As an information systems consultancy dedicated to successfully delivering lab-based information systems, we help our clients to overcome many different challenges. There are some important questions that we are frequently asked to evaluate.
In part one of this blog series, we’ll summarise the considerations to make when answering 3 common questions about lab informatics systems, all in the theme of ‘is a single system better than multiple similar systems?’
1. Should R&D labs use the same informatics systems as QC?
Here the context matters. If one were to generalise, R&D labs tend to be experiment-based, answering questions like ‘What ingredient changes in the product formulation will increase effectiveness and reduce environmental impact?’. On the other hand, QC labs are more focused on samples taken from production runs, and questions such as ‘Are the % composition of key ingredients within a production batch within specification?’
If we use the above generalisation and apply lab informatics thinking, in broad terms, ELNs are centred on recording experiments and therefore more suited to R&D. LIMS, being sample, test and results biased, are generally more suitable to QC labs.
However, it is not that simple. For example, perhaps one of the R&D labs provides analytical services to various teams executing R&D experiments – this type of ‘service’ lab is often better served by LIMS than ELNs.
The type of labs involved is not the only factor to consider. For example, CDS systems are generally applicable to both R&D and QC. The methods and use of the instruments may well vary across R&D and QC, but the instrument data systems can be exactly the same.
Finally, regulatory needs, specifically for QC can also be a driving factor in answering the question. We will consider this further in one of the following questions.
2. Should we implement a single global system or several more local systems?
When Scimcon first started nearly three decades ago, the focus within large multi-national companies was on implementing large, monolithic lab systems. This approach still has its place, particularly where the distributed labs are very close in terms of operational and analytical workflows.
Current thinking, however, looks to best support the diversity of lab workflows across global sites. While this should not mean a different system in every single lab, it should ensure some flexibility in selecting systems locally. This has several benefits, including a better informatics fit for each lab, and the increased local user buy-in gained by allowing flexibility.
However, against the background of the drive to increased data aggregation, data science and analytics, and AI/ML, this local approach can be counterproductive. It is therefore important to set standards and guardrails about how these systems are implemented, and how the data is structured and linked via reference data, so that consolidation into centralised reporting tools and data lakes is facilitated.
3. Should I have different systems for GxP and non-GxP work?
There is a well-used saying within regulatory-compliant organisations: ‘If a system contains just 1% of GxP data, then the whole system is required to be implemented, managed and maintained in a regulatory compliant manner.’
This statement leaves compliant organisations questioning:
Is it easier to run one regulatory compliant system, that contains both non-GxP and GXP data, and accepting that the non-GxP will also be subject to the associated GXP administrative overheads?
Or is it easier to have two systems, one GxP and the other non-GxP, the latter of which is subject to less rigid controls?
The first step to answering the question is to determine the delta between administering a GxP system, and administering a non GxP system. LIMS, ELN, SDMS, CDS and other lab informatics systems are often classified by labs as mission-critical. Most organisations wouldn’t countenance a lack of system administration rigour or releasing untested changes to mission-critical systems, so this delta may be lower than it first seems.
The next step is an open conversation with QA teams about the types of data being held, and the control systems that will be put in place. In the past, we have successfully taken a two-tier approach, where the administration procedures for non-GxP are simpler than those for GxP data in the same system. However, for this type of arrangement to be viable, a detailed risk assessment is required, and the ongoing management and control of the administration has to be very well executed.
Finally, before making the decision, it’s worth considering whether there are shared services or functions involved. For example, if the GxP and non-GxP work uses the same inventory management, it might be complex to get the inventory system interfacing and updating two systems simultaneously.
Summary
Hopefully, we have illustrated the importance of being clear about what your requirements are before answering these key questions about lab informatics systems. Each case is unique, and your decision will usually be based on a wide range of influencing factors. We help organisations to consider all of the options and roll out their chosen model.
Stay tuned for part 2 of this blog series, where we will look at the key question of how you can prepare your data for AI and machine learning.
Our latest industry leader interview is with Pascale Charbonnel, who tells us about how SCTbio supports customers through the cell therapy manufacturing chain.
In this instalment of our industry leader series, we speak to Pascale Charbonnel, Chief Business Officer of SCTbio. Pascale tells us about the work of SCTbio, how they collaborate with biotech developers, and why they are a great choice for outsourcing cell and gene therapy (CGT) manufacture.
Tell us about SCTbio
SCTbio is a cell-based therapy and viral vector contract development and manufacturing organisation (CDMO). Originally part of the SOTIO group, we spun out in 2022 and operate a Good Manufacturing Practice (GMP) facility in Prague, Czech Republic. Recently, eureKING, a French special purpose acquisition company, or SPAC, has signed an agreement to purchase full ownership interest in SCTbio, which will further bolster our position as a leading CDMO service provider.
As part of SOTIO group, we were developing our own cell and gene therapies for 13 years, so we have a lot of experience in manufacturing for clinical trials from phase I to phase III across multiple geographies. Given this expertise, customers trust us to guide them through the development process as they navigate the GMP world and clinical development.
What kind of customers do you support, and how do you support them?
Our target customers are mainly early-stage biotechnology companies, who typically outsource all their production needs. We are sometimes also used as an additional facility to absorb around 20-30% of the production needs for large Ph II / Ph III phases. Our main goal is to establish trust with customers right from the beginning, so we can then support them as the project progresses through later clinical phases. The average customer project takes about two years.
With our history in SOTIO, we can ensure GMP compliance for the full drug development life cycle as we have also faced some of those same hurdles associated with developing therapeutics. Our team understands the importance of saving time and costs, and maintaining momentum to ensure approvals run smoothly and that we can move onto the next clinical stage. We use this experience to create optimised development plans, which give customers the assurance that we can support them and hopefully go on this journey with them for many years to come.
How do you manage the data you generate for customers, and what formats do you report in?
We are still very much in a mixed model – so we have turned to electronic systems in some cases, but we do still have paper-based approaches too. It’s useful to have both, as it means we can tailor our approach depending on customer requirements. We’ve built our own data management system, which has been developed specifically to fit our operation here – so while there is scope for us to move to a full digital system, it will take time and our customers’ current requirements do not warrant that.
When it comes to customer data, we typically start by storing the raw data in a validated platform which we can then manage regularly. We then export it to the customer in whatever format they wish. As each customer’s requirements differ greatly, there’s no need for us to move to full digital systems yet, but it’s definitely something we’re bearing in mind for the future.
What does a typical audit look like, and how do you ensure success?
Since last year, we’ve run four audits – three by customers, one by a regulatory body. They all follow a similar process, where we will receive a request or announcement about two weeks in advance that an auditor is going to visit, and they usually request specific documentation which of course we already have to hand. During the day they will look at everything in our facility, speak to some of our technical staff, and then make a report outlining any observations.
GMP culture is very deeply rooted in our company, to the point where our recent regulatory audit returned no observations at all! While this shows everything was as expected, our customers were particularly impressed. One of our customers came back to us following their audit to say that they can see we go above and beyond the standard for GMP, and that our team is clearly well organised and collaborative.
How does SCTbio stand out as a CDMO?
One thing I think really makes us special is our people. We are a team of about 80 people, many of whom have been with us since the inception of SOTIO, and the staff turnover rate is very low indeed. It gives our customers a great deal of assurance that as well as having far-reaching experience in developing drugs and a deeply rooted GMP culture, our people are committed to our customers and get to know them and their needs.
What set us apart is our 13 years expertise in the CGT field and our flexibility to accommodate different sizes/stage of projects. We plan to stay very flexible, so that we can continue to take a bespoke approach to supporting our customers.
In addition, we offer a really wide range of services. We can collect the starting material, process it in our facility, release it under quality assurance / qualified person (QP/QA) and GMP conditions, and we have a logistical advantage as we’re based in central Europe, so close to a number of key markets. Being able to offer a full start-to-finish process in one place is quite unusual, so it gives us a strong advantage.
The recipe for success as a CDMO in my eyes is to have mutual trust and transparent communication with partners and customers, so with highly skilled people and low turnover, as well as the cost benefits of our location, our customers rely on us for consistency, reliability, and quality.
What do you think the future holds for cell and gene therapy?
The market has faced many challenges over the last few years, but we’re now starting to see an upturn. Funding is becoming available again, and we believe that ‘the good science’ will prevail. We’re excited to see what projects will come our way and to keep supporting customers to develop life-changing medicines.
Scimcon is proud to showcase CDMOs like SCTbio, and we’re looking forward to seeing how the company will grow over the coming years. To contribute to our industry leader blog series, or to find out more about how Scimcon supports organisation with lab informatics and data management solutions, contact us today.
In February and March 2023, Scimcon is hosting panel discussions at both SmartLab Exchange Europe and SmartLab Exchange US. The events, taking place in Amsterdam, The Netherlands and San Diego, North America take place on an annual basis as a forum for scientists in the modern lab to interact, form new connections, and learn more about the evolving technology that is disrupting the lab. Attendees and speakers will debate themes including: Lab of the Future, Data, Digitalisation, Quality Management and Standardisation during the conferences.
As a sponsor and panel chair in 2023, Scimcon’s opening panel discussion ‘What Is The Future For Human Scientists as AI & ML Deliver the Promised Step Change in Laboratory Practice?’ explored the future of human input in the lab, and how artificial intelligence (AI) and machine learning (ML) could impact the structures and processes in place.
AI and ML are everywhere we look – in the news, on our phones and other smart devices, and are increasingly making their way into other areas of our daily lives. In transport, we’re seeing steps being made towards self-driving vehicles. But what will happen to those engaged with the transport sector when human input is no longer required?
The same questions can be asked about the lab. We have seen similar disruptions in the past, and many scientists will still remember the days of cutting out chromatograms to weigh them and calculate peak areas – a task which now is fully automated. Through the employment of similar automated technologies – from sample prep, to HTS, and sophisticated instrumentation – we have been able to give more time back to scientists, to allow them to spend longer on the science that matters.
Our panel at SmartLab Exchange Europe and US will dig deeper into AI and ML, and how it will impact the role played by human scientists in years to come.
Tackling the AI & ML questions for scientists
The panellists will debate the big questions facing scientists on the topics of AI and ML during the sessions, including:
First; where are we seeing AI/ML making inroads into our labs today?
What will be the major areas of scientific workflow likely to be impacted by AI/ML?
What tasks do we believe will remain the domain of humans?
If AI/ML reduces the number of experiments that we need to do to achieve our lab targets, how will this affect the lab structure?
We have all heard the phrase “Enable Scientists to spend more time doing science”
What will scientists do with the additional time this next phase of technology will present us with?
Will lab sizes contract or will we seek to deliver ever greater results with similar sized operations?
What key skills will our scientists need in this new paradigm and how will this effect training and recruitment needs?
Scientists are exposed to rapidly changing technologies and work practices more than most people. How do we feel they will react to the likely significant change triggered by AI/ML?
Predicting timelines of future technology is never easy; when do you see the major impact of AI/ML occurring in your labs?
Following the SmartLab Exchange, Scimcon will summarize topics of key interest to the audiences in a future blog.
To join the discussion and hear more how AI/ML will impact laboratories and scientific operations, contact our team for more information.
We’re kicking off 2023 with a new industry leader interview, and shining a spotlight on Jana Fischer, Co-Founder and CEO of Navignostics.
In this blog, we speak to Jana about Navignostics’ mission, and how the team plans to revolutionise personalised oncology treatments with the help of data and AI.
Tell us about Navignostics
Navignostics is a start-up personalised cancer diagnostics business based in Zurich, Switzerland. Our goal is simple – we want to revolutionise cancer treatment by identifying a highly personalized and thus optimal treatment for every patient, to ensure that each patient’s specific cancer is targeted and fought as needed. Our capabilities allow us to do this by analysing tumour material, through extracting spatial single-cell proteomics information. and using this data to analyse many proteins simultaneously in individual cells within the tissue.
What is spatial single-cell proteomics?
Single-cell proteomics comprises of measuring and identifying proteins within a single cell, whereas spatial proteomics focuses on the organisation and visualisation of these proteins within and across cells. Combining these two research tools allows the team at Navignostics to characterise tumours on a cellular level, by identifying the proteins present across cells in a tumour, and also how these proteins and cells are organised. This means that the team can provide a more accurate estimate for how certain tumours will respond to different medications and treatments.
Proteins are typically the target of cancer drugs and measuring them on a cellular level allows us to identify different types of tumour cells, as well as immune cells that are present and how the two interact. This data is highly relevant to inform clinicians of the best form of (immuno-) oncology and combinatorial treatment for individual patients. Also, this information is highly relevant to pharma companies in order to accelerate their oncology drug development, by providing insight on drug mode of action, and signatures to identify responders to novel drugs.
The kind of data that we are able to extract from different types of tumours are monumentally valuable, so the work doesn’t stop there. All of the data we harness from these tumours is stored centrally, and we plan on utilising this data by building it into a system we refer to as the Digital Tumour, that will continuously allow us to improve the recommendations we can make to our clinical and pharma partners. Our journey has been rapid, though it is built on years of research and preparation: we founded the business in 2022, as a spin-off from the Bodenmiller Lab at the University of Zurich.
The dream became a reality for us in November 2022, when we secured a seed investment of 7.5m CHF. This seed funding will allow us to pursue our initial goals of establishing the company, achieving certification for our first diagnostic product and developing our Digital Tumour. By extension, collaborating with pharma and biotech partners in oncology drug development. It has also given us the resource we need to move to our own premises. We are due to move off university campus in May 2023. This offers us great opportunity to push forward with the certification processes for our new lab, and it gives us to the chance to grow our team and expand our operation. We will be located in a start-up campus for life science organisations in the region of Zurich, so we’ll be surrounded by companies operating in a similar field and at a similar capacity.
Tell us more about the Digital Tumour – how does it work?
The Digital Tumour will be the accumulation of all the molecular data we have extracted from every tumour that we have analysed to date, and ongoing. Connected to that, we store information on the clinical parameters and patient response to treatment. Over time, our aim is to utilize this central data repository to identify new tumour signatures, and build a self-learning system that will provide fully automated treatment suggestions for new patients, based on how their molecular properties compare to previously analysed patients that have been successfully treated.
Sounds interesting – are there any challenges to working with a database of this size?
Our data storage is quite advanced, so volume isn’t really a challenge for us. Our main focus is standardising the input of data itself. The technology is based on years of research and the data analysis requires a great deal of experience and in-depth expertise. In order to extract the full value from this data, it must be completely standardised. Data integrity is therefore vital to our work, and allows us to get the maximum value from past analyses. Our past experience in the Bodenmiller Lab allowed us to develop standardised processes to ensure that all of our data is fully comparable, which means that we can learn more and more from our past data, and apply this to new cases that we analyse.
It is also important to report on our complex data in a comprehensive but easily interpretable manner to the clinician/tumour board who needs to organise a treatment plan. We’re currently working with our clinical collaborators to develop readily understandable and concise reporting outputs. Unlike genomics analysis, our reports focus on proteins in tissue, which is the same information that clinicians are used to working with. So, there is a common language there that offers us the unique opportunity to provide clinicians with data they can easily interpret and work with.
What does this kind of research and data mean for oncology, both in terms of pharmaceuticals, biologics, and healthcare?
It’s important to note that personalised treatment approaches and precision medicine are not new concepts in the diagnostics space. However, our technology and algorithms allow us to extract novel types of biomarkers which were previously inaccessible or unknown, so we’re helping to level up the playing field and give clinicians and drug developers’ comprehensive information to individualize therapies.
Comprehensive tumour data is truly at the heart of what we do, and one key benefit of our technology is that we’re able to analyse very small amounts of sample – such as fine needle biopsies – to provide therapy suggestions. We can also analyse bio banked tumour material, so if there is any old material that has been stored, we have the ability to analyse those samples retrospectively. Not only does this help us to fuel our Digital Tumour with more data, but it also allows us to examine new fields such as long-term survival rates of patients with these tumours. This is of huge value to fuel our product development pipeline because it allows us to identify different molecular properties between individuals that may not have been considered on a clinical level, but may have played a role in patient responses to treatments and survival outcomes in the long-term.
This kind of retrospective data also plays a key role in the evolution of healthcare and drug development, as having the technologies available to acquire this sort of data and mine it to our advantage will provide enormous benefits. These include improving individual treatment courses for patients, as well as expediting the development of novel cancer drugs so pharma companies can get more effective treatments to market sooner.
For example, one commonly cited statistic is that 90% of clinical drug development fails during phase I, II, III trials and drug approval. Often, this may arise from a lack of available information to identify the subset of patients most likely to benefit from a novel drug. Having access to Navignostics’ technology and algorithms and a database such as the Digital Tumour will offer the potential to pre-select the right patients to enroll in clinical trials, and more easily identify the patients that do respond to the novel treatment, which could substantially expedite the speed of drug development in the trial stage, and help bring more effective drugs to the market.
Even unsuccessful trials offer valuable opportunities: it is possible to repurpose and reanalyse material from previous failed trials. Such high rates of failure in clinical development means that there are a large number of companies that have invested $millions in developing drugs that have not come to fruition, so if companies want to re-mine their data, our team can reinterpret the existing work into identifying more successful strategies, so we can give those drugs another chance and offer a better chance of Return on Investment.
A failure no longer needs to be a failure. Navignostics and its offerings can bring value to our pharma and biotech partners, and will also bring direct benefit to patients and clinicians once we launch our diagnostics product. So, data from every facet of the oncology industry, from curing a patient to halting the development of a drug, can offer us valuable insight that both we and the Digital Tumour could learn from when developing treatments.
What does 2023 and beyond have in store for Navignostics?
The next three years will be critical for our work, and we have projected timelines and key milestones for our diagnostics developments that we will achieve until our next funding round. Along the way, we are actively speaking to biotech and pharmaceutical organisations to identify projects and build the foundation for long lasting collaborations. We are looking forward to a successful continuation of the Navignostics development in 2023!
Scimcon is proud to showcase start-up companies like Navignostics, and we’re looking forward to seeing how the company will grow over the coming years.
To contribute to our industry leader blog series, or to find out more about how Scimcon supports organisation with lab informatics and data management solutions, contact us today.
Maintaining data quality is critical in clinical trials. As a follow up to his first blog, we have worked with Industry Leader Mark Elsley to create this infographic, outlining Mark’s top tips for managing clinical trial data.
Mark Elsley is a Senior Clinical Research / Data Management Executive with 30 years’ experience working within the pharmaceutical sector worldwide for companies including IQVIA, Boehringer Ingelheim, Novo Nordisk and GSK Vaccines. His specialist area of expertise is in clinical data management, and he has published a book on this topic called A Guide to GCP for Clinical Data Management.
2020 has been a difficult year for most industries, not least for event and tradeshow providers. Luke Gibson, Founding Director of Open Pharma Research and Lab of the Future, shares his experience of running events in the laboratory industry, and what makes Lab of the Future such a unique event.
Luke, please tell us a bit more about yourself and Lab of the Future
My name is Luke Gibson, and I am one of the three founding directors of Open Pharma Research. I have 30 plus years of experience in developing and running events, primarily in the financial and trade and commodity sectors. My colleagues Kirianne Marshall and Zahid Tharia bring a similar level of experience to the company.
Kirianne has had many years of experience in managing the commercial side of large congresses, such as Partnering in Clinical Trials, and research and development congresses. Zahid has 30 years of events experience too, particularly in running life science portfolios, and launching congresses/events. Our paths have crossed many times throughout our years working in events, and we eventually hit a point where all 3 of us had the capacity to try something new – something that was worthwhile, fun, and different to the corporate worlds we had become accustomed to. So that was why we created Lab of the Future – with a view to running events in a different way.
Did you feel that there was a gap in the market for this type of event?
I’m not sure if I would describe it as a gap in the market, more an ambition to do things differently. There was a desire from all of us to build an event with a different approach to the one we would take when working for large organisations, because when you’re working on a large portfolio of global events that cover a variety of topics, you and your team are always looking ahead to the next event, and the focus on the longevity of a single event isn’t always there.
We wanted something that we can nurture and grow, something that we can work on year-round without getting distracted by the next thing on our list. It also allows us to stay within this space and build our community, without having to face pressures such as a year-on-year development strategy or diverse P&L. Our desire was to avoid these constraints, and create an event that we can continue to work on for a long time.
Are you building just the one event, or are you looking at hosting a series? Has your business plan changed since starting?
We want to be able to live and breathe Lab of the Future, but one of the interesting things about it is that it’s such a broad concept. On the one hand we deal with informatics, but on the other hand, we deal with equipment, technology, and all the connectivity between them – but even that’s just one part of it. We are not an informatics conference; we are not strictly an instrumentation conference; we also look at the innovation side of things.
I think the best way to describe how we see Lab of the Future is as a proxy for how you do science in the future. Everything pertains to more efficient processes; better results; or ways of creating breakthrough innovation, and these are all part of the picture of science in the future. And that is the lab of the future – where the lab is the proxy for the environment where you do the science that matters.
So what is the main focus for Lab of the Future?
When we started off, we found we received a lot of queries from industry contacts who wanted to get involved, but certain topics they wanted to discuss didn’t necessarily pertain to the physical laboratory itself. But if it was relevant to science, then it was relevant to us. Things like data clouds and outsourced services may not be directly linked to the lab, but they still relate to how you work. So, within that, the scope for the Lab of the Future gets wider still, looking at areas such as how we can create virtual clinical trials, or use real world-data to feed back into R&D.
People are also keen to learn more from their peers and from other areas of the industry. Lab of the Future allows us to host senior speakers and keynotes who can tell us where we’re heading, and show us how the efforts of one area within life science feed into other areas. It presents us with an almost ever-changing jigsaw image, and it’s this strategic element that I think sets us apart from other events.
Who is your main audience for Lab of the Future?
We attract a real mix of attendees, and that’s what I love about it. You can run a conference for people in a specific job function, such as a data scientist or an R&D manager, but what people really want to know is what the people around them are doing, to almost give them context of the industry as a whole. So, our conference doesn’t just exist to help you do your own job better, but it helps you to develop a concept of where your department is heading in the future, and what you should think about longer term. We aren’t telling scientists how to do their job today; we’re helping them think about their responsibilities for delivery in the future. Lab of the Future is about the delivery of science of the future.
Our sponsors and solution providers that support the conference are also very much part of our community, as they’re all innovating and making waves in this space as well. They’re in a space that’s always evolving to build the Lab of the Future; and they are part of that solution. So, we don’t merely facilitate a conference of buying and selling between providers and services, we offer a space where everyone is evolving together. It’s a real melting pot, and that’s the fun bit really.
How do you build the Lab of the Future Community?
Zahid’s background in life sciences definitely gave us a starting point. Further to that, we’ve found that every time we put something out, that our community engages, and as a consequence we’re introduced to people we never expected to be introduced to. The fact we’re always talking to people enriches our content – the people we meet and conversations we have change our way of thinking, and shape what we’re doing.
Although I’m in charge of our marketing operations, I have to say I’m not always sure where some of our contacts come from! One thing I’ve found quite surprising is the lack of reliance on a database – there’s a lot of power in word-of-mouth, especially in this space where everyone is working on something – why not share that? As we’re seen as adding value to the conversation, it allows people to find us through their connections and our supporters.
Scimcon is proud to sponsor Lab of the Future, and we can’t wait to see you at the Autumn virtual congress on 26 – 27th October 2021. Contact us today to learn more about our participation in the event, and stay tuned on our Opinion page for part 2 of our conversation with Luke.
From trends and challenges in the informatics space, to data decoupling, and the impact of COVID-19 on digital transformation projects, Geoff’s conversation with Rizwan and Dave draws on years of experience in informatics consulting and helping scientists to transform their workspace and operations.
I am Mark Elsley, a Senior Clinical Research / Data Management Executive with 30 years’ experience working within the pharmaceutical sector worldwide for companies including IQVIA, Boehringer Ingelheim, Novo Nordisk and GSK Vaccines. I am skilled in leading multi-disciplinary teams on projects through full lifecycles to conduct a breadth of clinical studies including Real World Evidence (RWE) research. My specialist area of expertise is in clinical data management, and I have published a book on this topic called “A Guide to GCP for Clinical Data Management” which is published by Brookwood Global.
Please can you explain what data quality means to you?
Data quality is a passion of mine and now receives a lot of focus from the regulators, especially since the updated requirements for source data in the latest revision of ICH-GCP. It is a concept which is often ill-understood, leading to organisations continuing to collect poor quality data whilst risking their potential rejection by the regulators.
White and Gonzalez1 created a data quality equation which I think is a really good definition: They suggested that Data Quality = Data Integrity + Data Management. Data integrity is made up of many components. In the new version of ICH-GCP it states that source data should be attributable, legible, contemporaneous, original, accurate, and complete. The Data Management part of the equation refers to the people who work with the data, the systems they use and the processes they follow. Put simply, staff working with clinical data must be qualified and trained on the systems and processes, processes must be clearly documented in SOPs and systems must be validated. Everyone working in clinical research must have a data focus… Data management is not just for data managers!
By adopting effective strategies to maximise data quality, the variability of the data are reduced. This means study teams will need to enrol fewer patients because of sufficient statistical power (which also has a knock-on impact on the cost of managing trials).2 Fewer participants also leads to quicker conclusions being drawn, which ultimately allows new therapies to reach patients sooner.
Why is data quality such an important asset in pharma?
I believe that clinical trials data are vitally important. These assets are the sole attribute that regulators use to decide whether to approve a marketing authorization application or not, which ultimately allows us to improve patient outcomes by getting new, effective drugs to market faster. For a pharmaceutical company, the success of clinical trial data can influence the stock price and hence the value of a pharmaceutical company3 by billions of dollars. On average, positive trials will lead to a 9.4% increase while negative trials contribute to a 4.5% decrease. The cost of managing clinical trials amounts to a median cost per patient of US$ 41,4134 or US$ 69 per data point (based on 599 data points per patient).5. In short, clinical data have a huge impact on the economics of the pharmaceutical industry.
Why is the prioritization of data quality so important for healthcare organizations?
Healthcare organizations generate and use immense amounts of data, and use of good study data can go on to significantly reduce healthcare costs 6, 7.Capturing, sharing, and storing vast amounts of healthcare data and transactions, as well as the expeditious processing of big data tools, have transformed the healthcare industry by improving patient outcomes while reducing costs. Data quality is not just a nice-to-have – the prioritization of high-quality data should be the emphasis for any healthcare organization.
However, when data quality is not seen as a top priority in health organizations, subsequently large negative impacts can be seen. For example, Public Health England recently reported that nearly 16,000 coronavirus cases went unreported in England. When outputs such as this are unreliable, guesswork and risk in decision making are heightened. This exemplifies that the better the data quality, the more confidence users will have in the outputs they produce, lowering risk in the outcomes, and increasing efficiency.
Data quality, where should organisations start?
ICH-GCP8 for interventional studies and GPP9 for non-interventional studies contain many requirements with respect to clinical data so a thorough understanding of those is essential. It is impossible to achieve 100% data quality so a risk-based approach will help you decide which areas to focus on. The most important data in a clinical trial are patient safety and primary end point data so the study team should consider the risks to these data in detail. For example, for adverse event data, one of the risks to consider could include the recall period of the patient if they visit the site infrequently. A patient is unlikely to have a detailed recollection of a minor event that happened a month ago. Collection of symptoms via an electronic diary could significantly decrease the risk and improve the data quality in this example. Risks should be routinely reviewed and updated as needed. By following the guidelines and adopting a risk-based approach to data collection and management, you can be sure that analysis of the key parameters of the study is robust and trust-worthy.
If you were to give just one tip for ensuring data quality in clinical trials, what would it be?
Aside from the risk-based approach which I mentioned before, another area which I feel is important is to only collect the data you need; anything more is a waste of money, and results in delays getting drugs to patients. If you over-burden sites and clinical research teams with huge volumes of data this increases the risks of mistakes. I still see many studies where data are collected but are never analysed. It is better to only collect the data you need and dedicate the time saved towards increasing the quality of that smaller dataset.
Did you know that:
Approximately 30% of data points collected in Pivotal medical device clinical trials do not support core endpoints or regulatory submissions for marketing approval.10
And only 18% of clinical trial data in cancer studies is ultimately used in peer-reviewed journal manuscripts.11
In 2016, the FDA published guidance12 for late stage/post approval studies, stating that excessive safety data collection may discourage the conduct of these types of trials by increasing the resources needed to perform them and could be a disincentive to investigator and patient participation in clinical trials.
The guidance also stated that selective safety data collection may facilitate the conduct of larger trials without compromising the integrity and the validity of trial results. It also has the potential to facilitate investigators and patients’ participation in clinical trials and help contain costs by making more-efficient use of clinical trial resources.
What is the role of technology on data quality?
Technology, such as Electronic Health Records (HER) and electronic patient reported outcomes (ePRO), drug safety systems and other digital-based emerging technologies are currently being used in many areas of healthcare. Technology such as these can increase data quality but simultaneously increase the number of factors involved. It impacts costs, involves the management of vendors and adds to the compliance burden, especially in the areas of vendor qualification, system validation, and transfer validation.
I may be biased as my job title includes the word ‘Data’ but I firmly believe that data are the most important assets in clinical research, and I have data to prove it!
Scimcon is proud to support clients around the globe with managing data at its highest quality. For more information, contact us.
3Rothenstein, Jeffrey & Tomlinson, George & Tannock, Ian & Detsky, Allan. (2011). Company Stock Prices Before and After Public Announcements Related to Oncology Drugs. Journal of the National Cancer Institute. 103. 1507-12. 10.1093/jnci/djr338.
4Moore, T. J., Heyward, J., Anderson, G., & Alexander, G. C. (2020). Variation in the estimated costs of pivotal clinical benefit trials supporting the US approval of new therapeutic agents, 2015-2017: a cross-sectional study. BMJ open, 10(6), e038863. https://doi.org/10.1136/bmjopen-2020-038863
5O’Leary E, Seow H, Julian J, Levine M, Pond GR. Data collection in cancer clinical trials: Too much of a good thing? Clin Trials. 2013 Aug;10(4):624-32. doi: 10.1177/1740774513491337. PMID: 23785066.
6Khunti K, Alsifri S, Aronson R, et al. Rates and predictors of hypoglycaemia in 27 585 people from 24 countries with insulin-treated type 1 and type 2 diabetes: the global HAT study. Diabetes Obes Metab. 2016;18(9):907-915. doi:10.1111/dom.12689
7Evans M, Moes RGJ, Pedersen KS, Gundgaard J, Pieber TR. Cost-Effectiveness of Insulin Degludec Versus Insulin Glargine U300 in the Netherlands: Evidence From a Randomised Controlled Trial. Adv Ther. 2020;37(5):2413-2426. doi:10.1007/s12325-020-01332-y
11O’Leary E, Seow H, Julian J, Levine M, Pond GR. Data collection in cancer clinical trials: Too much of a good thing? Clin Trials. 2013 Aug;10(4):624-32. doi: 10.1177/1740774513491337. PMID: 23785066.
12FDA. Determining the Extent of Safety Data Collection Needed in Late-Stage Premarket and Postapproval Clinical Investigations Guidance for Industry. Feb. 2016.
A few months ago I read an article on bioprocess 4.0, which discusses how combining AI and ML with extensive sensor data collected during biopharmaceutical manufacturing could deliver constant real-time adjustments, promising better process consistency, quality and safety.
This led to a discussion with some of my colleagues about what the future of Lab Informatics could look like when vendors start to integrate AI and ML into products such as lab information management systems (LIMS), electronic lab notebooks (ELN) and others.
What is AI and ML?
AI: In simple terms, AI (artificial intelligence) makes decisions or suggestions based on datasets with the ultimate aim of creating truly instinctive system interfaces, that appear like you are interacting with a person.
ML: ML (machine learning) is one of the methods used to create and analyse the datasets used by AI and other system modules. Crucially machine learning does not rely on a programmer to specify the equations used to analyse data. ML looks for patterns and can ‘learn’ how to process data by examining data sets and expected outcomes.
How does ML work?
The following example is extremely simple, but it helps to illustrate the basic principles of ML. The traditional approach to adding two values together is to include the exact way the data should be treated within the system’s configuration.
By using ML, the system is given examples, from which it learns how the data should be processed.
Once the system has seen enough datasets, the ML learning functions learn that A & B should be added together to give the result. The key advantage of ML is its powerful flexibility. If we feed our example system with new datasets, the same configuration could be used to subtract, multiply, divide or calculate sequences all without the need for specific equations.
Where can we see examples of how ML and AI are used in everyday life?
Possibly without realising it, we already see ML in everyday life. When you open Netflix, Amazon Prime Video or Apple TV+ the recommended selections you are presented with are derived using ML. The systems learn the types of content each of us enjoy by interpreting our previous behaviour.
Most of us also have experience of personal assistants such as Amazon’s Alexa and Apple’s Siri. These systems are excellent examples of AI using natural speech to both understand our instructions and then communicate answers, or results of actions. ML not only powers the understanding of language but also provides many of the answers to our questions.
The fact that we all can recognise such an effective and powerful everyday example shows just how far AI and ML have come since their inception in the 1950s.
How will AI and ML affect the day-to-day operations of the lab?
Voice recognition software has been available for decades; however, it has not made large inroads into the lab. It has been used in areas where extensive notes are taken, areas such as pathology labs or for ELN experiment write ups. These are the obvious ‘big win’ areas because of the volume of text that is traditionally typed, the narrow scope of AI functionality needed, and the limited need to interface to other systems.
However, companies such as LabTwin and LabVoice are pushing us to consider the widespread use of not just voice recognition, but natural language voice commands across the lab. Logging samples into LIMS, for example, is generally a manual entry, with the exception of barcode scanners and pre-created sample templates, where possible. Commands such as “log sample type plasma, seals intact, volume sufficient, from clinic XYZ” is much simpler than typing and selecting from drop downs. Other functions such as “List CofAs due for approval”, “Show me this morning’s Mass Spec run” would streamline the process of finding the information you need.
Opportunities to take advantage of AI and ML within lab systems.
Take stability studies where samples are stored in various conditions (such as temperature, humidity, and UV light) for several years and ‘pulled’ for analysis at various set points throughout the study.
The samples are analysed for decomposition across a matrix of conditions, time points and potentially product formulations or packaging types. Statistics are produced for each time point and used to predict shelf life using traditional statistics and graphs.
Stability studies are expensive to run and can take several years to reach final conclusions.
AI and ML could, with access to historical data, begin to be used to limit the size of studies so they can focus on a ‘sweet spot’ of critical study attributes. Ultimately, this could dramatically reduce study length by detecting issues earlier and predicting when failure will occur.
Moving on to lab instrumentation
Instrument downtime, particularly unscheduled, is a significant cost to laboratories. Using ML to review each new run, comparing it with previous runs and correlating with system failures, could predict the need for preventative maintenance.
AI/ML interventions such as these could significantly reduce the cost of downtime. This type of functionality could be built into the instruments themselves, systems such as LIMS, ELN, Scientific Data Management Systems (SDMS) or instrument control software. If this was combined with instrument telemetry data such as oven temperature, pump pressure or detector sensitivity we have the potential to eliminate most unplanned maintenance.
Another major concern with instrumentation in labs today is scheduling and utilisation rates. It is not uncommon for instruments to cost hundreds of thousands of pounds/dollars/euros, and getting the highest utilisation rates without obstructing critical lab workflows is a key objective for labs. However, going beyond the use of instrument booking systems and rudimentary task planning is difficult. Although it is not hard to imagine AI and ML monitoring systems such as LIMS and ELN, there is far more that can be done to ensure this functionality can go even further. Tasks such as predicting workload; referring to previous instrument run times; calculating sample / test priority; and even checking for scientist’s free diary slots are all tasks that can be optimised to improve the scheduling of day-to-day laboratory work. The resulting optimisation would not only reduce costs and speed up workflows, but would dramatically reduce scientists’ frustration in finding available instruments.
Data integrity
Over the last few years, there has been a massive focus on data integrity within regulated labs. However, many of the control mechanisms that are put in place to improve integrity or mitigate issues are not real-time. For instance, audit trail review is often done monthly at best, and generally quarterly. Not only is it tedious, it is all too easy to miss discrepancies when reviewing line upon line of system changes.
ML could be used to monitor the audit trails of informatics systems and instrument systems in real-time and AI could report any out of the ordinary actions or result trends that do not ‘look’ normal to managers. Where appropriate, the system could interact with the corporate training platform and assign specific data integrity training to applicable teams. The potential increase in integrity of data while reducing the headcount needed to do so could be significant.
Final Thoughts
Lab directors, IT professional and the Lab Informatics industry are quite rightly focusing on the digital lab and digital lab transformations. Done right, this will form and excellent platform for the next level of informatics development using AI and ML to propel not just digital science forward, but to revolutionise the everyday life of scientists. Personally, I cannot wait!
To find out more about how Scimcon can support your informatics project, contact us today.
As a leader in a pharmaceutical or life sciences organisation, getting the most out of your team and resources is always a top priority. After making the decision to proceed with a critical investment in consulting services, there may even be more pressure to find the optimal use of these time-limited external resources. So, how can you make sure you are using these resources to their full potential? In this blog, our industry expert Micah Rimer will show you how.
During Micah’s 20 years’ working at big pharma & vaccines corporations, including Bayer, Chiron, Novartis and GSK, he has successfully deployed consultancy groups within lab informatics and clinical projects. Micah has worked with Scimcon to support his teams on high profile critical projects
Frame the problem – what does your implementation project need to achieve?
As with any business situation, it is important that there is a common goal that everyone is aligned around.
It is essential that you do not waste valuable time revisiting the same conversations. Ask yourself: “Is it obvious what problem we are trying to solve?” Often, issues can arise when people are arguing about implementing a solution, whilst losing sight of the challenge at hand.
Take the example of Remote Clinical Monitoring: You might decide that it would be beneficial to have your Clinical Research Associates (CRAs) track and monitor the progress of a clinical study without traveling to clinical sites. That sounds like it could be very promising, but what is the problem that needs to be solved?
Are you trying to reduce the travel budget?
Do you want to increase the number of studies your CRAs can cover?
Are you trying to get access to data more quickly so decisions can be taken faster?
Or do you want to increase the quality of the study?
Without clear goals on what you want to accomplish with Remote Clinical Monitoring, it will be difficult to declare an implementation a success. In addition, if you and your organisation do not know what you are trying to achieve with a particular technical solution, it will be impossible to give your informatics consultants a clear set of deliverables.
So, first things first, agree on the problem statement!
One of the first times I hired Scimcon to support me with an informatics project, I had recently joined a pharma company and found myself in the middle of conflicting department objectives, with what seemed to be no clear path out of the mess I had inherited. The organisation had purchased an expensive new software system that had already become a failed implementation. After spending a year continuously configuring and programming it, it was no closer to meeting the business needs than when the project had started. There were two loud criticisms to address on that point:
those who were upset that the problem had not been solved, that the business was still hampered by not having an overview of the laboratory data they needed; and
the second being the appearance that the department had spent money on software that was of no use.
This also highlighted a far wider range of issues, such as some people who felt their skills were not being properly utilised while problems went unsolved, and that the bioinformatics department might not have the right goals to begin with.
To solve this challenge, we sat down with Scimcon to identify all the different problems associated with the inherited project, and to clarify what we needed to do to turn it into a success. In taking time to review the situation and without too much effort, we were able to come up with four key areas to address:
Understand how bioinformatics/ IT priorities should map to the organisation’s priorities – before we spent any more time and money, what did the organisation actually need?
Solve the bioinformatics problem that the software had been purchased for (assuming that was indeed a verified need).
Determine how roles and work could be shifted and changed so that we were utilising the talents and the resources in the department.
If possible, put the purchased software to use!
With the help of Scimcon, we were able to define these problems and then focus on finding answers to each of the questions. In the end it turned out to be one of our most successful engagements together, award winning even. By just asking senior management what their biggest challenge was, we found their overriding priority was to have an overview of all the R&D projects going on. And while the new software was not particularly well suited for solving the bioinformatics problem that it had been acquired for, it could easily be used to map out the R&D process for portfolio tracking. Then, we turned our attention to the bioinformatics problem, which was easily solved by a bit of custom code from one of the bioinformatics programmers who felt that previously his skills were not being properly utilised.
Once we knew where we were, and where we wanted to get to, all we had to do was get there one challenge at a time.
Manage internal expectations – how will the informatics consultants work with your clincial/analytical teams?
Once you have identified and agreed on the problem that you want to solve, the next step is making sure the organisation is ready to work with your consultants. As with all relationships, business or otherwise, a crucial step is to make sure that everyone has the same expectations, and that all the relevant stakeholders are on the same page.
People have many different perspectives on why consultants are brought in.
Sometimes you need to bring an external voice to convince management of something most of the organisation already knows, as the conversation coming from an external party may be able to have more of an effect.
Sometimes you are looking for external skills and experience that you do not have internally that are needed to achieve a specific goal. Maybe you are trying to implement some new workflow or technology, or you are moving into a new area of responsibility with which you are not familiar.
Other times you may have the skills internally, but not the resource availability. You and your key people are all tied up in the daily operations and have no time to change your systems and processes into a more efficient way of working.
As there can be so many different roles and perspectives on the use of consultants, you need to make sure that you address all the different stakeholder perspectives. It is important to establish a positive situation, as you want the consultants to be able to work with your teams without unnecessary tension.
When I was just starting out with my first LIMS implementation (Laboratory Information Management System), I remember being impressed that you could hire someone who had the specific experience and expertise to guide you on something they had done before but that was new to you. I wondered, “why was that not done all the time? Why do so many implementation projects fail when you can bring people in who had solved that particular problem before?” When I asked Russell Hall, a consultant at Scimcon for us on that first project, he said that not everyone is comfortable admitting they need help. As my career has progressed, I have come to value that feedback more and more. There are many people who are highly competent and effective in their jobs, but are not comfortable with the appearance that they are not sufficient on their own. It is always important to manage for those situations, rather than assuming that everyone will welcome external help.
Lastly, it is also critical to manage expectations, regarding the use of consultants. Your boss may need to defend the budget, or be prepared to stand behind recommendations or conclusions that are delivered from people outside of the organisation. It should also be considered that management might not readily accept something that might seem obvious to employees working at a different level. By liaising with senior leaders from the outset, you can make sure both parties are aligned how the consultants will interact with people in the company, and what their role will be. This is important both to achieve what you want internally and also to make sure the consultants have a proper expectation of how their efforts will be utilised.
Communicate and adjust – how is your information managed between your team and consultants?
While it can be very tempting to feel that you can leave the majority of the project to the experts, the reality is things rarely go as smoothly as planned. As the life science business and information management have advanced over the last few decades, the amount of complexity and details has grown tremendously. It is more and more difficult for a single person to maintain an overview of all the relevant facts. The only way to be successful is to communicate and make sure that the right people have the right information at the right time. Your consultants are no different.
Many organisations have challenges in terms of taking decisions and communicating them effectively. For your consultants who do not typically have all the same access and networks in the organisation that internal staff do, it is imperative that you make sure they are kept up to date. You want to avoid them spending valuable time on focusing on areas and deliverables that have shifted to being less important. Finding ways to keep consultants informed on all the latest developments is absolutely necessary for them to be able to deliver successfully. Figure out what makes sense by considering the organisation culture and the consulting engagement setup. Whether it is by use of frequent check-ins or online collaboration, be prepared to put in additional efforts to make sure that the information gets to where it needs to go.
As well as good communication, organisations have to be able to adjust as needed. Occasionally everything does work out according to plan, but that is more the exception than the rule when it comes to complex life science informatics projects. While timelines and commitments are critical, it is important to view any project as a collaboration. There will be unexpected software issues. There will be unplanned organisational changes and problems. People get sick, life happens. By having open and continuous dialogue, you can be best prepared to make the adjustments needed to find solutions together to unexpected problems.
Ensuring success in your informatics projects
Consultants can be hugely valuable to you and your organisation.
But you have to setup the right conditions for everything to work out well.
Know what the problem you are trying to solve is, and make sure you have as much alignment around the problem statement as possible.
Make sure the organisation is ready for the collaboration by ensuring that your team and management know what to expect out of the engagement, and that your consultants similarly know the scope and what their mission is.
Lastly, you need to keep in constant communication and make sure that you are ready to work together to adjust to the inevitable bumps that will come up on the road.
Working together, you can get to where you need to go.
If you’re interested in working with Scimcon on your upcoming informatics project, contact us today for a no-commitment chat about how we can help you succeed.
In order to work as intended, this site stores cookies on your device. Accepting improves our site and provides you with personalized service. Click here to learn more
