Using AI to Improve Clinical Development

Automatic TRANSCRIPT

The california technology council's new c._d._c. Benefits trust combines groups of emerging technology companies to offer large company benefits fits to small businesses. This approach delivers employee benefit programs with better choices and lower cost. What's included medical. Dental and vision options are available with additional employer and employee online resources to support simplified enrollment and ed administration to learn more go to california technology dot org forward slash join. That's california technology alleged dot org forward slash join. I'm daniel levin and this is the bio report <music> <music> artificial intelligence is all the rage drug discovery today but there are efforts to apply the technology to other aspects of the drug development continuum to improve decision making and increase efficiency one such example is fi cy which has developed an artificial intelligence platform to improve improve clinical site selection of valuate protocol designs and patient inclusion criteria and manage the implementation of trials. We spoke to gently gently founder and president of fee cy about the company's a._i. Platform how it works and the case for the use of artificial intelligence as a tool tool to improve drug development jen. Thanks for joining us. Thank you dan for having me. We're going to talk about why it's artificial intelligence platform and how you're using it to help clients make better decisions decisions around the clinical development of experimental therapies. Perhaps you can begin with a description of and what it does who its customer is and how it makes money. <hes> says i was created twelve years ago. <hes> we are a data driven can communicate development organization we <hes> leverage our database <hes> which is big data <hes> and also our statistical artistic model and outward and those are artificial intelligence to help our cons to <hes> more objectively planning and executing clinical trials it was <hes> a created <hes> based on the situation that <hes> despite communicative elephant as an industry was spent ninety one hundred billion dollars. I year <hes> level the company next. The of the operations management was never a proper <hes> handled the simpleton. We we are seeing he either. Inside industry or external industry are those <hes> prolonged your cycle time <hes> blooming projects <hes> and and so never <hes> easy to come on pre trial originally planned those are the simple send sympathies being treatment by the fact that we really have little understanding think about <hes> how those <hes> clinical development process <hes> from the operational management poem view being structured and implemented so the fast i was created at the to reveal those patent and those fundamental structures driving those deliverables from the clinical development in point of view <hes> from the <hes> so so that's where the way can <hes> fess is working decline spending fading us from several different ways i fall we are helping them to looking at hey aurenche maly <hes> mystified the picture of planning economic trial <hes> to a more structured took the way so for example the always conflict between the clinical trial teams and the senate and management in terms how long picking a clinical trial to complete that conflict also happens between the the the pharmaceutical industry companies and so the result <hes> all fit. It's usually descend your management wanted to have an clinical trial to come to complete much faster <hes> to <hes> getting a job to be approved with the much quicker and this year owl's <hes> not <hes> <hes> understandably often one into have more sites and they are expecting too much longer time which defense i platform we we're allowed them looking at it the picture in much objective fashion that way we'll be able to <hes> implement it at a <hes> at its plant. The clinical development of drug involves many decisions that a a drug developer must consider along the way this includes lose things such as education assessments protocol evaluation clinical trial site selection trial management. How are these things generally done by drug developer's today do they rely on existing relationships recommendations gut feelings or is this a highly data driven process. Unfortunately the it's anything but driven data driven process a highly highly relies on the level of the experience of those individuals in those <hes> he being assigned with those tasks asks for example a medical professional also play a very important role of defining had talked <hes> is so cutic- product product profile and development and other elements of development plan withing a protocol <hes> in etc so so what so that's where the fundamental reasons of <hes> a lot of uncertainty <hes> commun- from and also a all of mistakes been embedded in those because those are really much empirical <hes> derived <hes> solutions not a data driven solutions. What's the consequence of these decisions. How much is at stake and what can the impact on the costume duration and success of clinical trial. Be it impacting in in many ways <hes> a connecticut trial in in today's <hes> setting most of the times of for example phase three clinical trial that's based on the results of face to clinical trial aw by anaemia extrapolation and as a result of it <hes> because the relationship it's not actually near <hes> relationship from our <hes> understanding <hes> but if you to use them ne-near extrapolation at today's a practical practice practice today the result is that <hes> face reconnect trou- often much less much longer than original plan and and that means it's gonna be must must much more costly <hes> from the financial point of view but it's also going to delay <hes> the <hes> the time for patients to get in those much needed. Medicine feces has an artificial intelligence platform. I suspect it's value now. You can only be as good as the data and has what data does it include and where do you get the data from. That's a very good question. We are according ourself <hes> a big <hes> a big data driven and artificial intelligence in naples able to platform because we are collecting data from all four corners of the internet and be there already examples like communica- trial started off <hes> which is sanctioned by the u._s. Government then you have similar chuck registries all over the world in europe in japan in china in <hes> in many places so those are the type of the source of the data. We are getting from. We also getting data from <hes> many of communicable child sides the the posting their information in different places and <hes> like <hes> also the academic conferences meeting extract and share the information getting information from those places as well and there are some of the less known <hes> sources of information <hes> we are also getting <hes> from <hes> by the they are a passing the majority of the data. We're getting off from the public domain but we also <hes> purchasing southern data sets <hes> from <hes> different places. The idea of our perform is not necessarily the monarch data we are having with the key to our <hes> power of the pratt from lachey residing all capabilities to <hes> structure structure the data in a meaningful way and cost checking data from different sources <hes> with each other to making it can sure the quality are reliable and consistent and and <hes> in in all of those spaces are approach are absolutely <hes> you know vegas and and and <hes> the most the forefront of the artificial intelligence and the imperil was state extra extracting and struck and the process in a structured process is the interpretation of the data and those are having equally as much innovation in our platform and those are the <hes> the innovation are being in <hes> patent that is so serious of <hes> of patterns in the u._s. and the rest of the world. How does one go about using the system. We are <hes> having to <hes> mechanisms of our clients to you died in this system <hes> traditionally we have been a service based <hes> organization we are engaging our clients and <hes> <hes> in in in a human to human interaction and getting the understanding the projects and giving input to various elements from the canoe you integrate economic development process to <hes> to communicate the indication assessment to particle planning to <hes> the the community trial implementation so along the time we figured it out the need to product size some l. dollars services to make an easier for constant access and that's where we have developed a canonical <hes> investigator data site selection pat from we call the clean side. That's a platform. It's assault software for <hes> for a two services services <hes> so in that platform or clients can <hes> going into any of the computer with internet access access to search <hes> the investigator sites according to what they need and that's a very friendly interface ace and in many of our concord like hookah like <hes> fashion to to to to to work and and we're hoping with doubt <hes> coincide will allow many of more for all kinds of benefit hall platform. I'm hoping you can walk. Talk me through the different ways the systems used. Tell me what date is available. What customers using the system to figure out what let's start with the the development planning indication assessment we are working with our clients <hes> based based on the their product <hes> they're actually communica candidate profile and <hes> in checking <hes> the <hes> the view <hes> data enabled <hes> <hes> fashion. Let me give you one example. One of aw clients were in a mechanism of <hes> lowering <hes> dealing with iron overload patients so data type the patients having <hes> a disease called thalassemia so their original plan was to looking at <hes> a single indication and forecast the peak sales to be <hes> <hes> to be he <hes> about six hundred million dollars and when we being caught in with started looking to i'll over no the magnuson we recognize that that's that's a cluster of diseases or share similar mechanism so you said of talking only thalassemia and we're asking them. I'm actually targeting a group of indication by doing so they're they're orange. No projection of the peak sales rat race the from <hes> six hundred million dollars a year to three point two billion dollars a year so that's the type of the ways we help them injecting data driven perspectives to allow them council <hes> comfortably and confidently making solutions to maximize the value of their product or product development. How about protocol cessna talk about <hes> a specimen that it's a fascinating area and it's we are pushing out a site selection solution called the clinton side. Many of the enrollment related the challenges are knocked cost the by m._s._a. Decide to performance rather they were actually being caused by the inadequate design of the particles in the i pat phone we are in in the process of developing a similar product <hes> he in in that area which we're looking at it historically <hes> studied <hes> clinical trials in any of those indications we currently have over three hundred thousand products in our database based which allow us looking at the detail the patent in all of the design elements associated with a with a particular disease z's then allow us to build a modal value which is statistical time <hes> really means the most frequently used the value and and the using that frame structure to allow clients start at the enchanting innovative ideas to change and the modify those value but the benefit and taro fest. I platform to allow our clients not only just having an on could a structured structure to make those chance but also having the ability to understand that the operation impact because of those changes is being made and for example if you shanking the patient the popular <hes> population by restricting their age old old by the severity of their disease simpleton <hes> you're you're starting to looking the increase of the number of investigators sites needed to for their trial and the enrollment cycle times becomes longer those are obvious reasons but our systems will allow clients to making much more sophisticated assessment and decisions based are out to phone how about when it comes to the selection of sites for clinical trial all sides selection <hes> solution. I it's real solution. It's a very competitive space. <hes> from site selection dentists point a point of view but the generally speaking many of those <hes> pat forms are providing and they are not solution are how to form allow our clients to looking at a particular disease indication occasion then we will <hes> guiding them to the country distribution of those qualified sites around the world and then from this country picture we allow constant dive into specific country or a specific set of a set of for counties then <hes> we allow them to <hes> we actually assign a set of the school <hes> allow our constant make uh-huh decisions on those of ascites the needed before there can initiate trial then the <hes> coincide walk go no further to <hes> to provide the detail the profile for each of those investigator sites including from <hes> some of the men mechanic <hes> <hes> elements of the information the address and email and telephone etc too too much more soft could description for example historically. What did the trials to have run and what's the capacity they have you gotta sites to round southern child's and as well as their performance in running those clinical trials then <hes> go the further we started looking at it their medical knowledge profile so if they are a <hes> on college est <hes> they oncologist colonist specialized in particular type of cancer or a set of cancer if it's a set of cancer where other to exp expertise area resides so and we also looking at the experience in working with different types of <hes> <hes> jacker candidates is that small monarchial or vaccine or antibody and we will be able to making those those <hes> those experience experience visible in a quantitative and objective fashion so those are all of these things are being built the way based on massive amount of data rounding behind in real time and allow them to making the decision <hes> based on <unk> collective view we provide by this platform of the last one to ask you about was trial tation station management. What what does this including. How does the system do this. Could you repeat that again. Sure the last one i wanted to ask. Ask you about was trial implementation management. What exactly does the system do in this regard. Oh <hes> that it's so the travel implementation management. It's not a part of the coincide but it's a separate set of capability in the <hes> fence. I <hes> performs. The meaning of it is is that you can imagine the <hes> <hes> the the implementation of a clinical trial particularly a museum large scale of clinical trials can and be a very complicated in the sauce tecate it <hes> system system engineering projects so what we are allowing in <hes> all capability allow or clients to do is to looking at the space in a structured review for example <hes> who are potential high <hes> <hes> potential high enrolling sites heights and who are not but not just looking at a static picture but looking at in in more than amick away so so we incorporating semak muezzins to tie our ties those the size and the supporting them was different type the resources to allow two sides with the best potential to enroll patients get the most support that away will will be able to <hes> save money and resources and also what's the potential to shorten the cycle time and those saw all who we have patent the method to support that implementation process you mentioned a stunning example early on about the the patient <hes> the company that was developing a treatment for iron overload. Is there anything more generally you can say of. The pay off has been for people have used the system as case studies two point two. Are there actual savings or hard time savings that have been realized <hes> there are three major benefits from <hes> the the <hes> from our clients perspective. One benefit is just to gain the kind of <hes> visibility to the outcomes of their plan and which will be a very much improved the communication effectiveness evenness between the sponsor of <hes> of the the trucker candidate and in the seal and bettine senior management and a specific <hes> project team so that's one type of the the the benefit another type of the benefit is to allow the <hes> the the picture to be objective for assessed then to <hes> pull with different levers to <hes> to do either offered him is the woman the cycle time <hes> all looking at the most cost after effective way to implement the trial in some incidences we would be able to looking at a particular economic trial kyle just the from the design point of view. We come protected. The actually cannot be successfully implemented and one example four of so. It's a t._v. h._d. Which is a horrible disease <hes> associated with <hes> with transplantation and <hes> one type of intervention one. It's <hes> type of intervention. <hes> it's the treatment of those patients another one. It's the prevention of the patients from happening that these it looks like very similar <hes> trial by the by i using our product form we indicated that the prevention trial it's actually much more costly <hes> <hes> physical than betrayed them in the child. It's the matter of fact our model or clients. The <hes> betrayed treat him in the trial will take three times small decides and the words even long on the side cycle time to implement and the based on our assessment all clients actually <hes> basically they can't the trade them in the trial and focusing hang on the prevention trial by doing so <hes> the chief dissimilar marketing objective <hes> by the saved <hes> likely about thirty to forty million dollars and that's just one topic example another example is at that time i'm a rare disease in cancer world and our clients got a very promising clinical <hes> can develop in the candidate so the sentence senator management so eager to putting this <hes> candidate to to to patient so the <hes> the basically padding the portuguese team say tell us how much money you need to do this but i want you to be able to complete the trial in two thousand a month so so the trial team somehow understand that this is not to oval but did they don't know how to express it so we look at all aspects of that can nicotrol by looking at it the competition by looking at the business in this process which decided activation by looking at the decided performance by also looking at the product design we presented our our assessment management and all and in that <hes> assessment we concluded that trial cannot be completed could eat the shorter than thirty months so by the because of the <hes> reliability and objectivity our assessment the to send your management create with our print <hes> cycle time so at the end instead of everybody forty in the frustration fission <hes> and and <hes> and many of the other uncertainties associated the trial was successfully implement exact as we predicted addicted and the rockets actually now provide <hes> approved and also approved the medicine to the patients in that indication. There's there's a lot of discussion about the use of a._i. And discovery and in the clinic less oh about applying to clinical development and you see this changing or proving the clinical development process. It's a <hes> it's a <hes> at the complexity in the communica could develop <hes> comparing with <hes> with the discovery space. It's the it's the patience patience it's the the center in terms of our business but they are also fundamental reason of causing so much of the complexities associated with connecticut development element so we are chew believer that with the amount of data now we are having <hes> and with the soft occasion patient of all our capability of in interpreting data we are in a very good position and to allow <hes> artificial intelligence to be widely implemented and applied in various of communicate development scenarios scenarios <hes> we just talk about the clinical integrated community development planning. We're talking about polical design. We talk about vowed that the <hes> the site selection with we'll also talk about the clinical trial implementation management but i wanted to also should the excitement is is that we also accumulate over thirty million of patient data well-structured and and and and and derived the form femme does controlled clinical trials data will be the type of data. We are looking forward to us to to <hes> to do like synthetic on all even <hes> you know similar certain type so the community charles <hes> we are so much excited to see those other things being put into practice jann lee founder and president of jenn. Thanks so much for your time today. Thank you very much for having me <music>. Thanks for listening. The buyer report is a production of the levin media group automatically downloaded of this podcast each week subscribe to our r._s._s. feed or through itunes or other podcasts manager join our mailing list go to levin media group dot com com. We'd love to hear from you. If you want to drop us a line or interested in sponsoring this podcast send email to danny at levin media group dot com special facts joel levine who composed are themes and jonathan levin collective which performs um <music> <music> <music>.

Coming up next