As part of a series of articles based on a recent online roundtable event entitled ‘From research to manufacturing: Overcoming data challenges in the drug development cycle’, hosted by Scientific Computing World, we asked our panellists about the challenge of cultural change during the digitalisation journey.
You can register for the White Paper here.
The digitisation of data – turning paper records into digital ones – and then embedding digitalisation – the consistent use of digital data in an institution – can be a long and painful process, particularly if team members have been used to doing things in a particular way for a long time.
“Digital records mitigate the chance of human error, particularly around missing data in paper records,” says Jackie Lighten, Program Manager, Cell and Gene Therapy Catapult. “With digital records, you can put in safeguards that reduce the number of missing fields, as well as completely eliminate any risk from illegible writing.”
While the benefits of digitised data are hard to argue against, cementing the rules of engagement with that data is a much harder step.
Lukas Kürten, Digital Innovation Manager at CPI, says, “It’s an old adage, but I think the journey to digitalisation is as much an organisational cultural step as it is a technical one. A digital system can be as useless as a paper system if scientists don’t use it properly. We’ve had challenges in past projects where we had digital data capture processes in place, but the scientists didn’t understand, for example, that if you put a space in a spreadsheet cell somewhere, it’s going to throw a computer off.
“Typically, the people in the lab are biologists who are very smart people, but they’re not necessarily the most technically computer-savvy or digital people. There’s a cultural education process to go through to help the scientists understand why it’s important to enter data the same way every single time. They might feel that the process is restrictive, but they also need to understand the benefits of doing so.”
The cultural change required is also recognised by Moritz von Stosch, Chief Innovation Officer at DataHow, who recommends incentivising accurate digital records.
“The system should really be such that when a person has to enter data manually, they get something back,” he says. “If that person is not directly getting something back, there is no incentive to record that data. I’ve worked in places where we have introduced a new data gathering platform and discussed why we all needed to adhere to standards of data entry. People did it for a while, but then, after a period of time, standards began to slip. It’s not that they were purposefully forgetting or doing it wrong; they just had other things that were higher in their priorities.
“So, one needs incentives to record the data point, such as access to a graph that they can put directly into a technical report. We need to build ‘win-win’ mechanisms into our data systems to improve the quality of data recorded.”
Kevin Back, Product Manager at Cambridge Crystallographic Data Centre (CCDC), echoes the incentive sentiment. “Data dashboards are really helpful,” he says, “particularly if you can get data visible to managers. If they can look at the data that’s being produced, see its quality, and see the volume of it, then can use that to discuss their reports in meetings or reports back to the wider organisation on progress. That’s the user getting something back and encouraging them to really interact with that system.”
DataHow’s von Stosch suggests building KPIs around data. “If the quality of data is particularly high due to the efforts of a person,” he says, “they should be rewarded for that. Or you could take a completely different approach, where you put it much more into gamification. Duolingo’s streak function* is very addictive, for example, so adopting that to recording quality data in a lab would reward consistency.”
The cost of digitising data, particularly for early-stage companies that don’t have the internal resources or the budgets for the larger software packages, can also be a barrier to transition.
“In the cell and gene sector, there isn’t a reluctance for digitisation,” says Catapult’s Lighten. “Quite the opposite: everyone knows they need to do it. It’s more about picking the right moment in a company’s journey to make the transition, as it’s not a trivial task. In the first stage of a cell and gene therapy company, the challenge is getting the biology right so that you can perform experiments to test your hypotheses on therapy efficacy. That’s where the seed money goes, rather than on digitisation. I assume that VC boards would rarely ask for the latter. This is in spite of the fact that effective digital systems can ensure that the data is collected and utilised effectively, thus improving the power to demonstrate therapy efficacy and therefore derisk their financial investment.”
Indeed, the transition to digital has major benefits right up to the manufacturing stage. “A recent partner reported up to 70 per cent efficiency savings by going to digital BMRs (batch manufacturing records),” continues Lighten.
AI (Artificial Intelligence) is being used increasingly at most stages of the drug discovery journey but, as Darren Green, Director, DesignPlus Cheminformatics Consultancy, points out, it’s only as good as the underlying data. “Everyone’s fascinated by AI and machine learning tools,” he says, “but you need investment and prioritisation in data processing and posting systems so you get reusable data. The balance of investment tends to be skewed to the fancy stuff (AI), meaning the latter gets deprioritised.”
So, what can organisations do to ensure the widespread adoption of digital habits? Catapult’s Lighten says building trust is a good place to start. “It’s easy to take comfort in paper printouts when assessing data,” he says. “The moment you implement an interoperable digital system, the user experience can make many data points appear invisible, and that can be unsettling for those used to being able to see them directly on a paper report. That makes it doubly important that the backend system design is thorough and functioning properly, as you don’t always have eyes on every data point, and so you must have a trusted data traceability pipeline supporting the exchange between systems.”
Initiatives to standardise data in digitalisation journeys is nothing new, as von Stosch explains: “The QbD (Quality by Design) initiative that was launched by FDA in 2004, and has been widely adopted by the pharmaceutical industry since, is all about knowledge capture and the reuse of that knowledge. It is a big investment in experimentation, as it involves a deep understanding of the system in order to make riskbased judgments on the impact of a factor on critical quality attributes. Reusing knowledge, however, reduces the amount of experimentation required without limiting the capability of the model to judge how the factor will impact on the critical quality attributes.
“Similarly, digitalisation should help us to capture the knowledge in a systematic way and also allow the use of that knowledge in a much more systematic way. Using FAIR (Findable, Accessible, Interoperable, Reusable) principles for data is a basis for this, but there must also be principles applied to other knowledge capture, such as text, so that it can be reused. It’s not easy in practice – people are always busy, so they aren’t always tagging data correctly or capturing stuff systematically.”
Patrick Bossuyt, Quality by Design and Digitisation Manager, Pharmaceuticals and Life Science, Siemens, adds: ‘Development processes in the R&D space require support from a multidisciplinary team with expertise in various domains, such as biology, data science, processes and quality. Optimal performance is achieved when these experts work in full harmony with technology, accelerating new product and process development. The industry has an obligation to embrace technology to deliver (new) medicines to patients faster.’
The full report is available to download as a White Paper, which also covers: Data collection and formats; Data silos and how to avoid them; Data ontologies and efficiency of process development; The shift to in silico for experiments; and Process optimisation and technology transfer.
You can register for the White Paper here.
The roundtable and series of articles is sponsored by Siemens Digital Industries.
*The Duolingo app encourages habitual use by rewarding ‘streaks’ - sequential days of using the app
