Episode 51
Unmasking the true cause of quality issues in Pharma labs!
Category : High Variability Operations
Tune into this enlightening episode exploring USFDA warning letters affecting global pharma companies, with Abhinav Srivastava from Vector Consulting Group. We'll dissect analyst error, investigate its role in quality issues, and reveal Vector's innovative solutions. Don't miss these impactful insights.
Transcript
| Shubham Agarwal : | Hello and a very warm welcome to the counterpoint podcast I’m Shubham Agarwal. Now across the world, the US FDA is responsible for inspecting and reviewing the lab operations and drug development process at the pharma companies. In the event of a lapse in the process, or a concern that the FDA identifies they handout a warning letter to the company. Now, they are given a timeframe to rectify and submit the corrective actions within a within a stipulated time. Now, having done this over and over again, for years, one would assume that the pharma companies and industry in general is learning from their experience, and would have reduced the number of these letters year on year. However, the truth is entirely opposite. The warning letters have been consistently on the rise. But is it really such a big problem that needs addressing? Why does it happen in the first place? What is the core reason? And is it common across all pharma companies? Are some of the questions that we should explore today in our episode? And of course, how do we go about solving it? For the discussion, we have Abhinav Shrivastava with us today. He’s a consultant at the Vector Consulting Group. And he has, and he has helped many pharma companies set their processes right, and reduce the occurrence of these warning letters substantially. So let’s welcome him and discuss with him this topic. Hi, Abhinav. Welcome to the counterpoint podcast. How are you? Hi, Shubham, I’m good. How are you? I’m great. Thanks. Thanks for being here Abhinav. So let’s dive right into the discussion. And, you know, let’s start with understanding that, one could argue that some of the errors, some of the mistakes are unavoidable, you know, humans are involved in the process, there are human errors as well. And they’re quite complex processes that are being run by these pharma companies. So, so why make such a big deal of all of this, you know, how do these warning letters impact the organisation? |
| Abhinav Shrivastava : | Yeah, okay. So, there are two parts actually, one is, you know, so you said unavoidable, and why to make a big deal of this. So, so, actually, if you see, for the consumers, like you and me, who are consuming this drug, what is most important is can we trust it, right? The drug has to do the drug content, and, you know, whatever the impurities are, the manufacturer claims that this is the level and it should meet that level. So, and it becomes safe for us to consume a drug or, or trust brand or a trust a drug. And this is this is where these bodies come into play. And, you know, they come in they check they are, they give us the assurance that whatever you’re consuming, it’s been tested, it’s been approved, it’s, it’s compliant, okay. And that is why they are very much you know, into, into doing these audits to go and check around the world, all these manufacturers, whether whatever they are claiming is what is actually getting shipped, and manufactured. So, so any mistakes or any lapses in these process that they that they find, and that is why they, you know, they try to control it through observations through interventions through those warning letters. And that’s why they make a big deal out of this because any such lapses in this and people missing out such broad gaps in the quality can impact a lot of things in the market. So the trust can go away, you may have to from the from the manufacturer point of view, you will be imposed on lot of fines lot of restrictions to introduce new products in the market. Also, there’s |
| Abhinav Shrivastava : | A financial impact as well with these. Yeah, |
| Abhinav Shrivastava : | Significant. So, for the organisation, it will have a severe impact. So, that’s why they make a very big deal out of it and the first part that you said you know, these mistakes which are unavoidable. So, okay, we will come to what is unavoidable or, or avoidable? Yeah, what they are actually looking is whether the, the process is it trying to detect a mistake or not. So, there will be some mistakes naturally, it’s run by a human, as you said, there will be mistakes. So, is your process robust enough to catch or detect those? Right? And if let’s say they are, there are genuine mistakes in a drug, you have done something grossly wrong, maybe not intentionally. But let’s say there is something wrong. So is your process robust enough or good enough to catch it? Before it leaves your premises? Right. Alright. So it should be good enough. And that’s what they check. So, |
| Shubham Agarwal : | on the face of it, it looks really important because these drugs are being consumed by people like you and me. Yes, yes. Now, the way you put it, it feels that you know, yeah, it is important and they should really make a big deal out of this, right. So but, you know, by experience, the pharma companies must have become really adept at reworking on the identified problems, you know, isn’t it even though it takes a hit on the efficacy, it is a fact of life for them you know, like, like we discussed that these problems can happen The only important thing is that the process should be robust enough to identify and detect these Yeah, but is there a long term repercussion of these warning letters as well? |
| Abhinav Shrivastava : | Yeah, sure. So, one way of looking into these repercussions is what we just discussed that it may lead to your supply is getting restricted manufacturers not allowed to launch new products and sometimes in extreme cases, plants being you know, asked to stop production as well these are there, but the other way to look is you know, what does come what are the companies investing in these you know, over a period of time, so, huge investment has gone you know, into adding a lot of manpower, into their skills, trainings, various sources, you know, space infrastructure and all the fancy instruments or equipments. So, these are things which, you know, the companies are being asked to invest or you know, put in there as a necessary condition. So, to have minimal errors as possible, but only thing is they do not, you know, give that return, or of expected return of reduction of errors, so, that’s the problem for them. |
| Shubham Agarwal : | I think it is a multiple multiplier kind of effect that impacts almost everything in the organisation. | Abhinav Shrivastava : | Yeah. So, so, if you want to just summarise the problem for them as their future sales or address, you know, and current resources are always shown as gaps. So, either it is in terms of man skill, or you know, resources, we always want something to give more improvement, or you know, to reduce errors from the current to, you know, the desired levels. And whatever corrective and preventive action they do, or they, whatever they are taking, it adds more and more burden to the existing capacity and forms of SOPs checklist COPPA so, so it adds a lot of pressure on their existing capacity. And then again, they have to come out of his own way to come out of that as adding more resources. They’re caught into this vicious loop. And then no significant reduction of human errors is what they feel problem, a big problem for them. |
| Shubham Agarwal : | this is serious, this is serious stuff then. Yeah, actually, right. So, if the problem is so, rampant and if this is such high concern, you know, the brand equity the company’s financial the launch of new products in the future is all at stake sort of at stake, then the companies and the industry in general must have come together and found a found a way as to why this is happening right and probably work on at least those occurrences which are very common or are happening again and again. So, have they been able to make any progress on that front |
| Abhinav Shrivastava : | Yes, they have, like they have made a lot of progress over the years. So see, it’s an outcome of continuous improvement on the current processes. So they they come from level, you know, where a lot of basic steps are missing, you know, simple STPs were not clear. They were transcription mistakes, simple, smooth transcription mistakes, typographical errors, you know, the lab practices, were not that clear, okay. And they’ve come very far from that, and to a level where at least some basic necessary conditions are they are not failing, put in place. But still, when there are some errors, which are recurring, which are related mostly to human errors, that’s not there, you know, in a finding to resolve that, and they’re not able to show the next level of improvements. They’re stuck. And right now, they’re not able to see the visible pattern in this. So they’re able to quickly able to find out, you know, what are the major contributors, but what if the, there are no contributors, so major contributors, so, if they appear, even if they are appearing, they do not have the interlinkages which makes the Pareto Analysis not very helpful. |
| Shubham Agarwal : | So you mean the 20% could actually be the real reason why the 80% is happening. |
| Abhinav Shrivastava : | Yeah, yeah, so, you’re just looking at a very small point and that is also revealing a huge thing. Now, which can affect a lot of problems to you, you can spend a lot of time in fixing 100 problems, but you just see two or three and you can find you know, maybe resolve all the all the problems that you are observing Oh, interesting. And the second part what you just asked see, if you see the how do they do the analysis once they know okay this system is not giving me the result you know, as per the standards, the post facto all these analysis post facto right? So then they start discovering okay, what went wrong? Which step did not work as per the defined document? Yeah Is it the solution problem is it the man problem or is it you know, simple mistake in instrument I have maybe something an environment or maybe something is contaminate. So, so, there can be many possibilities you know, which can lead to this error and you do not know it is not very clear. So, what do they do? They try to invest time and capacity to do small iterative test to just to locate just to isolate where the problem could have come from, once they do Okay, okay, this is the area then. Okay, you have to understand that a lot of effort has gone into this right to reach up to this stage. Now, once they reach to this stage, what happens is they can see okay, now what could explain this simple mistake, human? Right, so it’s now the energy does not go beyond this, you know, asking the further level of reasoning of why this human made a mistake. But okay, it’s good to you know, it’s very easy to just put the label there and move past it. Yeah. Or the other way or the other way, some, some ways of, you know, resolving it, you find five possibilities, you fix all, you put all the SOPs in every possibility areas, you have five checklists, you have 10 more new SOPs, you have global CAPAs means CAPAs is corrective and preventive actions. Yeah. So you implement everything something might work. |
| Shubham Agarwal : | But that’s probably, you know, bringing in even more variables to probably give way to even more problems in the later stages. |
| Abhinav Shrivastava : | Yes, yes. And that’s how that’s where the problem starts to people start realising Okay, some things have been solved, some things have not. And over the period of time, they realise, okay, I’ve invested a lot of capacity into it, you have now used then you start questioning your own method of, you know, analysis. Yeah. Yeah. And FDA, as you were, you know, looking at, you know, all these warning letters and observations, when they approach okay, they see errors is happening, fine. But what is your way of investigation? Is it a thorough investigation? Or are you just labelling it some kind of an error and not even spending time or energy to deep dive into it? Okay, what went wrong? And is this the right root cause or not? Right, is that corrective action and preventive action good enough, justified? So they also want the manufacturer or the quality lab to do a thorough analysis, a scientific approach towards, you know, solving these problems? Right. So, yeah, so that’s their approach, overall, |
| Shubham Agarwal : | the method of investigation itself could be questioned at times, right, isn’t it? |
| Abhinav Shrivastava : | Yes. Yes. So let’s say it will not happen that rampantly in you know, big pharma companies, because they have been refined over the years. So they have they have invested a lot in, you know, adopting those methodologies, investing in people, training them, you know, segregating the entire workforce, just for the investigation. So they have invested a lot, maybe in other companies, you may find such mistakes, also, maybe not even a simple process of investigation. But a lot of big pharma’s, if you see, they have at least reached to a stage where they start doing those, you know, way of investigation, but even for them, they are not saved from the two vicious loop that we’re talking about, you know, adding a lot of capacity just to solve some problem, what is that we do not know? So, at least some problems should get solved. That is the problem. |
| Shubham Agarwal : | Right. So, you know, you mentioned the human error again and again, now, at counterpoint and at Vector Consulting Group, we take the human error at the last we, you know, that’s one of our core principles that you know, the inherent goodness of people are never wrong, essentially, right. And any company attributing humans as the core reason for the error might not be right. But is it really the root case here in these cases and pharma companies? Or is it is it again, for the lack of a better reasoning that we attribute the errors to humans? |
| Abhinav Shrivastava : | See, people that all the processes are run by humans so even if they say it’s all human error, it’s from some point standpoint, it’s right kind of a paradox or making mistake. But yeah, when you say okay humans are not making so they do not get up in the morning and come to the plant and you know, thinking okay, I will Today I will make a mistake with this mistakes. Yeah. So what is leading them to make these mistakes is what you know, we try to understand, okay, and that’s where we try to Yeah, that’s why we try to find okay. I’m sure something else is also working at play as the that human or that analyst is trying to work his way through the shift. So we try to ask those questions, you know, just to understand, Okay, why did this analyst made a mistake? Right? and not someone else? Is he doing the same this product for the first time? Or this test for the first time? Or the step is very new to him? Is he experienced analyst or very fresh around a new guy who’s just doing the lab? Let’s see if there is any yes to it. Then a similar analyst who did the same product by dint he made a mistake. Yeah. So, when you ask these questions, you of course, these questions are very complicated, and you do not get a pattern. Because if let’s say, someone says, going to my lab, you know, these new guys come in, and they make a mistake, they don’t even understand STP. And then you say, Okay, can you do you have a pattern? Then they say, okay, you know, sometimes my experience analyst doing a make me making a similar kind of mistake, then how do we how do I explain these two phenomena’s? So there is no clear pattern to it. But then we try to come in and we say, Okay, why do we also make a mistake in our lives, right? We try to avoid these mistakes, we’re not talking about some deliberate mistakes, we’re talking about simple misses and skips, right, I wanted to do this step, but somehow I missed, maybe I would want you to write this in in a, you know, in a, while I’m doing a calculation, I had to do this step, but I somehow I it just skip, you know, I just skip those steps. So these things can happen. And that’s where we try to understand, Okay, why in our lives we make a mistake is it sometimes we are in rush, right, you have three hours of time, and we have a lot of work to do in that amount of time, you try to you try to rush through it, and you try to make a mistake, sometimes we also, you know, do a typographical mistake, while rushing. And this, this creates an environment where I try to complete my work, okay. And in doing so I try to switch between a lot of things in between and remind, let me remind you, these are a lot of complicated tasks, you have to learn, you have to do a lot of entries in in a booklet while you’re doing it happens concurrently, while you’re doing the test, you have to make a lot of solutions, you have to prepare a lot of solutions, you have to search and you know, in short, all the consumables are also there, you have to ensure that the instruments are right, working in right condition, all that you have to ensure in just eight hours, so and this is all happening parallelly. Yeah, this is all happening. Something’s happening in series. Something’s happening in parallelly. And while you do this, you’re, you’re this we use that word cognitive overload. Yeah, so this cognitive ability to handle all these tasks keeps on |
| Shubham Agarwal : | getting overloaded, but one could say that if it’s just the cognitive overload, you add more people to the same process or the same step. And that way, you can distribute the work among probably two or three people and then get the work done. Easy. |
| Abhinav Shrivastava : | Yes. So, yeah, so we can we can see, let’s say, I added two or three people just to support them, right. But you do not know which guy requires the support right now. So at the start of the test, you do not know that, so what eventually would happen is if let’s say you have 10 analysts in the lab, you would require 10 additional people just to support them, and how do you do that? So, so it requires a lot of capacity again, so is there a way out which, would not, you know, deal with a lot of addition of capacity in forms of steps, checklists, physical checks, all these SOPs and also help me in you know, reducing the errors, if that is what we can, you know, find it’s, it’s a win for everyone, you don’t have errors, you don’t invest in a lot of capacity, |
| Shubham Agarwal : | right? Just for the sake of our listeners, could you simplify and tell us probably in just a one liner as to what really is the concrete reasoning or the core reason that we at Vector have identified with these problems or these concerns? Okay, |
| Abhinav Shrivastava : | so, I will not do it in one line. Okay, so, yeah, so, actually, if you see there are two types of errors one is type one and type two. So type one is very simple. Okay, I have I know the step and I have somehow missed it. Okay, okay. It’s not that it’s a knowledge problem. And for that simple answer is yes. The it’s a rush it’s switching it’s because of these things I am I am able to I’m subjected to such an environment where I’m able to make a mistake or where I’m where some mistakes happen, you know, in the due process, why those rush and switching elements are there we can discuss maybe at some data point Sure, but this is what the elements are a key ingredient is to make a mistake, right. The second part which we were talking about type two, okay. These are the mistakes which are Today is you know, because of a knowledge gap so, I do not know okay this product has a different behaviour to it okay. Now, now it is not it is not mentioned clearly in the STP sometimes it is not known also for the lab that this product has this kind of a behaviour and this error is also can arise because of certain steps which was not defined in a in a clear manner. So, for example 10 minutes of sonication is required, it might not be adequate for this particular process. So, in this environment it might be required 20 minutes although the entire process takes care that such kind of cases do not occur in the lab in the in the in the commercial labs, but yes, the time and multiple iterations of batches such kind of mysteries do also appear. So, what we have to do is we have to segregate the type one has a simple question or simple cause, okay, I have missed it just because of rush or switching, I have done that, yeah, type two is very serious, okay, I have to find out, I can’t blame the analyst or something else, or the instrument or the some behaviour, I unless I know that product behaviour, I can’t. And the problem is, today, I can’t segregate between type one and type two. So that’s, that’s an added problem for the |
| Shubham Agarwal : | but why is that?
because in a in a in a chaotic world, where you have 100 errors, everyone will always try to blame product or a machine or an environment, you do not know whether it is actually a human problem or a product problem. Right. And, and the investigation teams keeps going round and round in circles to you know, sometimes it’s a product problem, then they try to do a lot of testing product, they don’t find anything, then they come back, even if they find something they fix it, but actually if it is not a problem. The problems reoccurred. |
| Shubham Agarwal : | Right. So that’s even a bigger problem, I would say because, yeah, the you know, disability of identifying if it’s a type one or type two problem is a bigger problem in my head at least to solve first. |
| Abhinav Shrivastava : | Yeah, so, two simple problems, but in combination no one knows which is |
| Shubham Agarwal : | right. So, then my obvious question is, you know, we have solved these problems for a lot of our clients in the pharma industry, what is the approach what is the direction that we take with some of these clients to solve this? |
| Abhinav Shrivastava : | So, okay, so, as we were talking about type one and type two errors, so, we try to see that unless Okay, unless we try to bring the noise or this kind of chaos down okay. And we try to resolve those systemic problems till the time we will not be able to pinpoint actually whether it is a product problem or a behavioural problem or a training need, which is the gap or something else right. So, what we try to do is we try to solve type one problem first and then move once the noise is down significant type one errors goes out of the lab, those things are not happening now, you can see a clear pattern which can you know, which you can take to resolve those type two errors right, right. So, this is the broad approach that we take. Now, if you talk about, how do we solve type one errors okay rush and all these skips and switching cognitive overload that yeah, the cognitive overload problem. So, now, then we ask ourselves, okay, why this lab is subjected to these rush and you know, these kinds of environments, then we see okay, there are two possibilities one is an external and one is internal, external would mean because lab is not working in isolation, right, they are working with procurement alongside the procurement and manufacturing, procurement is bringing in the material asking the QC to work in an in a defined SLA to test it and keep it in the store warehouses, the manufacturing is behind them to get those tested materials which are actually required for dispensing or you know, manufacturing. So, the QC is always stuck in between whether to should I service this SLA need or should I service this manufacturing need, right. So this is the conflict, right, two actions. And this, this manifests somehow in, you know, in the way in which they are planning their campaigns, they are their way of planning the analyst when they come in the morning. So, today the analyst just as an example today the analyst comes in the morning, and what he has to do, let’s say it is been told or it has been displayed in a in a board moment he starts doing the work, if let’s say something urgent comes in, he has to stop that he has to start something else. It’s
Shubham Agarwal Abhinav Shrivastava |
| Shubham Agarwal : | You’re the entire plan gets rolled over and your plan goes for a toss. |
| Abhinav Shrivastava : | Yes, yes. So as the day progresses, you start seeing such kind of variability is hitting your plan. And half a day past you see your clock, it is three o’clock, and you’re still two hours behind schedule. So the moment of Rush moment of urgency starts coming in. And that’s a it’s a very good environment to make mistakes at that point of time. |
| Shubham Agarwal : | So Abhinav I am assuming that we can’t really control these variables, because you know, something expanding the time expanding is something which is beyond anyone’s control, I guess, right? That’s, that’s just Murphy hitting the system. So, I can’t solve that, but what is it that I solve in these cases in these implementations that we do that really make the difference? |
| Abhinav Shrivastava : | Yeah. So, we try to ensure that these variabilities do not hit that analysis okay. So, for example, all these variabilities can be so, first is can we decouple the lab okay from the external part, can we cushion it in such a manner that okay even if those emergencies appear in some manner, it does not reflect an immediate change in the plans for the planner particular for the particular day or for the next day, right. So, if you if you kind of decouple your lab to from your procurement or your supply variations and your demand, you know, if that is from your manufacturing, then you are working peacefully at least right you can make your optimal campaigns you can make you can define you can plan your maybe those necessary conditions or you know, full kits, well ahead in time and then there will be no surprises right. So, that is one part. And the second part is all the internal challenges. So, so, all those things which can take away or you know, the additional time from the analysis, days’ work, can we can we reduce it, okay, can we ensure that someone else takes care of those things, while the analyst whenever he comes or approaches to a particular desk, he finds all those things there, you know, with his desk, or at his desk, so as the prerequisites are in place, |
| Abhinav Shrivastava : | yeah, there’s all those prerequisites are there in place, and that’s what we try to try to see that these things are in place and of course, when the need for the switching comes, because this may take away the rush element or the time element from the entire day, but there might be still some switching which you know, which can still be there Yeah. So, can we take away that also, because if you are still subjected to switch between two kinds of tasks, then you will always try to you know, remember something then go to somewhere someplace else and then that work may take more time than you have to come back to your existing setup. So can we take away that part also. So we do not want any switching any rush to for the analyst to for the entire part of the day? At least when he’s working on those critical aspects of you know, solution preparation, all those which can, you know, end up in making those invalid OSS, right, those make mistakes. |
| Shubham Agarwal : | Great, I think easier said than done, but I think that’s where our forte lies, as Vector Consulting, that’s what we really bring to the table. What kind of improvements have we observed across our clients with some of these interventions some of these solutions that we have implemented. |
| Abhinav Shrivastava : | Yeah, so, some of these implementations have actually started as a struggle but eventually when you when you go past all this you know you make such kind of change across the lab you see a tremendous increase oh sorry the benefits coming in with respect to the type1 errors first you see the dip in type 1 errors so a lab which is doing let’s say 10, you know 10 kinds so there are two kinds of errors, one is OSS which are invalid OSS and one are lab incidents so these are the lower forms of errors or you know the simple mistakes which can get acknowledged and does not have impact on the sample or you know bigger sample results so if I just measure these two, so 10 invalid OSS have reduced to just 3 or 4, wow, in in a in a quarter, okay and in the incidents lab incidents which are which are which you used to happen 150 or 200 that has gone now to a level of 60 or 70 per part so this amount of reduction, which is more than 50% is what you can see not just by adding the capacity and adding lot of CAPAs but by just simple two steps avoid rush elements to come into the lab or to into the day’s work of the analyst and ensure that he does not spend time in switching between you know set of activities critical activities just by doing this and then the second set of benefits come in you know where in you start attacking your type 2 errors then you start seeing the holistic way of you know looking both type 1 and type 2 so you see clear patterns now ok this products is making mistakes every time in all the labs then you can start creating those development long term development activities either to you know improve your documents either to improve your processes and that becomes very clear for the organisation to start investing time and money and resources just on those problems its very clear for them now that’s the second benefit for them. |
| Shubham Agarwal : | Great. I think you know great points that for the fact that we could simplify and clearly distinguish that this is type 1 error which is predominantly because of cognitive overload and there are type 2 errors which are predominantly because of probably skill gap or knowledge gap and typically these two errors are interrelated and we could distil them down and work on each piece and then in the second phase work on them together as well has shown some considerable results in the projects, right. I think that’s a great piece of content for all our listeners and lot of food for thought I think to simple things like switching losses could make so much of an impact. Right. So thanks a lot Abhinav for that discussion, anything you want to add or conclude with. |
| Shubham Agarwal : | I’m great. Thank you so much for being here. |
| Abhinav Shrivastava : | Thanks for this opportunity. |
| Shubham Agarwal : | Okay great. Thanks a lot for giving us time Abhinav, it was great discussing with you, I am sure the listeners would love it. For all the listeners any questions on this topic if you want to go deep dive on this topic ask something understand something further please write to us the details of this podcast of this episode has all the links.
Until next time we’ll keep bringing such interesting topics and episodes for you. Thank you, this is Shubham signing off. Bye. |
Comments
Related Information
Get in touch
Registered Office:
Vector Management Consulting Pvt. Ltd.
10th floor, Thane One, DIL Complex,
Ghodbunder Road, Majiwada,
Thane (West), Maharashtra - 400610, India.
Vector Management Consulting Pvt. Ltd.
10th floor, Thane One, DIL Complex,
Ghodbunder Road, Majiwada,
Thane (West), Maharashtra - 400610, India.