skip_previous play_arrow skip_next
00:00 00:00
playlist_play chevron_left
  • Home
  • keyboard_arrow_right Featured
  • keyboard_arrow_right Thermal Mapping
  • keyboard_arrow_rightPodcasts
  • keyboard_arrow_right Revolutionizing Temperature Control in Pharma [Jakob Konradsen]


Revolutionizing Temperature Control in Pharma [Jakob Konradsen]

Jakob Konradsen June 12, 2024

share close

Yan Kugel⁠ is joined by Jakob Konradsen, the Co-Founder & Head of Quality at Eupry and here he discusses the importance of moving away from paper or PDF calibration certificates and manual work in the GXP industry. They explore the idea of using machine learning algorithms for auditing and the need for a business continuity plan for temperature control. The speaker emphasizes the value of being prepared for potential issues and encourages professionals to focus on improving emergency response plans. The episode ends with a recommendation to connect with the speaker on LinkedIn for further discussions on temperature control.

Jakob’s Career Journey

Jakob started his career as an electrical engineer and later co-founded Eupry, focusing on temperature compliance and mapping. His journey from engineering to entrepreneurship in the pharmaceutical industry is both interesting and inspiring.

Challenges and Innovations in the Industry

The pharmaceutical industry is constantly evolving, and Jakob discusses the challenges related to temperature control and insights that will benefit GMP professionals. He emphasizes the importance of data integrity and the need for more advanced statistical analysis of data.

  • Moving from manual processes to automation
  • The shift towards digital calibration certificates
  • The future of AI and machine learning in the industry
  • Predictive maintenance and its impact on GXP decisions

The Role of AI in GXP Decisions

Jakob raises thought-provoking questions about the role of AI in GXP decisions and the acceptance of risk in processes. He discusses the concept of using AI to find answers faster, while still requiring human validation.

  • Validating AI as a tool in the pharmaceutical industry
  • Accepting risk and the future of GXP decisions
  • The balance between human validation and AI assistance

The Impact of AI on Auditing and Compliance

In this segment of the podcast, Jakob delves into the potential impact of AI on auditing and compliance within the pharmaceutical industry. He discusses the use of AI-driven auditor apps and the potential for machine learning algorithms to assist in the auditing process.

  • The potential for AI-driven auditor apps
  • Machine learning algorithms in the auditing process
  • The evolution of auditing and compliance with AI

In conclusion, Jacob’s insights shed light on the future of AI and data in the pharmaceutical industry, paving the way for innovative solutions and advancements in compliance and quality management.

Stay tuned for more episodes of the Qualitalks Podcast, where we bring you the brightest minds in the pharmaceutical industry to share their wisdom and expertise.

Episode Chapters:

  1. Introduction and motivation for improving industry standards – 0:00 -6:45
  2. Discussion on the potential of artificial intelligence and machine learning in the industry – 6:46 – 15:20
  3. Transition to cloud-based data storage and its impact on the industry – 15:21 – 21:10
  4. Use of machine learning algorithm in pre-meeting with auditors – 21:11 – 27:45
  5. Importance of being prepared for temperature control deviations – 27:46 – 35:30
  6. Conclusion and contact information for the guest – 35:31 – 39:10
  7. Closing remarks and podcast outro – 39:11 – 40:30

Podcast transcript:

Please be advised that this is an AI generated transcript and may contain errors.

00:27 – 01:09
Yan Kugel: Welcome to our podcast episode featuring a prominent figure in the pharmaceutical industry, Jacob Konradsen. Jacob is the co-founder and head of quality at UPRI, bringing a wealth of experience and expertise in compliance as a service. With a background in electrical engineering and a role as an advisory board member of performance contract within digital Metrologue, Jacob has made significant contributions to quality management and calibration within the industry. And today we will delve into Jacob’s career journey, industry challenges related to temperature control, and insights that will benefit GMP professionals in the pharmaceutical field. Jacob, welcome to the podcast.

01:09 – 01:21
Jakob Konradsen: Thank you very much, Yan. I’m very excited to be here. I’m very excited to be part of this. There’s been some great talks. I hope I can fill in the shoes of all the great people you invited on the show. 100%.

01:22 – 02:01
Yan Kugel: So we did quite a lot of events together. You spoke on a lot of the webinars that we ran together. And I think it’s very nice to have you, only you, to chat with us and share a bit about your personal story because I think it’s quite interesting your background and the industry where you’re at right now. So you started as an electrical engineer, and then suddenly you found yourself in a GMP compliant environment where you have co-founded Hubris. So how did it all happen?

02:01 – 02:39
Jakob Konradsen: Yeah, I think it’s an interesting story at least. I thought a lot about this. How do you end up in a position like this? But yeah, as you said, for me, it was really a starting point in electrical engineering and within the space design industry where I studied at university. And we co-founded, I co-founded together with my great colleagues, UBRI, focusing on temperature compliance as a primary focus point, and also this mapping that we’ve been having a lot of webinars on. But I think my sort of intro into the compliance field was really an interest in

02:40 – 03:21
Jakob Konradsen: what I think a lot of engineers can feel like the sensation you get when everything is right. And when everything is controlled, when you try to eliminate errors and really quantify some of all the things you’re doing, trying to quantify how things are stored and how you can enhance that relationship between operations and something that’s really deeply technical with something that people do every single day. So I often like that I came from a technical background because there’s a lot of the things that to me makes a lot of sense, which I sometimes find it hard

03:21 – 03:46
Jakob Konradsen: to understand why that is so difficult. But then I remembered that not everyone comes from electrical engineering background. But yeah, surely it’s been a very interesting journey for me for the last 10 years to go from electrical engineering into compliance. I think some people would say that those are related and some would say they’re not. But I think to a great extent, they are related by last degree. Absolutely. Right.

03:49 – 04:18
Yan Kugel: So you started the UPRE as a project in the university. So in the university, you can try out and do whatever you can imagine and make it happen. So do you feel that changing from this, everything possible to this very compliant industry, so do you feel it’s like constraining you a bit with what you can imagine, or it has a really nice, you know, limitations that actually help you? So how do you see it?

04:18 – 04:54
Jakob Konradsen: Yeah, but that’s actually a great question. And it’s 1 of the things that I like about my own journey is that when you come from a startup perspective, as you said, everything is possible. You don’t have any constraints other than your own imagination. And I think also as part of my job is also to dealing with the calibration and that being an industry which is very, I wouldn’t say old school, but it’s very, you should say, tuned into how you’re used to doing it. And then coming with a fresh mindset has really been sort of a

04:54 – 05:30
Jakob Konradsen: motivator for me, because a lot of the things that I saw in that industry were really classical methods for how to do stuff. It involves a lot of manual work. And as someone who was trained in partly in automation and electrical engineering, that was really where I saw that this could be an industry that could be really interested in trying to digitalize and trying to automate larger parts of it. But surely going back to your question, going from the startup world where everything is possible into a world where much more needs to be documented, validated, and

05:30 – 06:08
Jakob Konradsen: needs to have a much higher degree of certainty that works reliably and for intended use and so forth. I think a lot of people will talk about the whole validation journey. That has been truly interesting because it’s really 1 of the places where you get to see the quality of your own work and the work of others and really trying to help them reduce errors, but also elevate the level of quality of the products and the services that people provide. I think that’s a really interesting part of it. Having experienced both sides of GXP, both believing

06:08 – 06:38
Jakob Konradsen: that it should be something that’s fairly simple to starting to understand it and understanding how difficult it can actually be to ensure that all is documented and all is working as it should. I think sometimes we tend to forget when working within the compliance and quality field that we’re just trying to ensure that things work as intended and that we can document that. It’s not much more complicated than that. But, you know, of course it’s difficult.

06:39 – 07:00
Yan Kugel: Right. So I think coming from engineering, I think it’s something quite familiar because engineering must be very precise. So making it document and then even documenting each stage in a better way actually helps you develop and develop and create a better product actually. Right.

07:00 – 07:39
Jakob Konradsen: Yeah. And I think also 1 of the things that I’ve really come to love is that, or at least 1 of the realizations is that often engineers are considered to be rather uncreative or sort of following the strict path of how things are developed. But I think engineers from my experience, at least, are some of the most creative people. And when I’m truly motivated, or when I’m doing work that I’m truly motivated about, it’s really where you can combine that creativeness with a firm technical background and firm knowledge background in order to sort of be creative

07:39 – 08:08
Jakob Konradsen: about the solutions but still stay within the compliance field and all the regulatory requirements. I think That’s a really interesting area to be in because if you know what you’re doing, if you know the processes on how to document and validate and ensure that what you’re doing is actually of good quality, you actually have a huge amount of creative liberty. That’s 1 of the things at least I take from my background, the sort of the creative approach, but based on a knowledge foundation, you could say.

08:11 – 08:56
Yan Kugel: Great. So that’s very good to hear that as an engineer you don’t end in a complex, you don’t use your creativity at the end. So throughout this podcast season, I’m interviewing quite a lot of entrepreneurs and people who founded their companies, even if it’s small consultancies. So I think it’s quite interesting to hear also your personal view and personal career experience and every advice or everything that you learn in your path. So for me, it’s interesting to hear your thoughts about how your journey started. So you pre was founded from the university. So you really quickly

08:57 – 09:28
Yan Kugel: went through being an engineer and to an entrepreneur where you co-founded a company and you don’t have any business degree in your background, so you had to learn a lot by yourself. So how was it this journey for you? Is it something that you find really fun and you’re really helping doing it? So what were your challenges there personally and so on? So that would be very interesting, I think, to hear.

09:28 – 10:09
Jakob Konradsen: Yeah, absolutely. I think 1 thing that’s common for most entrepreneurs is, at least that’s true for myself, is that I really get motivated by seeing some issue that’s maybe not being handled, but where you can feel this burning sensation inside of yourself saying, I can do this smarter. There is something that must be done smarter or more clever in this way. I think that’s true for all of the co-founders at Jupyter. I sometimes phrase it as I really hate, utterly hate manual processes. I think those are error-prone, those take up valuable time that could be used

10:09 – 10:46
Jakob Konradsen: much smarter elsewhere. And I think that that’s really been my key motivation from the beginning. I mean, Going back from the start, we started at university with a project where we investigated the transport of vaccines that were being produced in Europe and the US, and then being transported to Africa and being transported on ground vehicles throughout Africa. And oddly enough, we found that most vaccines were actually being way too cold. They were being stored way too cold. You could imagine it might have been too hot. But just out of the fear of getting the vaccines too

10:46 – 11:20
Jakob Konradsen: hot, they were freezing them instead, making them really not that useful. And that was really something that opened our eyes, saying, OK, there’s something here where just by getting this data back, you can make much smarter decisions. We sort of went away from that idea and looked towards the first, the Danish and then the European and now international market in terms of temperature compliance. And what we really saw at that point in time, which is, I mean, quite a few years back now, was that a lot of the solutions out there were not really being inventive

11:21 – 11:56
Jakob Konradsen: in terms of how you managed temperature compliance. I think back then we were just looking at making digital thermometers. I think at some point in time, it dawned on us that the issue is not necessarily that people would normally like to say that people want a faster horse, want a smarter, more precise the moments. I know they want a solution that could take care of their needs and really delving into over those first years what is truly the need of the GXP industries and all the other related industries is not only to have sensing of your

11:56 – 12:33
Jakob Konradsen: materials or products or whatever it might be, but having that whole sensation of being absolutely sure that those are stored correctly and the documentation is there, that it’s being calibrated, that you can note all your deviation. All that stuff that’s surrounding storage compliance is really something that I found to be not really been invented on for a lot of years and that really intrigued us into doing… If we were to invent the perfect solution, what would that be? I don’t say that we have the perfect solution today, but It’s that constant drive towards doing something that

12:33 – 13:11
Jakob Konradsen: is smarter. I talk about it a lot with my fellow co-founders, is that when we sleep at night or when we wake up in the morning, we’re just constantly thinking about how can we improve something. And it’s not always something that might be a major issue, but just small annoyances just keeps you driving towards something that’s more interesting. And as you mentioned in the intro as well, I’m part of this advisory board for digital metrology and where we’re looking at digital calibration certificates. And it’s really been a dear sort of idea child of mine that this

13:11 – 13:43
Jakob Konradsen: should be something that should be much more out there. I mean, just the concept of having calibration certificates on PDFs, as we do as well. Today, it just blows my mind that the industry hasn’t moved forward. I mean, the whole idea of a paper or PDF calibration certificate is just odd. I mean, if you were to come into the industry, just not knowing anything about how it used to be, just coming in from the side, you would say, why are you still storing those calibration certificates in such a stupid way where no machine can read them

13:43 – 14:21
Jakob Konradsen: and you really can’t get data out of it? So that sort of constant drive is what motivates me at least to still keep on this journey and keep on getting it out to more people and talking about it and really trying to get it out there. That’s really my largest motivation. I’m honest to say I’m not being the most motivated by getting a large business. I’m really motivated by all the people who have to do manual work. I mean, I truly honestly feel for them, for everyone who has to do manual work. And in the end,

14:21 – 14:48
Jakob Konradsen: I do honestly believe that it creates a better world. It might be a big thing to say, but I honestly think that if we can get more of these manual processes out of the way, we can focus on developing new types of drugs. We can actually help people rather than just doing manual tasks. Yeah, I get excited talking about it, but it’s really what I feel is that There’s so much wastage of time and resources in this world.

14:48 – 15:40
Yan Kugel: You can shift so much resources to doing the important things instead of manual documentation or manual review, etc. So it makes so much sense. And just yesterday I recorded another podcast with Paul Suresin, who is a GDP expert and consultant. And 1 of the things that came up when I asked him, OK, what do you see shifting in the industry? So when he started his journey many years ago, he seen especially the technological change and shift to automation and he mentioned specifically also temperature controls. And so that was a very interesting. So to build upon this,

15:40 – 16:22
Yan Kugel: to understand where the industry is at right now. So I know that before that, and still a lot of companies use temperature control units which are working with USB. So they need to copy the information they need to take the manual USB, plug it in. So much can happen with this drive until it comes to the computer. And looking at this, we still have companies that work like this, but a lot of companies are shifting to other solutions. So where do you see this industry? Where is the change in this industry when you’re looking at from

16:22 – 16:33
Yan Kugel: your perspective, from your competition and what is happening right now in the industry and what do you think is coming soon also as well?

16:34 – 17:08
Jakob Konradsen: Yeah, it’s a really good question. I think the only thing we could be uncertain about is the future, of course. But It’s often funny for me to look back at 10 years ago when we started this company and where were we at, not only us, but the whole industry. And I think at that point in time, a lot of industry was looking for a faster horse. They wanted USB data lockers that had longer battery life, that had more memories, that you could store more. And then I think around 7, 8 years ago or something, we started

17:08 – 17:44
Jakob Konradsen: to see a lot more of these wireless or wired lockers and monitoring systems that would allow for getting rid of a lot of the manual work. And in my view, it’s kind of been sitting there for quite a bit of time now. But now I’m starting to see that we’re getting past the hurdle of automatic data collection. That’s 1 of the parts that’s pretty much closed down now. And I’m really happy to see that because, I mean, just the work of emptying USB data logger is just something that really does not make sense in 2024. There’s

17:44 – 18:18
Jakob Konradsen: still a few out there, but I think most what we see both with us and our competitors is really that USB and this whole manual processes, those are being outphased. And now, in my view, there’s other topics. Right now, we’re seeing the hot topics is data integrity and how do we ensure that the integrity of the data. And I really see that as a symptom of us having moved forward as an industry, really having gone away from using all of our time on these manual tasks to now starting to look at, okay, the quality of the

18:18 – 18:52
Jakob Konradsen: data that we’re getting. How is that both in terms of integrity? But I think also 1 of the things we’ll see in the coming years is, first of all, I think for me right now, the hot topic is really digital calibration certificates. I know that Germany by the PTB and Denmark is really on the forefront on trying to roll this out in terms of digital calibration certificates. And as a sort of a side note to that will also be the whole digitalization of calibration and all those activities related to that. I think also another major part

18:52 – 19:30
Jakob Konradsen: we’re going to see in the coming years is really also more advanced statistical analysis of our data. 1 of the sort of, you could say, side effects, I think, of an industry that’s been laying dormant in terms of innovation is really that today we’re only looking at whether our temperatures are within or outside our limits. And another factor is that world calibration needs to be done once a year, on some cases every third year, whatever. It doesn’t state in this way specific other than the GDP. And there’s some few regulatory guidelines that kind of tell you

19:30 – 20:03
Jakob Konradsen: the calibration interval as an example, but it’s really based on guesswork today and sensors are so good today you don’t need to calibrate them once a year if you don’t have something where you need a very precise measurement. But what I want to get to is that I think more advanced statistical analysis where we can actually have a risk-based approach to whether we need calibration is 1 of the very interesting projects, at least that I see ongoing, for example, at the Technological University of the Institute of Denmark, for example, is looking at, which is a collaboration

20:03 – 20:40
Jakob Konradsen: between more countries. I need to say that. But we’re really looking at, can we use data to tell us something more intelligent about both our storage conditions, but also how our sensors are doing to evaluate whether we need calibration? But this is probably not something I should say, being in a company that sells calibrations, but often it’s not really technically founded why you need calibrations every year, as an example. So I think a lot of this whole using data in a more intelligent way is going to be something that’s part of where we’re going as an

20:40 – 21:14
Jakob Konradsen: industry, because we’re sort of moving away from where it’s really non-digital. I would say USB loggers are non-digital because you just get the data and you can do something with it, but it’s not really automated to getting into a scenario where we have vast amounts of data and we need to decide what to do with that data to make intelligent decisions. And that for me is really 1 of the things that keeps me motivated to be in the industry, because there’s so much data in there. I mean, we’re even talking like artificial intelligence and machine learning.

21:15 – 21:35
Jakob Konradsen: Can we do something to detect whenever we have machines that is about to break down? Can we sort of predict non-conform temperatures as an example? There’s a lot more things I think we can do with the data. And that’s going to be so interesting seeing the next years, what’s really going to happen in the industry.

21:36 – 22:24
Yan Kugel: Yeah. So this is exactly what I wanted to ask you more about. And you just already mentioned it, especially the AI. So we know that in terms of devices that store information, so the transformation was incredibly quick, right? So we had the old music records and within some years, then we had the cassettes and then the videotapes and the videotapes then turned into CDs and everybody thought, okay, what can you think about? And from the CDs we had the USBs and then you had the hard drives and then you just have the cloud. So, and all

22:24 – 23:01
Yan Kugel: this happened within, I don’t know, like within a span of like 30 years. So this is just incredible, right? So do you think it’s difficult to predict the future, but do you think is it possible that some kind of an additional tool will arise that will be better than the cloud? Or it’s probably already we reached this point where it’s already in the most efficient place. It’s more on how quickly and safely you can retract the data, I guess. But I guess what is better or quicker? You just can maybe improve the servers, right?

23:01 – 23:39
Jakob Konradsen: Yeah, yeah, yeah. I think also that’s also actually also part of where the industry is moving. It’s been really interesting for the last 10 years seeing that the companies are moving from what we need on-site data centers and so forth to actually, I think, acknowledging that cloud is in many ways more secure if you do it correctly. I think a large part of at least what we’re seeing that customers are asking for in the future is really How can you prove, how can you convince us that cloud storage is still safe? And in some senses, it’s

23:39 – 24:16
Jakob Konradsen: a step back because I don’t think anyone, I think most companies don’t want to step back to onsite data storage or at least not fully. I think what we’re seeing at least is there’s a large demand for having a backup on an onsite storage. But I think the whole backup part is really going to be interesting to see how that problem is being solved. Of course, the traditional backups today, and I think it’s very few systems who experience any sort of loss of data or backup in that sense. I don’t think that’s a place where we’re

24:16 – 24:54
Jakob Konradsen: going to see a large amount of improvement over the next decade or 2. I think it’s really going to be on how do we use the data, how do we ensure the integrity of the data, because introducing new models and machine learning and decisions really being made automatically without us truly understanding the underlying statistical tools being used is going to be something that’s going to be a great gray area. Because 1 thing is that we understand, as most people working with culture, we understand the statistical models we use to understand, for example, temperature data. We use

24:54 – 25:30
Jakob Konradsen: mean kinetic temperatures. We can go forth and back between different SI units and the metric and the period system and so forth. There’s a lot of things that we can intuitively understand, but I think in the coming time, it will be really interesting in how much can we rely on stuff that an AI has generated. I mean, how much can we truly appreciate and how much value can we gain from that without losing control? I think that that’s truly going to be 1 of the few things that we’re going to be looking into in the future

25:30 – 26:01
Jakob Konradsen: is what can we trust and what can’t we trust in order to make GSP decisions? I talked with someone on the topic from the industry, and I think he very cleverly said that, well, every operator or every person working within a GSP facility needs GSP training. How do you make sure an AI has gotten GXP training, right? Right. I mean, it probably has to some extent. I’m not an AI expert, but how do you truly understand whether that model has understood what is important in our industry?

26:01 – 26:38
Yan Kugel: Yes. So that’s a very interesting point. So a few podcasts back I recorded 1 with Fabrizio who is the director of business development in Sparta Systems Honeywell, right? And We talked about AI and he has a very interesting point of view on that because he said you cannot validate an AI as a tool, especially the LLMs because the large language models like Chia GPT because they can always give you a different result when you ask them the same

26:38 – 26:39
Jakob Konradsen: question.

26:39 – 26:46
Yan Kugel: But you actually need to validate them and qualify them as if they are an employee.

26:46 – 26:47
Jakob Konradsen: Yeah, Yeah, exactly.

26:48 – 26:59
Yan Kugel: Because each employee has its own risks, and it all depends how well you train them and make sure they’re within the lines of the operations. So what do you think on this?

26:59 – 27:43
Jakob Konradsen: Well, I think this is such a great statement as well. I think it’s really interesting. 1 facet of this whole GXP industry that I always found interesting is that we allow humans to do a great amount of mistakes, but we allow humans to do 0 mistakes. And I think the true question is, will we gain more value whilst accepting that there might be mistakes or failures in systems than the mistakes and failures that humans do. I think 1 of the concepts that we’re working with a bit here at Uber is something I know a lot of

27:43 – 28:26
Jakob Konradsen: companies have been working with, which is predictive maintenance. Can we predict whether fridge or freezer or whatever might fail ahead of time? And 1 of the interesting discussions I’ve had with colleagues in the field is, is that a GXP decision? The famous example is always, you don’t need to validate the computer system that tells everyone what’s for lunch today. You don’t make any GXP decisions on that. But on the other hand, if people don’t get lunch that day, they’re more likely to do mistakes. So in the same way, can we add to the value of GXP

28:26 – 29:12
Jakob Konradsen: industries as a whole without AI? What is the logical next step? And when will we start differentiating between what is a true GXP decision and what is not? 1 of the things we’ve been experimenting with here internally in EUPRI is basically using a variant of chat GBT, which has been trained on all of our quality procedures, all our SOPs and quality manuals, in order to find answers faster. Not to give the answer because the system basically tells you where that is stated in an SOP. But I would argue that getting an answer faster, making it less

29:12 – 29:45
Jakob Konradsen: troublesome to search for this than typical search methodologies is something that adds to the value whilst you can look at the answer from CHAP DBT, but you still have to look it up in an SOP. So who’s the smarter person here? I mean, I think it’s going to be truly interesting. What is, Who do we train and why? Where do we accept risk? I mean, we’re used to accepting risk in a lot of other processes. I’m sticking my head out there. I know suggesting that both that humans can make mistakes and that we should accept some

29:45 – 29:47
Jakob Konradsen: risks. But it is true. Part of

29:47 – 30:36
Yan Kugel: the risk, right? And it all can come down to ForEyes methodology that you still have a human checking on the AI, right? So it’s the same way. So you need to review it. So it’s quite, indeed, an interesting point here. And by training the language models on specifics, because if you take the Chai GPT itself, if you ask it questions about GMP, it will tell you. But researchers found that since GMPT went live, it became more dumb. And It’s funny because the more in the general Chai GPT, right on the server interacts with more people and

30:36 – 31:16
Yan Kugel: gets more information, somehow it gets a dumber instead of smarter, right? So the solution for that is actually using local separate Chai GPT servers, right? And training them on your specific needs. Like you said, like, okay, you know nothing about what’s happening right now in China, but you know only what’s happening right now in your group. So then you can all the knowledge is here. And the data retraction, I heard it from a lot of people with whom I was speaking from all areas in the GXP, like auditors. So sometimes they go in an audit, they

31:16 – 31:53
Yan Kugel: need really quickly to know the found deviation, but they don’t remember where exactly this regulation is quoted. So then instead of starting, you know, browsing through the whole list of standards or GMP regulations, because everybody cannot remember word to word every page. So you can just ask, okay, this deviation, where can I find it? It gives you the, as you said, the quote, okay, it’s here in this chapter. And this is already saves, can save hours upon hours within your work. So that’s quite amazing, I think.

31:53 – 32:23
Jakob Konradsen: It’s actually quite interesting because 1 of the ideas that I’ve kind of been tumbling around with and trying to just set up slightly on a test base is really that how can we, because we get tons of customers auditing us, which is, I mean, that’s what it’s like being part of this industry, right? But a large part of the audits that we get are sort of getting to know each other, how are systems built up, which kind of products do you have specifically and which ones are we using and so forth. So I’ve really been actually

32:23 – 32:55
Jakob Konradsen: been trying out what if we set up a pre-meeting with the auditors where I’m not present or any of my colleagues are not present, but a machine learning algorithm is present. It could give a lot of the foundations already, which kind of which SOPs are relevant to look in, which kind of systems, which kind of documentation do we need to look at? What do you have and how do you formulate it? I think that could be a very interesting point. Maybe we’ll even end up at a point where 2 language models are auditing each other based

32:55 – 32:59
Jakob Konradsen: on the material they will get. There’ll be no people involved. I don’t know.

32:59 – 33:33
Yan Kugel: That’s very interesting. So usually, they’ve been in audit for myself, right? I’ve been doing audits quite a lot and a lot of the hassles is giving the audit, the list of SOPs you want to read or regulations and data and that and learn about it. And of course, but the problem is companies, they don’t want to reveal too much when they don’t have to, right? So if you give them the access to the whole system, it’s clear that a lot of questions can be asked, a lot of answer can be given. So you need to make

33:33 – 34:10
Yan Kugel: sure, of course, so you need to trust your compliance. And it’s of course, and I think this can be 1 of the most honest way to do it for an auditor is great. So you can just ask, okay, how do you deal with deviations? What is so piece? So I think your idea is quite incredible, can save so much hours for auditees, because I’ve been also on the auditee side where I had to host audits by clients. And for each audit, So I worked in an API manufacturer and there we had almost every week an audit,

34:10 – 34:46
Yan Kugel: right? And you had to prepare for each as if it’s your first audit, right? So you need to do it from scratch. And what you’re saying is actually a great idea. It’s just, you know, the company is not everybody, every company will be ready to just give away because it’s 1 thing to sit and approve which document they get and be near the auditor because you know what they’ve seen and what they haven’t seen. Right. But it’s a very interesting concept. But I think the idea of just being open, it’s revolutionary, but it can be quite

34:46 – 34:48
Yan Kugel: great. I think FDA should make it obligatory.

34:51 – 35:20
Jakob Konradsen: But I think it’s again, it’s another symptom of this where we’re going away from some of the manual labor, which I mean, auditing to a large degree is manual labor, looking through, reading through documents, trying to summarize them your head and then getting to the actual value. And that’s really what drives me Atlanta. What I find so interesting is that how can we get value from these things, which are basically a lot of within temperature compliance is a check or no check, right? Of course, there are gray zones and there’s a lot of things. The interesting

35:20 – 35:54
Jakob Konradsen: parts are where you need to make an informed decision on something, whether it’s okay, do we need to do a recall, whatever it might be. But we’re getting to a point where we’re actually getting value for the human hour. I mean, because there’s so many things. And also I think 1 of the, exactly as you mentioned, 1 of the back sides by being very open, which is also true when you have a digital system for storage of temperature compliance, which has been for many, many years, where it’s easily searchable by more statistical means, it’s also that

35:54 – 36:33
Jakob Konradsen: you’re opening up 100%. And I think 1 of the things that industry needs to learn as well is that No company is perfect. I mean, I heard someone say that if you’ve been audited and there’s been no nonconformities, the auditor wasn’t looking well enough. A healthy patient is just a patient who hasn’t been examined well enough, right? There will always be something. And I think it’s part of the improvement. Yeah, exactly. In a very beautiful way that goes down to the continuous improvement idea. Right. It’s really about improving. And then you have experienced this as well

36:33 – 36:42
Jakob Konradsen: as both audit and auditors. The best audits are the ones where both parties learn something. Yes. Right. And

36:45 – 37:32
Yan Kugel: The biggest faults you see with companies that become arrogant with time, right? Because they feel they are the best and everything is great and you see it in the biggest company in the world. So You’ll see it in Pfizer, the pharmaceuticals buyer, where they had to shut down the whole plants because they just got very comfortable for years. And they just stopped doing the changes because they’ve seen, okay, everything is great. And sometimes the issue here is sometimes they’re getting audited always by the same regulator within the agency that comes to them and they just already,

37:32 – 38:04
Yan Kugel: it’s a people world, right? So they get friendly, but then the person changes, his attitude changes. He’s the way of looking at things changes. And then you auditor comes suddenly, they find something completely different. And then it’s a curve ball. So you should, I think, you know, as an ODT, you should never be arrogant to it. And the people in those companies really need to implement better internal audits and really take better care of having good internal auditors that really scrutinize everything, right? And give suggestions.

38:05 – 38:16
Jakob Konradsen: Absolutely. I think it’s going to be really interesting what we can do with technology, right? Because auditing is really 1 of the last domains where we haven’t seen automation just yet.

38:18 – 39:01
Yan Kugel: Yes. So I see it a bit. So a while back, I was talking with some great guys who had a startup on AI driven auditor app that actually walks you through the audit and reminds you, okay, check this, check that. So I need to check with them what’s going on with that. So hopefully they’re still in business. But Going back also to your activities, right? So you are hardware and software provider, let’s say, but you also give services in terms of calibration. So what have you learned by working with a lot of clients on the biggest

39:01 – 39:07
Yan Kugel: challenges that the companies face in terms of temperature control?

39:08 – 39:41
Jakob Konradsen: I think it really comes down to exactly what you said just there. That what we found, at least when we started out, was that the challenge was not really in the software. It was not really in the hardware. It was not really in the calibration. It was in the service that people needed. I still see to this day that some of the customers we have with the highest or the most challenges are the ones that have difficulties in managing your whole fleet. And when is something due to be calibrated and when is it not, and what

39:41 – 40:19
Jakob Konradsen: is in compliance and what is not. I’m starting to see a shift away from those being the major challenges, but now it’s more of a question of finding out the scope of temperature compliance. For some companies, it’s very clear which areas need to be temperature monitored and not. And I think that’s true for the most of it. But really understanding which kind of calibration do I need? What is a calibration? Sort of in some sense, some of the more basic of information. What is it that we need? What should we set as acceptance criteria? What can

40:19 – 40:53
Jakob Konradsen: we tolerate? I think, as I said earlier, it can often be sort of a check-no-check scenario where either it’s within the limits or it’s not within the limits, but the more deeper understanding of which deviations can we actually accept and which can’t we. That’s really 1 of the places where I see that people don’t prepare enough for. Because the 1 thing that is true and always will be true is that there will be mistakes made. And not only mistakes, there will be failures, whether those be due to systems or hardware, whatever it might be, which causes

40:54 – 41:28
Jakob Konradsen: deviations in, for example, temperature. But what do you do in that case? That’s 1 of the places where I see people don’t truly understand their products and what they’re storing, what they’re actually capable of. I mean, just because you have a small deviation in temperature does not mean your product temperature has changed, as an example. And even if it has, Is it really critical? I mean, 2 to 8 degrees, is that really, really critical that it’s 9 degrees then? And where does that responsibility lie? But I would say, and I have to give kudos to the

41:28 – 42:04
Jakob Konradsen: part of the industry that we’re touching base with, there’s a really high knowledge level in general. 1 of the things we’re starting seeing is that, and this is my own theory, that because the knowledge level is so high on temperature compliance, and what needs to be monitored is that those requirements start to trickle down into other parts of the industry, the more transportation part, the same point of sale part, that those need to be monitored as well because I see there’s a good level of control in the sort of the production and the primary distribution, but

42:04 – 42:42
Jakob Konradsen: the last parts of the distribution chain and the point of sale, we start to see those regulations being trickled down. I think 1 of the things that started our company a lot of years ago was that pharmacies had to start monitoring temperature of medicinal products, which is something I mean, that’s not 10 years ago that they needed to do that. And that’s kind of a bit of frightening. But really, you can start seeing that governmental institutions start to understand that it’s something that’s easily enough done, that more parts of the industry and larger life cycle of

42:42 – 42:49
Jakob Konradsen: the products need to be monitored. It’s always difficult to point to 1 huge challenge.

42:51 – 43:06
Yan Kugel: Yeah. Right. So they feel that companies underestimate a bit the importance of temperature compliance and monitoring for their product. And also some of the personnel are not trained enough to understand?

43:06 – 43:47
Jakob Konradsen: That’s a really good question actually. I think, let me put it this way, I think a lot of companies understand that it’s really important. But what is it actually that is really important is sometimes where it lacks a little bit. What actually is critical and what is not critical. It’s always fun to see we have a large amount of our customers at distribution. Customers still handle intermediate storage or longer-term storage of medicinal products. And they have a huge degree of control with their products and how they are stored. But then if you go to somewhere lower

43:47 – 44:20
Jakob Konradsen: down the chain, you’ll see that there’s less control and you might think, well, is it becoming less important? I don’t really think it is, but I think the understanding of what the limits should be and what we are allowed to do differs a lot from when the life cycle a medicinal product is. That’s really interesting to see how things are handled differently in the whole cold chain of a product, which in my mind should be the same. I know that there’s a value add down the line because it’s been transported and more cost has been added,

44:20 – 44:24
Jakob Konradsen: but in theory, it should be the same. Yeah.

44:24 – 44:56
Yan Kugel: Right. So you feel that the quality people are very concentrated on the manufacturing itself and they feel less responsibility the moment the product leaves the manufacturing floor or you know even stores in their in their warehouses because they feel okay my manufacturing part is over this is the most critical we don’t we don’t have cross-contamination or even further when it leaves the plant and they say, okay, my hands are clean. That’s it. It’s not my business right now. Yeah,

44:56 – 45:33
Jakob Konradsen: actually, I think that’s 1 of the things that I really start to notice is that the responsibility stays with the product a lot more with the people who are in the manufacturing side. I see that manufacturers are starting to take responsibility further down the distribution chain. They start to really ask the distribution supplies, how are you storing this and how well is it being kept? How well is your level of quality? And I think with the distribution companies that we worked with, they really walked to the, or stepped up to the challenge. I mean, the, the,

45:33 – 46:09
Jakob Konradsen: the, the level of quality I’m seeing is just rising in the industry in terms of the distribution. I think it’s, it’s like, I mean, hands down, huge kudos to the distribution industry, because really it’s like you’re, you’re handling on 1 side, this whole logistic, from my point of view, logistic nightmare of keeping thousands and thousands of pallets on the road and getting fast from A to B, while keeping this really sensitive thing with temperature really tight and controlled. I’m truly impressed by that industry specifically. Also because it’s not an industry who has had decades to learn

46:09 – 46:31
Jakob Konradsen: from best practices and how to do this. A lot of the distribution customers that we’re getting in, cold chain management is something that’s not brand new, but relatively new within the last decade or so. And I think it’s actually quite impressive to have built up such competencies in those companies to be able to fully comply with all the regulations there.

46:32 – 47:17
Yan Kugel: Right. So that’s great. So you basically, you know, confirm what Paul said to me yesterday, Paul Serres, in the GDP expert that he also sees an incredible shift in the responsibilities that the transfer providers take. And a lot of them seems to be more involved to understand GMP and GDP and train their personnel much better. And you see it and you feel it because I think the competition between the suppliers gets bigger because you as a pharma company, API company, you will come and you will choose the providers can showcase that they have the right certificates

47:18 – 47:23
Yan Kugel: and that they train, right? So this place is big. Right.

47:23 – 47:32
Jakob Konradsen: And absolutely. And I think it also, to a large degree, it has something to do with that being part of the selection criteria for the distribution partners, right? I mean, let’s

47:32 – 47:33
Yan Kugel: be

47:33 – 47:54
Jakob Konradsen: honest, price is a large part of choosing a provider like this as well. But actually, it makes me kind of proud being in an industry where price is not the primary point, we start to see quality and the amount of control you have over temperature compliance being a very large part of

47:54 – 47:54
Yan Kugel: who

47:54 – 48:00
Jakob Konradsen: is selected for distributing this. And I think that was a great amount of responsibility, to be honest, from the manufacturer. 100%.

48:01 – 48:37
Yan Kugel: And I think it’s also because companies understand that 1 fault can cause them, again, a lot of money because if they need to close their facilities for even, I don’t know, a couple of weeks, they lose millions. It’s better to make sure that everything is controlled. So even if you’re going to the management side and not only the quality, that they should always take care of the quality and consider it their highest priority. But even the management, they understand, Okay, if we close the plant or if the supplier is not reliable and the product is lost

48:37 – 48:46
Yan Kugel: and they need to recall the batches because of some temperature issues. So, it’s better to invest the money now than be sorry.

48:46 – 49:12
Jakob Konradsen: Yeah, absolutely. Yeah, I think also we’re also sometimes part of sort of the selection process. How can we as a company live up to and exceed those requirements? I think it’s also a large part of is, is what I see is also not only due to risk, but also a true desire to keep a high level of quality. And that gives hope to humanity at least.

49:12 – 49:13
Yan Kugel: Yeah, exactly.

49:14 – 49:21
Jakob Konradsen: I mean, people feel a true responsibility to the inpatient. I mean, that’s very much in the center of what we see, at least. Right.

49:21 – 49:52
Yan Kugel: That’s what we need to keep in mind. I think that’s why you also have so many certificates in UPRI. Right. So, lastly, If pharma professionals want to take their temperature control to the next level, want to improve, take the next steps, what would you say would be the main steps that they should take with the least energy to make the most improvement?

49:54 – 50:30
Jakob Konradsen: I think this is often what I tend to answer. So what is the lowest hanging fruit in terms of staying temperature compliant? I think I won’t fall into the trap of saying just buy Ubrim. But I think having a system, of course, in place, but it’s very easy to get a very reliable system, being it also a competitors or whatever today. The lowest hanging fruit for me is really to be prepared for when something happens. We start seeing customers who will get non-conformities from auditors, not because their temperature compliance is not in place, not because they

50:30 – 51:05
Jakob Konradsen: don’t have monitoring, not because it’s not calibrated, all those basic things, but not being prepared for what happens if something goes wrong. I mean, it’s 1 of the questions that is truly good to answer yourself. What happens if we get a temperature alarm on Friday evening at 11 o’clock. It’s easy to answer it if it’s a midday Wednesday, but what happens if it’s Friday at 11 o’clock in the evening? Who will respond? What will we do and how will we evaluate those changes? It’s something that in my mind, at least, it is relatively simple to go

51:05 – 51:33
Jakob Konradsen: through. What is our emergency plan? What would it happen to do? And how will we evaluate those results? We also being used sometimes in the evaluation of this data comes out as saying we’ve had this deviation of temperature. What would you suggest and so forth? Those are really some… It’s really a place where you can easily upfront take those evaluations and sort of make a decision tree. Of course, there will always be gray zones. What should we do in this case? What should we do in this case and what should we do in this case, but

51:33 – 52:05
Jakob Konradsen: really being prepared for what happens if something goes wrong. That would be the easiest step for me. Keeping your, then all the basic stuff, of course, keep your calibration certificates in order, review them, keep your deviation management in order, make sure your alarms set correctly. But I think most of those things are pretty basic, but really preparing for what is going to happen. That’s a place where we start to see auditors really look for Sort of like a business continuity plan, right? But but with even more specific to this field

52:06 – 52:40
Yan Kugel: Yeah, that’s great Great cool Jacob. Thank you very much for you know, having this chat with me very interesting Topic or at least you’re making it very interesting. So it was a pleasure. So if you have questions about temperature control, I think Jacob is really your guy, so you can connect with him on LinkedIn. His information will be also in the post about this podcast episode. So feel free to reach out. I think Jacob will be happy to chat with you, right?

52:41 – 52:50
Jakob Konradsen: Absolutely. And thank you very much for allowing me on the show, Yan. It’s been a true pleasure. I think it’s always great when you and I keep talking. We could have used a few more hours.

52:51 – 53:01
Yan Kugel: For sure. So we will consider another episode for another topic within temperature controls. Super fascinating. Thank you, Jakov. Have a great day.

53:01 – 53:02
Jakob Konradsen: You too.

Tagged as: , , .

Rate it
Previous episode