ARTIFICIAL INTELLIGENCE SHOWCASE Driving Drug Innovation With AI The global AI market has been showing rapid growth as major pharmaceutical and medical technology companies invest heavily in internal and external AI partnerships.

ccording to some reports, more than key objective is to predict which patients are 50% of executives expect broad scale at risk of cardiovascular disease and then de- How Pharma Can Use A AI adoption by 2025. Further, reve- termine the treatments that might work best. nue generated through AI-based solutions in Scientists at Janssen Research & Devel- AI for Drug Innovation the industry is projected to rise at a CAGR opment have found a way to use AI to speed Create internal expertise and of 21.94% and reach $2.199 billion by 2022, the process of discovering new potential treat- according to Frost & Sullivan. Meanwhile a ments. They are using AI to sort through support AI employees with the right report by Global Market Insights forecasts how cells representing certain diseases react resources that the U.S. healthcare AI market will exceed to a variety of compounds. Then, through Collaborate with start-ups and $10 billion by 2024. With total investment collaborations with academic partners, they others using AI in drug discovery exceeding $7.2 billion across 300-plus deals are using computer algorithms to predict how Partner with academic groups that between 2013 and 2018, the pharmaceutical other cells are likely to react to those same focus on AI R&D industry continues to lead the healthcare sector compounds. The benefit for scientists is they in terms of attracting AI-related VC funding. don’t have to start over for every study, which Consider open science projects or The COVID-19 pandemic has propelled means they can potentially get treatments to various R&D challenges growth in AI investments with a new wave of patients faster. Source: A Guide to AI in the Pharmaceutical Industry, transformative technologies being developed Many other companies are partnering with Stefani Group in an effort to contain the pandemic, such as AI leaders to drive innovation. For example, smart machines and robots. AI is also being Pfizer formed a partnership with Atomwise put to use to determine the effectiveness of to help identify potential new drug targets by nology to find treatments that will mitigate existing drugs in the treatment of the virus. using deep neural networks toPharmaVOICE quickly analyze disease complications that arise during and millions of molecules. This further expands after infection. Its Project Prodigy platform Pfizer’s AI-led initiatives, which have included maps all known drugs against every possible Transformation Through AI partnerships with IBM Watson and invest- mechanism in which the COVID-19 virus ment in MIT’s Machineof Learning for Pharma- operates and causes complications, and seeks to Many pharmaceutical companies see huge ceutical Discovery and Synthesis Consortium. build scenarios to identify non-obvious medi- opportunities to transform their operations Other pharmaceutical companies have cal solutions as well as non-linear risks. with AI capabilities, from productivity im- formed partnerships with Atomwise to drive provements to better outcomes across the drug discovery, including Merck, Bayer, and value chain to using AI to create new offerings. AbbVie. In addition, AbbVie is working with Challenges to AI Innovation AI is top of mind for GSK, which an- AiCure to monitor medication adherence in nounced plans to recruit 80 AI specialists by patients with schizophrenia. One of the biggest barriers to growth in the end of 2020. The company’s AI head- One small biotech company that has been AI capabilities is the limited number of AI ex- quarters will be in San Francisco, but experts putting AI into practice in its drug develop- perts. According to a report by Tencent, there will be spread across Europe and the United ment process is Verge Genomics. The com- are about 300,000 AI researchers and practi- States. The goal is to use AI in the search for pany uses similar processes to those that power tioners worldwide but the market demand is treatments in various therapeutic categories, Google’s search engines to map out hundreds for millions of roles. Other barriers include including cancer and autoimmune diseases. of genes that cause disease, and from there regulatory pathways. Since innovation must The company is also working with several AI find drugs that target all the genes at once. By be balanced with safety risks, there is a need platform developers to advance drug discovery. using AI, Verge is taking products into clini- for robust strategies to enable well-considered Another company that is investing heavily cal trials rapidly thanks to its approach. For ex- market surveillance for AI-led innovations. in AI is Bayer. Together with Merck, Bayer ample, it is expected to start a small molecule AI is only as good as the innovators behind has developed the CTEPH Pattern Recogni- ALS program in 2021, followed by four more it, with experts noting that to deliver results, tion AI Software to help radiologists recognize programs. It has achieved this by building a users must ask the right questions and work chronic thromboembolic pulmonary hyper- massive database of brain tissue sequences. with clean data. In Verge’s case, the com- tension, a rare form of hypertension. FDA has Another biotech company committed to pany started its work in human rather than granted breakthroughCompliments designation for the soft- AI is Healx, which has created an AI platform animal models, combining its brain sequence ware, which will be deployed through Bayer’s for rare diseases. The company’s purpose built database with human genetics, overlaid with Radimetrics platform. AI platform, Healnet, serves to fill crucial gaps human-derived neurons. This let researchers In addition, Bayer is using a combination in knowledge about rare diseases and speed the identify gene signatures that appear in pa- of real-world data and AI-driven solutions to discovery of new treatments. tients. improve the characterization and stratification Another AI-driven bioscience company, Despite the challenges, the importance of (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] of disease as well as patient populations. A Biovista, is applying augmented and AI tech- AI in drug development is now well-under-

24 October 2020 ● PharmaVOICE SHOWCASE ARTIFICIAL INTELLIGENCE

EXECUTIVE VIEWPOINTS

Rich Christie, M.D., Mike Weber On-Demand Reproducibility Ph.D. VP of Product Pharma companies work in a highly Chief Development Management regulated environment, and in addition Officer Aktana to using AI to drive data insights it AiCure is vital to ensure compliance with regulations, especially computer systems validation requirements. For this purpose AI Provides Patient Support It’s All About Balance traceability and reproducibility of data AI-powered solutions are playing a Implementing AI effectively is all about is key. Hence, being able to capture and pivotal role to keep patients engaged balance. The key is finding an AI partner reproduce on demand a trail of how in their trials, ensure visibility into their who excels at integrating all data an AI system predicted an outcome is well-being, and maintain communication assets, coordinating all stakeholders key. This places an additional onus on by providing multiple avenues of and leveraging existing technical and implementation and effective use of AI support. Sponsors of clinical trials and analytics investments. A well-rounded systems. sites need data mining and advanced AI implementation that can be deployed analytic capabilities to safeguard trials quickly, adopted effortlessly and can Jerome Premmereur, with meaningful insights and limited data generate practical suggestions and insights M.D. variability, especially when supporting for users allows the organization to focus VP, Patient Safety trials remotely. on agility, continuous improvement and Solutions and extracting the most value from their AI. Adjudication Michelle Marlborough PharmaVOICECovance Chief Product Officer Agility and Vigilance AiCure Production-grade AI models take work Time, Resources, and Quality to maintain. As markets shift, models AI is helping a lot in the equation of drift, especially in theof face of unexpected three variables that safety is dealing change, like COVID-19. Pharma companies with, wondering how we can get must establish a diligent practice of balance. The three variables are time, Enhancing the Patient Experience ingesting recent data, retraining AI models, resources, and quality. One cannot be Partnering with AI vendors that have and regularly evaluating their predictive perfect without sacrificing the other solutions that truly reach patients where health to ensure they haven’t degraded. two. Time is regulated and cannot be they are, and that use technology already Companies must also remember that as AI adjusted above the traditional reports at a patient’s disposal to enhance patient changes, so do the workflows of its users. of 7/15 days for expedited safety reports engagement will empower study teams This agility requires eternal vigilance, but (ESR). Resources are very substantial at sponsors, sites, and CROs to focus the competitive advantages that result can if the quality of the report is at the on the patient experience. Furthermore, be remarkable. level expected and on time. AI is fast. A this focus on partnering with technology learning machine on natural language vendors will also give study teams the Dinesh Kasthuril processing (NLP) can read hundreds of patient data visibility needed via a single Global Head, Client pages much faster than any physician dashboard to help them make data- Relationships and Project and can extract the elements that are driven decisions that can greatly impact Management important in a narrative for the ESR. the successful administration of an Covance Patient Safety Therefore, time, resources, and quality ongoing study.Compliments are able to achieve a new and better s

stood. In the future, some predict that chem- ally to try to predict the 3D shape of proteins While AI is just one element in the search ists who use AI to support their research will in the human body. The contest’s winner: for innovative new drugs to address unmet replace traditional ways of working. This was DeepMind, which made the predictions using health conditions, increasingly companies underscored during a 2018 protein folding neural networks — the same technology that large and small recognize that it is a crucial (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] competition, which attracted scientists glob- recognizes photos in Facebook posts. part of the equation.

PharmaVOICE ● October 2020 25 ARTIFICIAL INTELLIGENCE SHOWCASE

EXECUTIVE VIEWPOINTS

balance with AI. For these reasons, AI industry knowledge will be a force to be new technologies to replace real- has attracted a lot of attention with reckoned with. life interactions/decision making. regard to facilitating the capture of More importantly, patients need the adverse events and the traditional Max Divak to be comfortable with allowing pharmacovigilance operations within Interactive Group medical decisions to be made based the pharmaceutical and CRO industry. Supervisor on a program algorithm instead of This scope of the AI development is Ogilvy Health on thousands of years of medical now very active. In the near future, AI experience. is going to be very efficient on safety signal detection for clinical trial benefit Malai Sankarasubbu risk evaluation on an ongoing basis AI in Diagnostics Gaining Traction VP of AI Research — and certainly before the end of the The pharmaceutical industry employs Saama Technologies study even if the recruitment is very fast the use of AI across many categories, such as vaccination of a large cohort or including epidemic outbreak prediction, the seasonal aspect of allergy patient drug discovery, diagnostics, improving the recruitment — as well as postmarketing quality of treatment, optimizing medical safety for products on the market. cost, and even sales and marketing. The AI and Clinical Trial Data effectiveness of AI in diagnostics is quickly Management Claire Bonaci gaining traction. For example, diagnosing AI is yielding highly valued results Senior Director, US conditions through the analysisPharmaVOICE of imaging. for pharma as related to clinical trial Health and Life Sciences A trained algorithm can compare scans data management. Using AI models Microsoft against massive databases of imaging to predict discrepancies and manage

libraries and diagnoseof conditions within queries results in significant process seconds with a high level of confidence. improvement that reduces time to The AI could also cross reference patient database lock. Expediting conversion AI and Clinical Trials information to rule out certain criteria. of raw EDC data to analysis-ready With Covid-19 permeating every aspect Currently, the fate of a patient is not SDTM data in near real time is a of our lives, the pharma industry is left to autonomous machines. But when tremendous win for biopharma from a no exception. Digitization and AI are healthcare professionals use AI as their time perspective and can have positive uniquely positioned to be front and helping hand, or even as a second opinion, implications for clinical trial budgets as center to positively affect how clinical the effectiveness in diagnostics and quality well. trials are executed and how drugs come of treatment drastically improve. to market. This will not only benefit Walk, Then Run with AI biotech companies but also the many J.P. Maranzani Managing the pharma industry’s patients with unmet medical needs. VP, Account Group expectations of AI is a major Supervisor - Point of Care challenge. AI can facilitate remarkable Scaling AI Across Pharma Practice (EHR/ABM) achievements, but underlying business Cross-industry partnerships will be Ogilvy Health processes to support AI implementation crucial in the journey of implementing must be in place to ensure its success. AI at scale in pharma. Collaboration In order for AI to enable the industry between technologyCompliments companies, biotech Partnerships Are Key to AI to fly, it must first walk, crawl and then startups, and industry leaders have Implementation run with AI. This progression can be the opportunity to lean in and design There needs to be alignment and achieved in rapid succession with a tech ethical and innovative solutions using AI. partnerships with everyone. Obviously partner that can align internal business Leveraging technology companies’ AI “big tech” and pharma need to strengthen processes with the right state-of-the-art expertise and pharma companies’ deep their trust and relationships to allow AI automation platform. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected]

26 October 2020 ● PharmaVOICE Be thought-

PharmaVOICE provoking.of

Compliments Be here. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected]

THE FORUM FOR THE INDUSTRY EXECUTIVE ARTIFICIAL INTELLIGENCE SHOWCASE

Case Study: Leveraging AI for Meaningful Clinical Trials

linical trial sponsors seeking to de- latter with the study team. This helps save velop new drugs and therapies face time, as it helps to decrease the amount of C unique challenges, particularly when human review time typically needed with it comes to understanding patient dosing and high-risk patient populations. response to therapy. This aspect can be diffi- cult, as patients often: Optimized Patient Pools Via Predictive Do not take their medicine regularly, or at Technology all AI technology is also useful before the trial Take creative measures to hide their non-ad- begins. Using facial and voice measures, this herence technology can be utilized to screen poten- Have symptoms that make it difficult for tial candidates and build accurate, predictive clinicians to monitor and measure adher- models that help to optimize the patient pool ence with those patients most likely to be reliable followers of the care plan. Traditional standards for understanding Rich Christie, M.D., Ph.D. patient dosing are a) inefficient (enrolling Chief Development Officer additional patients increases costs), b) unreli- AiCure PharmaVOICEReal Impact for Your Studies able (pill counts can be easily falsified), and c) inconclusive (blood draws for PK measures). By integrating technology into therapeutic Because of these challenges, solutions must their own smart devices, patients can follow studies, teams can expect real, measurable be identified and implemented that support simple instructionsof for taking their medi- benefits: patients and trial sites as well as help address cations per protocol, and an application can Increased patient engagement and adher- the unique approaches necessary for successful utilize the device’s camera to recognize: ence: Real-time recording of each dose, clinical studies. That the patient is the correct person along with easy access to site support, build That the dose being taken is correct: it is a foundation for compliance the correct drug and the correct amount Optimized patient groups for more effi- Building a Solution That the dose has been swallowed cient, faster trials: Predictive technology That Works lets you limit the number of patients en- This, coupled with dosing reminders and rolled to those most likely to deliver reliable Problems with patient compliance spurred simple-to-use tools for receiving site support, data the need to develop ways to accurately moni- help to keep even the most challenging patient Decreased site burden: Using AI and tech- tor and measure medication dosing. By using populations more adherent. nology enables the ability to flag possible problems in real-time, and site teams can Negative Symptom Detection focus on providing the right levels of sup- Additionally, leveraging AI and computer port Having the right patient pool, vision on a cellphone platform enables the along with the technology to detection of a range of digital biomarkers that Having the right patient pool, along with are well-suited for addressing the specialized the technology to help accurately address the help accurately address the challenges presented by patients across thera- unique challenges of patients, will produce peutic areas. This is particularly true of nega- better data and decrease burden on sites while unique challenges of patients, tive symptoms, such as lack of facial affect, the saving sponsors both time and money. will produce better data and inability to show emotion and apathy. New Complimentsapplications, via artificial intelligence, moni- decrease burden on sites tor changes in facial affect as well as changes AiCure is an AI and advanced data in vocal characteristics via the smart device’s analytics company that monitors patient while saving sponsors both microphone. The AI also works to help iden- behavior and enables remote patient time and money. tify doses that were confirmed to have been engagement in clinical trials. successfully taken, and flag any doses where For more information, aicure.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] patients require further support, sharing the

28 October 2020 ● PharmaVOICE SHOWCASE ARTIFICIAL INTELLIGENCE

How AI Can Improve Michelle Marlborough Chief Product Understanding of Patient’s Officer AiCure Behavior and Symptoms unning and managing clinical tri- site, when this interaction is restricted to just predictable reasons: They lack a fundamental als, at the most basic level, is about the small amount of time the patient spends at understanding of what is required of them R creating a foundation to get the best a site visit, the likelihood of patients dropping during a trial, or they struggle to stay orga- possible data. While a great deal of thought out, or losing motivation increases. While nized and fit the requirements of the study goes into recruiting the right patients and more and more interaction with consumers is into their already busy lives. Leveraging AI developing the most effective protocols, ini- going mobile, the clinical trial industry is still and predictive analytics enable study teams tiatives designed to improve the experience for struggling to provide sites and patients with to predict how a patient will behave during the patients while at the same time gathering the tools to stay suitably connected between a clinical trial with high accuracy within the cleaner, more accurate data are relatively new. site visits. The lack of visibility into how a first few days. They can use this data to recruit There are many exciting potential applications patient is doing with things as simple as tak- patients who are more likely to comply with of AI in clinical trials. In terms of understand- ing the medication each day makes it hard for the study protocol or can proactively provide ing patients’ behavior and symptoms, there are sites to provide the right support to the right support to ensure patients stay on track. a couple of very impactful opportunities. The patient and the right time. Furthermore, patients can be assessed re- first is the opportunity to use tools like com- motely between clinic visits using digital puter vision to objectively assess symptoms biomarkers. Using facial and voice recognition that have traditionally been very subjective, A Better Way and the front-facing camera of a smartphone, especially in areas such as depression and other patient symptoms can be assessed, measured, CNS disease. The other is that it can be used to The world is getting more and more and recorded. For example, speech patterns extend a site’s ability to support, monitor, and connected, which means that therePharmaVOICE is a high and facial expressivity can indicate if a patient assess patients outside of the clinic, in a way probability that patients have a smartphone is experiencing depression, or the smartphone that is simpler and less intrusive for patients with them at all times. This provides a unique camera can measure and record a facial or hand but that can lead to better compliance and opportunity to not only collect new and better tremor. A computer-vision and AI platform better data, which is exceptionally important data but also to assist theof patient at every step can enable study teams to collect, curate, and especially during COVID-19. in their clinical trial journey, not just when annotate these patient insights extracted with they are physically on-site. Patients can receive cutting-edge analytical tools. regular reminders and guidance on taking What Makes Patient their medications, completing an assessment, Engagement Difficult? or attending clinic visits. Applications can A Focus on the Patient utilize a phone, and the in-built cameras to use Helps Us All How engaged a patient is in a clinical trial AI and computer vision to confirm activities will be dependent on many factors, and is such as taking their medication or to record A positive patient experience can go a long likely to change over time depending on their more frequent, less intrusive, interactions for way to improve data through better reten- experience within the trial. The main way of longitudinal assessment of disease symptoms. tion and dosing support. A patient who feels ensuring that a patient stays connected and This always-on link between the patient and supported is more likely to remain engaged motivated is through the interaction with their site allows the site to know who needs help through the course of the trial. and when. Patients are able to receive support Technology can be extremely useful in when they need it, often eliminating the need helping patients to manage their study par- for additional clinic visits. This can be a breath ticipation along with the litany of other life Technology can be extremely of fresh air for patients who no longer need to responsibilities. At the same time, it can ease useful in helping patients to worry about remembering each assessment or the burden for study teams, allowing them manage their study participation dose, whether they took it correctly or if it was to have more meaningful interactions with recorded correctly. Sites and sponsors get the patients. All this leads to a happier, more en- along with the litany of other benefit of complete and accurate data, all in gaged patient pool. life responsibilities.Compliments At the same one place. time, it can ease the burden AiCure is an AI and advanced data for study teams, allowing The Power of Remote analytics company that monitors patient them to have more meaningful Engagement and Assessment behavior and enables remote patient engagement in clinical trials. interactions with patients. Patient retention is a top priority in every For more information, aicure.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] clinical trial. Patients who drop out, do so for

PharmaVOICE ● October 2020 29 ARTIFICIAL INTELLIGENCE SHOWCASE 5 Steps to Intelligent Omnichannel Engagement

he future of HCP engagement in the field to understand how their efforts support post-pandemic world isn’t clear. But brand strategy and capturing on-the-ground T in the wake of pharma’s forced digital intelligence to optimize future interactions migration, the need for an effective omnichan- across channels. An intelligence engine can nel strategy is. also function as a commercial command cen- Luckily, most commercial teams already ter, bringing all channels into one interface Matthew have a strong data and analytics infrastructure that provides the necessary perspective for Van Wingerden in place and are beginning to activate non-per- managing complex campaigns. sonal channels at scale. But for an industry GM Advanced that’s more complex than most, what does a Services transition from the legacy commercial model 3. Bake Agility into Aktana actually look like? Your Brand Strategy AI will be indispensable as pharma brings example, instead of automatically sending an intelligent omnichannel engagement to life, With so many moving pieces, conflict be- HQ email to all conference attendees two days but implementation isn’t as simple as flipping tween channels and analytical inputs is likely. later, task AI with evaluating channel affinity, a switch. Follow these five steps to ensure your To mitigate this, add business rules that clarify engagement timing, and other interactions to omnichannel approach balances impact with when certain triggers should take priority, select the next-best-action for each HCP. scalability for maximum results. enforce compliance with local regulations, With this approach, you can orchestrate and set blackout periods to prevent redundant hundreds of highly targeted, personalized cus- contact, constraining AI-driven suggestions tomer experiences from one campaign — 1. Create a Clear Picture within the bounds of what’s reasonable for the without manually micro-segmenting your au- of Your Customer brand and HCPs. PharmaVOICEdience. It’s a win-win: less work for you, more This is critical for producing suggestions value for your HCPs. Successful omnichannel campaigns require and insights that “feel right” and driving a detailed understanding of each customer. adoption among your commercial team. Lack of data isn’t the problem. The challenge of Campaign Objectives is considering all data sources simultaneously Are Brand Objectives while separating the signal from the noise. 4. Balance Optimal Working within the context of your brand Experience with Omnichannel doesn’t need to be an all-or- strategy, machine learning algorithms can nothing proposition. In the interest of long- scour mountains of data to predict which Optimal Resource Use term adoption, it shouldn’t be. By taking these combination of channel, engagement time, measured steps, you’ll not only ease change and marketing message will yield the greatest When communicating with one HCP, management for your team, but also lay the impact. optimizing brand interactions is simple: the groundwork for a successful strategy that can Along with evaluating past brand inter- most preferred channel + the optimal time easily scale. actions, AI can use individual healthcare pro- to engage = the ideal touchpoint. Machine As you increase the frequency of launching vider (HCP) attributes to make predictions, learning allows you to expand that approach to discrete campaigns and begin to reshape your evaluating every characteristic — from patient multiple customer journeys, making it easy to internal processes in parallel, leveraging AI to coverage data to when they graduated medical weigh the value of an action against its cost to chart the best path forward throughout the school — against a large dataset of HCPs determine the most efficient next step. entire targeting process will be a natural next to determine communication preferences and In addition to an HCP’s bandwidth for step. likelihood to engage on a one-on-one level. “information overwhelm,” be sure to consider each channel’s limitations. How much email content do you have? How many virtual visits Aktana is a pioneer in intelligent 2. Break down Silos with can a rep reasonably accomplish in a week? engagement for the global life sciences Real-Time Visibility Automate and learn from these tradeoffs to industry. Its Contextual Intelligence continually optimize your approach. Engine leverages a proprietary blend of An AI-drivenCompliments intelligence engine is your AI, human insight and other advanced best ally for ensuring every touchpoint en- technologies to help life sciences teams hances what precedes it and sets the stage for 5. Start Small and Scale seamlessly coordinate and optimize what follows, aligning teams in two critical personalized omnichannel engagement ways. A great way to incorporate intelligent with healthcare providers. As a next-best-action tool, it links sales and omnichannel engagement in your daily work- For more information, visit aktana.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] marketing in a virtuous cycle, empowering the flow is on a campaign-by-campaign basis. For

30 October 2020 ● PharmaVOICE The shift to digital just became critical. To fully support HCPs when you can’t visit, the right intelligence partner makes all the difference.

Your customers deserve a seamless omnichannel experience that speaks to the topics they care about most. AI alone can’t get you there, but Contextual Intelligence can.

Using the right blend of machine learning, explainable AI, human intelligence and other advanced technologies, Aktana’s Contextual Intelligence Engine gathers insight from every interaction to make those that follow more personalized and precise. PharmaVOICE Find out more at aktana.com/ci of

Compliments (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] ARTIFICIAL INTELLIGENCE SHOWCASE

Technology-Driven Pharmacovigilance Operations

ogy1. The agency has also indicated an Improving Data Quality and openness to using AI-driven tools and Compliance while Reducing technologies to support new platforms2. Cost with AI Another initiative is the WEB-RADR, an EU-wide mobile phone app enabling Dinesh Kasthuril raditionally, the pharmacovigilance users to report adverse drug reactions Head of Global (PV) function has been responsible for (ADRs) directly to the national compe- Client Relations T collecting, assessing, and reporting tent authority. These efforts have shown & Project safety information to health authorities, sites, progress in improving both data quality Management, and other stakeholders. Companies have in- and greater insight by leveraging tech- vested in safety systems to gather, organize, nology. Patient Safety and search data. Yet, these closed systems There are significant gains from ma- Solutions often have lacked the real-time integration and chine learning and AI-based automation Covance sophistication needed to offer the insights that tools, which offer solutions to chal- patients, physicians, providers, and regulators lenges that existing systems cannot solve. For industry lacks integrated solutions that com- demand. Technology and automation, includ- example, the cost and volume of individual bine activities such as case prioritization and ing AI, will enable the transformation to more case safety reports (ICSRs) increase yearly3, processing, aggregate reporting, and signal personalized and proactive PV. This article but it is estimated that more than 90% of detection. discusses the complex journey. adverse events (AEs) go unreported4. The Automation offers significant operational uptick in informal reporting, especially from and strategic benefits, but full transformation social media, has encouragedPharmaVOICE pharmaceutical entails a multi-year journey leveraging various Technology Across the companies to look for cost-effective solutions technologies as outlined in this transformation Value Chain to handle structured and unstructured AE roadmap (Figure 1). data. As reports indicate that ICSR processing Basic process automation involves tracking Regulators are exploring technology con- contributes to 50%of of PV spending,5 there is and monitoring tasks with continuous met- cepts across the pharmaceutical value chain. justifiable interest in automation technologies. rics collection and real-time reporting. The For example, the FDA recently permitted de- benefit is increased efficiency in time- and vices for detecting diabetes and predicting the cost-savings. chances of a stroke based on cognitive technol- Transforming PV Robotic process automation (RPA) helps reduce manual effort and errors, resulting in As of today, despite safety databases, sig- automated data entry for objective informa- nificant components of PV operations are tion in structured formats. Benefits include Real-World Examples effort-intensive and spreadsheet-driven. The increased quality and efficiency. A Fortune 50 pharmaceutical client implemented Covance RPA solutions to augment manual PV processing. Figure 1: Pharmacovigilance Transformation Roadmap The goal was to transform quality and improve efficiency in two areas: PV Case Processing: Our RPA Case Processing Assist Tool improved quality with 100% accuracy for automated fields and increased productivity by 15% on 125,000 cases. Compliments PV Intake Processing: Our RPA Intake Process Assist Tool increased productivity by 18% on a volume of 14,000 cases/month, with 100% accuracy on automated fields. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected]

32 October 2020 ● PharmaVOICE SHOWCASE ARTIFICIAL INTELLIGENCE

from this data. Cases identified would undergo a physician’s as- Figure 2: Covance Suite of Automation Solutions sessment to drive both regulatory reporting and real-time signal de- tection. Both the data sources and technology to enable this are real and achievable today. These solu- tions would also be extended to prescribing physicians, allowing them to make better decisions and develop a tailored and per- sonalized treatment plan based on both the patient’s medical history and real-time data from other patients. This vision moves beyond un- derstanding safety-related prob- lems to transformational PV: a next-generation, patient-centric system that uses deep learning to Cognitive automation leverages natural lan- Envisioning the Future of PV improve a product’s benefit-risk profile, help guage processing (NLP) to help humans PharmaVOICEpatients select an optimal treatment, and en- make decisions, often combined with RPA. hance patient safety. Cognitive computing automates subjective Covance has a vision of bringing value to content and supplements human efforts for patient safety by leveraging delivery excel- References: data interpretation. A key benefit is in- lence, automation capabilities,of process trans- 1. fda.gov/news-events/press-announcements/fda- creased medical insights but note that cog- formation expertise and use of analytics on permits-marketing-artificial-intelligence-based-de- nitive intelligence may still require human real-time data. vice-detect-certain-diabetes-related-eye interaction. For the past few years, we have been part- 2. fda.gov/news-events/press-announcements/ Artificial intelligence (AI) is envisioned to nering with clients to develop automation fda-permits-marketing-clinical-decision-sup- be more independent and predictive by le- solutions by streamlining processes, rapidly port-software-alerting-providers-potential-stroke veraging large volumes of information. prototyping automation tools and successfully 3. windwardstudios.com/white-papers/exist- Machine learning, a subset of AI, enables deploying solutions with short cycle times. ing-functions-of-pharmacovigilance-system data scientists and analysts to construct al- These solutions have provided efficiency 4. Mockute, R., Desai, S., Perera, S. et al. Artificial gorithms that can learn and make predic- gains of approximately 15-25% over two Intelligence Within Pharmacovigilance: A Means tions based on data. Deep learning takes years, resulting in over a million dollars in to Identify Cognitive Services and the Framework this idea further, processing information in savings for clients while improving case qual- for Their Validation. Pharm Med 33, 109–120 layers where the output from one layer be- ity scores. (2019). doi.org/10.1007/s40290-019-00269-0 comes the input for the next one. The ben- In the next phase of this journey, we an- 5. navitaslifesciences.com/implementing-ai-and-ro- efit is an increased focus on proactive PV. ticipate benefits from leveraging AI-enabled botics-in-pv-a-practical-approach solutions specifically to interpret subjective Leveraging the technologies mentioned information mimicking human intelligence in the transformation roadmap (Figure 1), during case processing and auto-assessing Covance is a business segment of Covance has deployed and continues to build ICSR validity for literature cases. For author- LabCorp, a leading global life sciences on a suite of automation solutions (Figure 2) ing, we are employing machine learning and company, which provides contract across the PV spectrum towards one-touch natural language processing to auto-gener- research services to the drug, medical case processing, intelligent authoring and pro- ating key sections of the aggregate reports. device and diagnostics, crop protection active signal management.Compliments These solutions For the signaling, we are looking at external and chemical industries. Employing are validated and work seamlessly with safety real-world data sources aided by visualization over 21,000 people worldwide, we are databases. and real-time analysis. the world’s most comprehensive CRO, Today, we leverage these solutions across Long term, we are working toward a plat- dedicated to improving health and multiple clients, resulting in efficiency im- form that integrates real-world data sources improving lives. provements of about 25% for the processes such as lab data and medical records, which in For more information, visit covance.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] where they are deployed. real time curates and identifies potential ADRs

PharmaVOICE ● October 2020 33 ARTIFICIAL INTELLIGENCE SHOWCASE

AI and Drug Discovery

iscovering and bringing new drugs to market is notoriously slow and costly D for a reason. Science moves slowly Although not a silver bullet, because of the incredibly complex nature of human biology and the difficulty of sluggish the use of AI is expected to recruitment of clinical trials. contribute toward finding And in spite of greater investments in R&D by pharmaceutical companies, the num- novel candidates as well as ber of new drug approvals has declined. Al- though not a silver bullet, the use of AI is offer cost and time savings expected to contribute toward finding novel for the life cycle of clinical candidates as well as offer cost and time sav- ings for the life cycle of clinical research. research. Today pharmaceutical organizations are getting a lift from artificial intelligence (AI) as it provides the ability to sort through massive volumes of data in shorter time periods to find cess rate of delivering life-saving treatment patterns that increase the efficiency of all major options. processes in the drug development life cycle. Claire Bonaci 4. Clinical Trial Design & RecruitmentPharmaVOICE Patient recruitment is tedious and one of the Senior Director, US Health and Six Drivers of Adoption major reasons for exorbitant cost and trial Life Sciences failure. AI and digitization can support trial Microsoft In looking ahead there are six key areas design, protocol ofdevelopment, and patient that research from Frost & Sullivan show will matching and stratification/enrollment. life-saving drugs to those with unmet needs increase the adoption and use of AI in the faster and more efficiently. pharmaceutical industry.i 5. Clinical Trial Optimization & Execution Microsoft has begun an AI-based collabo- Digitization supported with AI can resolve ration with leading global medicine company, 1. Target Discovery and Validation the many challenges resulting in high trial Novartis, to transform drug development and Identification of drug targets from net- costs. From protocol optimization, appropriate commercialization. work-guided machine learning/AI is on the site selection, drug adherence and compliance Similar collaborations and partnerships rise in pharma and biotech companies. Target tracking, to reduction of adverse events, the will continue to arise as the focus on data prioritization and validation of hypothesses use of AI can greatly reduce the manual and science and AI models become more prolific linking drug targets to disease mechanisms monotonous tracking that is needed during a and value is seen from the early collaborations can increase success rates for certain molecules trial life cycle. and partnerships. Goals of reducing time to for a given therapeutic area, reducing time to market, increasing operational efficiency, and human trials and ultimately to market. 6. Precision Genomics reducing costs can be aided by expanding re- The complexity of data coming from a variety lationships between life-sciences organizations 2. Focus on Candidate Discovery of sources can be overcome by applying AI and big tech companies. AI can be used for augmented design and con- and Big Data techniques. AI can be applied ditional automation, predicting efficacy and towards patient stratification, improving diag- Editor’s Note: i Growth Insight—Role of AI in safety of a new molecule and identification of nosis, and decision support that augments the the Pharmaceutical Industry Global, 2018–2022, secondary (life cycle) opportunities. capability of clinicians. Frost & Sullivan, September 2019

3. Drug Repurposing ComplimentsPartnerships Key to AI will enable extracting insights from the Microsoft Health and Life Sciences peer-reviewed, patent literature, patient data, Moving Forward empowers healthcare organizations of and other sources to identify the right already all sizes to reimagine the ways they bring existing compounds for a particular disease. AI Partnerships leveraging innovative pharma together people, processes, and health can aid in the decision making of combination and life-sciences organizations and technology data insights to improve care delivery. drug trials, multiple cohorts, and advanced organizations such as Microsoft with cloud- For more information, visit microsoft.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] trials, ultimately increasing the overall suc- based AI technology are crucial to delivering

34 October 2020 ● PharmaVOICE Better experiences. Better insights. PharmaVOICE Better care. of

Microsoft for healthcare provides innovative solutions that empower organizations to achieve more.

With Microsoft’s intelligent cloud platforms and powerful partner ecosystem, you can better enable personalized care for your patients, empower your care teams, optimize your clinical and operational effectiveness, and transform the care continuum for your organization. ComplimentsTogether, we all can realize a healthier future, today.

(c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] Learn more at Microsoft.com/health ARTIFICIAL INTELLIGENCE SHOWCASE The Accessibility of AI: Innovation at Your Fingertips he availability of datasets keeps grow- Synthesizing new drugs: through deep Max Divak ing exponentially. With this we are learning, AI can speed up the discovery process. Ogilvy Health T witnessing an escalation of available By programming a massive range of variables Interactive Group Supervisor new life-altering systems. Some of these tools into an AI, it can make better predictions in are highly autonomous and can perform func- turn sending new drugs to human trial in interventional cardiologist Dr. Paul Lee who’s tions with little to no human intervention. shorter timeframes. affiliated with the NY Mount Sinai School of Self-driving cars, rockets that land upright on Medicine. After realizing he could one day be drone ships, computers that write poetry and Clinical trial research, data analysis and replaced by AI, he made the decision to be the perform their own music, tireless warehouse recruitment for enrollment: patients usually one programming the AI and started taking workers, and medical tools that are capable of rely on their doctors to learn about applicable online courses in Python, deep learning and making diagnostic decisions. These technolo- studies. Or they search on ClinicalTrials.gov, computer vision techniques. Within two years gies use massive amounts of data, machine and which contains more than 300,000 studies, a of research and after compiling a reference deep learning with language processing, audio daunting task of trial and error. AI can help library of images, he was able to train an AI to and visual processing algorithms all working in filter out 60%-80% of trials that the patient analyze and interpret coronary angiograms, de- orchestration so they can form decisions on how wasn’t eligible for. tect blockages in patient arteries with high con- to perform primary functions effectively. fidence levels, and help those patients reduce Radiology and radiotherapy imaging di- and prevent heart attack. With improvements Articial Intelligence Innovation agnosis: trained physicians manually review made in the available libraries, he’s able to run medical imaging to make a subjective assess- his code on a mobile device. Time otherwise There is a fearful side of the AI conversation ment based on education and experience. AI, on spent analyzing these scans can now be spent when we mention the robots taking our jobs or the other hand, can recognize complex patterns focusing on other aspects of his practice. He how The Terminator movie is going to become and provide quantitative assessments almost was invited to present his research and findings reality. But the more realistic outcome for now instantly. When healthcare professionals are at the American Heart Association Scientific is that our AI counterparts are powerful tools we under intense workload, they PharmaVOICEmay not have the Session in Philadelphia last year. Lee PC, Lee program that require our command and enhance time required to interpret every image with the N, Pyo R. Abstract 12950: Convolutional our abilities. They make us better craftspeople, same care. AI can increase efficacy and reduce neural networks for interpretation of coro- military personnel, or healthcare professionals errors. nary angiography. Circulation. 2019;140(suppl and leave us available for higher-level intel- of 1):A12950-A12950. https://www.ahajournals. lectual tasks. Why does cyborg enhancement Wearables and Internet of Things: great org/doi/10.1161/circ.140.suppl_1.12950. come to mind? Maybe I read and watch too strides have been made in devices that can Those with technical acumen should ex- much science fiction and we’ve all had life-im- diagnose from home. The goal is to open up plore the many open platforms. Learn Python, itates-art hammered into our psyches as a part a HCP’s schedule by ruling out unnecessary at a minimum, it will help you work better of the culture. The technology we use every day visits. During a pandemic, higher risk patients with data. See if you can fold these techniques is composed of electronic components that have won’t need to risk the visit if it is not needed. into your everyday workflow. If you don’t have been miniaturized to a minuscule scale. What a desire to use these techniques, keep in the was once many different tools performing ded- Epidemic outbreak modeling and predic- know of the latest innovations in ethical AI by icated functions now can live on one powerful tion: a Canadian start-up that uses AI to detect checking out https://openai.com. Some of the device that easily serves up a laboratory of useful disease outbreaks was one of the first to initially ways we use AI in the Innovation Lab include functions at your disposal and fits in our pocket. raise concern for the outbreak of a respiratory sending sensor data, as part of one protype, to Because of this transformation, many tools are illness in Wuhan, China. Now AI is used to the cloud for further analysis to detect patterns now powerful enough to run the computations determine many factors concerning COVID- and improvement in patients’ health over time. required for AI models and algorithms making 19: where outbreaks will occur, what regions Another use has been to scan submitted video these everyday devices smarter and affordable. need hospital beds, patterns in spread and more. clips of walking patients to analyze and detect The pharma and medical field has the potential skeletal abnormalities. to be a leader in AI innovation. Man and Machine We always want to partner with our brands to find the best use cases for the technology that Disease identification and diagnosis: stud- What excites me about healthcare making make the most sense for their business and most ies show that AI is as accurate as healthcare pro- strides in AI innovation is that it’s at a time importantly their patients. fessionals at analyzing medical imagining and when the barrier of entry to learning the tech- diagnosing conditionsCompliments and identifying disease. nology and cost of hardware to run it for our personal use has become attainable to more Optimizing personalized treatment and people. Many of the libraries and methods Ogilvy Health makes brands matter behavioral modification: AI streamlines used for machine and deep learning are open- by keeping our audiences’ health, finding the appropriate drug combinations and source and backed by great communities. These healthcare and wellness needs at the dosing strategies based on the individual. This technologies are accessible to anyone with a center of every touchpoint. ultimately limits side effects and improves the desire to learn a human readable programming For more information, visit ogilvyhealth.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] quality and length of a patient’s life. language like Python. An example of this is

36 October 2020 ● PharmaVOICE SHOWCASE ARTIFICIAL INTELLIGENCE

AI-Driven Marketing

e have all seen those digital ban- approach can be implemented. Fewer impres- ners that are somehow related to sions to a customer base that shows interest in W what we’re looking at online and a brand lead to higher, more quality engage- appear to follow us wherever we are. Don’t ments. The data captured during these en- worry, there is nothing ominous at play and gagements allow the brand to make decisions the government is not listening to our con- based on real-life insights into where physi- versations through our devices. But this is not cians are engaging with content and — more completely coincidental either. Those “timely” importantly — what type of content they are and relevant ads are part of a larger marketing more likely to engage with. This is invaluable trend that has become commonplace among as brands continue to fight for a voice in a very our traditional B2B and B2C marketing coun- crowded marketplace. terparts. The use of artificial intelligence (AI) On a broader scale, the data from AI- and machine learning to identify, message, driven campaigns can be shared and imple- and convert customers based on their real-life mented across the entire digital ecosystem. behaviors has provided consumer marketers Integration with existing marketing automa- J.P. Maranzani the ability to broaden campaign reach and in- tion and salesforce automation systems (SFAs) crease efficiencies while continuing to provide VP, Account Group Supervisor - is commonplace in the B2B space. Providing a relevant and timely content. As pharmaceu- Point of Care Practice (EHR/ABM) field force representative with behavioral data tical marketers, we need to go beyond just Ogilvy Health PharmaVOICEwill allow them to cater every call to the spe- saying we recognize AI as an important trend. cific needs and interests of the HCP. This will Marketers need to embrace AI when develop- be invaluable as well when we start to navigate ing digital marketing strategies. Otherwise, (endemic vs nonendemic sites). Imagine if a post–COVID-19 world. we run the risk of being behind yet another you could now identify,of execute, and refine a Historically, the pharmaceutical industry marketing innovation. marketing campaign in real time and based on has been very meticulous and careful when real-life customer behaviors. As we enter a new implementing new strategies, especially when frontier, the advancements in AI and machine it comes to technology. There is usually a long Marketing Smarter learning provide marketers with the capacity phase of testing and vetting prior to a large- to do just those things. Not only can marketers scale acceptance. Getting brand leads, regula- AI and machine learning executed through message to a known customer base on known tory, and IT comfortable with the concept of tactics such as account/behavioral-based mar- sites, but AI now allows marketers to learn AI and machine learning early on is of utmost keting (ABM) have been mainstays in B2B from customers’ behaviors and send messages importance. Once the team has accepted the marketing for decades. In the late 1990s, to them based on seeing precisely where they idea of AI-driven marketing, the implemen- consumer brands started using early concepts are online and what content they are con- tation is easy. of AI to help deliver more personalized expe- suming. All in real time. The days of waiting As the world of pharmaceutical marketing riences to a customer base. This was normally weeks, if not months, to review, identify, and tries to keep up, it is imperative for brand executed through predetermined media place- retarget are gone. AI and machine learning teams to embrace technology as early as pos- ments designed to provide a specific customer take the guesswork and time lag out of our sible. Reactionary implementation generally base with relevant ads. Relevance and ad marketing. The more a customer engages, the leads to less than optimal campaigns and placements were based on a set of assumptions more AI learns, and the more we can adjust. less than optimal HCP experiences. Since an that would try to determine which content HCP’s first impression with a brand could be would cause a reaction in the customer. This the difference between an Rx and a compet- is still a common practice with the develop- Real-Time, Real-Life Insights itor, using AI to help create a personal and ment of customer personas. Although highly relevant experience is the way to go. successful, the driving strategy was still heav- So what does this mean for pharmaceuti- ily based on assumptionsCompliments and guesswork. As cal marketers? In short, a lot. AI provides us technology has advanced over the years, so has with access to information that we historically Ogilvy Health makes brands matter our ability to market smarter. Broad media have not had and the aptitude to act in real by keeping our audiences’ health, placements have given way to more targeted time. Marketers can use a brand’s budget to healthcare and wellness needs at the buys in which marketers focus on a physical be more precise and efficient with campaigns. center of every touchpoint. location (geotargeting), a more specific type of By serving up ads based on what a customer For more information, visit ogilvyhealth.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] customer base (based on sales data), or content base shows they like, a “quality over quantity”

PharmaVOICE ● October 2020 37 ARTIFICIAL INTELLIGENCE SHOWCASE Leveraging AI to Slingshot Drugs from Lab to Market

he science of clinical development Machine learning models, trained to predict continually astounds even those of us data discrepancies and manage queries using T intimately connected to it. We recog- historical clinical data, can offer a significant nize that our overarching industry mandate — process improvement that reduces time to to develop and deliver safe and effective new database lock. Sponsors and CROs can use cur- therapies to fight and cure disease — must rent clinical and scientific data from electronic be continually informed by new thinking and data capture (EDC) systems and third-party tools if we are to realize not just incremental, sources (labs, biomarkers, PK/PD) to almost but exponential success. The COVID-19 pan- instantly identify discrepancies and generate demic has underscored the need for this in an query text related to: unprecedented way. Baseline characteristics (e.g., inclusion/ex- Likewise, the engineering of artificial intel- clusion criteria violations) ligence (AI)-powered automation also contin- Randomization/stratification (e.g., multiple ues to astound. The pharmaceutical industry, enrollments for the same patient) Malai Sankarasubbu which always has its ear to the ground for ways Compliance (e.g., missing study visits) to improve the clinical trial process, is begin- VP of AI Research Disposition (e.g., discontinuation reasons) ning to embrace AI to a greater degree than Saama Technologies Efficacy and safety (e.g., data outliers) ever before. COVID-19 has highlighted the Data managers and medical monitors are astonishing applications AI offers biopharma, the industry to rapidly map raw data to STDM able to identify data quality issues more effi- and its almost limitless potential, while at the and expedite it for analysis. ciently and enable continuous improvement as same time underscoring the industry’s over- By working together, biopharma and SaaS machine learning models grow even smarter. whelming need to adopt and integrate AI to product companies can developPharmaVOICE and deploy AI as a tool can empower the industry to better accelerate clinical trials. However, in order to AI-powered analytical tools to clear many of understand its clinical trial models, helping to fully leverage AI for clinical development the the obstacles faced by study data managers speed up trials and get drugs to the market industry must trust AI, and we aren’t quite and monitors. Combining AI-powered tech more quickly. there yet. It is only through the collaboration platforms with biopharma’sof clinical data and of the great scientific minds of biopharma and domain knowledge facilitates the creation of AI During COVID-19 the great engineering minds of tech that such deep learning models trained to make predic- trust can be established, thereby enabling the tive analytics with unprecedented accuracy. A current real-world example involves a life sciences industry to effectively leverage AI to slingshot safe and effective drugs from lab to market in record time. Connecting Scientific and Engineering Minds Through AI

Connecting Scientific and Engineering Minds Thru AI Just as drug development begins in the scientific lab, AI development begins in the engineering lab. Though seemingly separate and distinct, the rigorous research that serves as the foundation for both is actually parallel and complementary. Unfortunately, at the early stages there is usually no line of sight from one to the other. Leading software-as-a-service (SaaS) prod- uct companies have created AI research lab- oratories thatCompliments focus on deep learning to opti- mize clinical trial technology and accelerate the drug discovery and development process. The AI algorithms and solutions born of the advanced engineering research conducted at these facilities bring immense value to clinical (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected] trial data review and management, enabling

38 October 2020 ● PharmaVOICE SHOWCASE ARTIFICIAL INTELLIGENCE

research is achieving a comfort level for entrust- ing valued clinical development efforts to this COVID-19: AI in Action amazing technology. This is where AI explain- ability comes in. Explainable AI describes the processes and methods used to apply AI, so that the results of an AI-generated solution can be understood by humans. It hinges on clear artic- ulation of how the various techniques involved in machine learning models generate their out- puts, thereby helping end users trust that the AI is making good decisions. Explainable AI is the antithesis of machine learning as a mysteri- ous “black box,” for which sometimes even tech engineers can’t explain the decisions produced. The focus of explainable AI is transparency and providing direct line-of-sight to under- standing the predictions at which AI arrives. It is crucial that AI tech partners to life-science companies ensure explainable AI in order to establish and enable its continued, effective de- ployment, as well as the ability to build upon and expand today’s successes into tomorrow.

state-of-the-art AI automation platform being ers, genome sequences, radiology scans, and The Next Frontier of AI leveraged to ensure data quality and manage- more, within and across studies ment for a COVID-19 vaccine trial involving Securely integrate internal, proprietary in- The road to precision medicine of the over 40,000 people. Because of the AI-pow- formation with externally sourced historical future, where a person’s genetic make-up, life- ered solution being utilized, data entry to and real-world data to improvePharmaVOICE patient en- style, and environment inform how researchers data cleaning took less than 24 hours. Natural rollment and site selection develop targeted therapies, is paved with AI. language understanding (NLU) and machine Leverage an integrated correlation engine The one-size-fits-all approach to clinical de- learning enabled these powerful results. (ICE) to correlate COVID-specific markers velopment will be cast aside in favor of a more In a more broad application to the industry with clinical outcomesof personal approach to disease treatment and during the pandemic, AI is now also being Search databases containing tens of thou- prevention. Existing and future generations of built into command center solutions targeted sands of scholarly articles about COVID-19, AI platforms can and will continue to facilitate at COVID-19 to combine the power of pre- SARS-CoV-2, and related coronaviruses dynamic visualization, analysis, and interroga- cision medicine with large-scale clinical trial Such command center tools also have broad tion of data across clinical research programs to operations. With unprecedented visibility into applicability to other therapeutic areas, be- further realize the promise of precision med- multiomics and clinical data sets — along with yond COVID-19, and offer the ability to icine. The seamless integration they offer can historical data from past trials — sponsors and reshape clinical research processes and timing. be leveraged to curate and animate clinical trial CROs can improve inclusion/exclusion criteria, data for more actionable insights and faster, monitor individual patients, and track trends Explainability more reliable decision making. in patient cohorts, all while ensuring patient Key to AI Trust The great scientific minds of the pharma- safety and drug efficacy. Such AI-powered ceutical industry can utilize these advanced AI command centers enable biopharma to: Success, however, is built on a foundation platforms, developed by the great engineering Analyze patient data, including diagnostic of trust. Critical to the industry’s willingness minds of the tech industry, to do their jobs markers, COVID-19 inflammation mark- to and effectiveness at deploying AI in clinical more efficiently than ever before. By learning to trust AI and leveraging it in clinical devel- opment, safe and effective new therapies can The Road to Precision Medicine of the Future be slingshot from lab to market for the people who need them, both during and after the COVID-19 pandemic. Compliments Saama Technologies is the No. 1 AI clinical analytics platform company, enabling the life sciences industry to conduct faster and safer clinical development and regulatory programs. For more information, visit saama.com. (c) PharmaLinx LLC. Rights do not include promotional use. For distribution or printing rights, contact [email protected]

PharmaVOICE ● October 2020 39