for the right ELN?
A:There were two things—the drivers that led us to look for an ELN and the key criteria for selection that we were looking for. In terms of drivers, we wanted to move from the paper-based system and increase efficiencies in the laboratory as well as increase quality and compliance. For selection approaches, you can either take an application and try to customize it to what your users expect, or you can pick a partner that can deliver an application that’s very close to your requirements and then do small tweaks and configurations for your users’ needs. We were really looking for the latter, and the key in that case was really asking our users what they were looking for and getting early user engagement.
Q:What kind of feedback did you get from the users?
A:The users were really looking for a few different things. One was usability, obviously. If we put an electronic system in, it had to be usable and couldn’t be overly complex. It needed to be just as portable and just as accessible as a paper notebook. It needed to be searchable. And finally, the users wanted an electronic system to be more collaborative than the current paper system.
When we went through the choice of vendors, there were four high-level categories considered— operation, quality, interface and integration, and other. The category called operation included experiment setup, ordering, the ability to handle chemical reactions and structures, the ability to apply workflows to things in the system, and being able to get metrics from the system. The quality category included things such as signature and reviews; approvals; and some way to address archiving, reviewing, witnessing, and reporting. The third area that we were looking to evaluate was what we called interfaces and integration. The user interface, the look and feel of the product—do they seem right? Is it something that’s very foreign to your environment within your company? What did the base configuration look like? Is it easy for users to make changes that they need? And then on the integration side, what are the available instrument connections? If we want to do direct instrument connection, is it available to us? And then, the last large category called “other” included the relationship with that vendor as a company and their size of scale and support, etc. What kind of relationship would you have with this company moving forward? Are they scaled to the size that could support your organization? Do they have appropriate training, organization, and staff to support your enterprise? Could they help with implementation as well as with maintenance and upgrades?
Q:Do you think that lab managers,although they are looking at the present when they buy an ELN, should really be thinking more futuristically when they actually go about implementing it?
A:That’s correct. It is important to consider “the ultimate total cost of ownership” because you’re really looking at ELN as a long-term tool. You have to look very far into the future to see whether ELN is going to be a key piece of your knowledge management puzzle. Could this system grow to a point where you really need to think about the total cost of ownership over a decade or two before you’d be moving on to some change in technology? Think downstream in your process and in time to see how other systems or users are going to want to consume information in the ELN. You may want to rethink how you implement your ELN, because you may want the data in the ELN to be structured in a way so that you can access it long term. So if you see ELNs as a key piece of your knowledge management strategy, then that should be part of your evaluation.
Q:If you were to recommend some improvements in the ELNs, what would those be?
A:We need to look at things like extraction tools and business intelligence tools that can extract real scientific information from the ELN and apply some sort of context to it so that we can begin to build predictive control to improve our processes. I have a colleague who says that if we put in an ELN but don’t build new efficiencies, the only thing we’ve done is put paper on glass. It should be adding benefits, particularly around knowledge management, to improve our decision making and our processes. We continually need to improve ELNs, driving toward that goal. This depends partly on how the vendors build the product, how they structure the data moving forward, but it also depends on the customers. It’s the customer’s long-term vision of how they want to extract data, what kind of context they want to apply to it, how they want to feed it back into their systems and send that information back to the vendors; that plays just as important a role.