Amar Drawid, Ph.D.

An Integrated Data Review Plan streamlines the clinical trial process by eliminating redundant data reviews, establishing clear quality objectives, and minimizing the overall burden of data review. #clinical #clinicaltrials #clinicalresearch #biometrics #statistician #data #healthcare #rbqm #rbm #innovation #sasprogrammers

48

Key Differences b/w Randomized Clinical Trial (RCT) vs RWE (Real-World Evidence): Learn more about Real World Data: https://lnkd.in/gRGZdJJM Reference: https://lnkd.in/gN2vf6Hp #clinicaltrials #clinicareserach #data #healthcare #innovation #patient #regulatory #india #rct #rwe

20

🌍 Everyone deserves a chance at a healthy life. But who you are, where you were born, or how much you earn, can determine how healthy you are. 💻 Data and health technologies like AI are powerful tools that can help address inequities by identifying communities that need better care, guiding healthcare resources toward them, and improving the quality of care received. 🤝 Partnerships play a pivotal role in bringing the skills and resources of the public and private sectors together to work with communities. To hear more, watch the video from our Head of Global Policy, Advocacy and Health Equity, Stefan Weber. #WHS2024 #HealthEquity #Interconnectedhealth

18

Thoughtful planning in the establishment of a commercial data strategy prior to brand launch can help to frame the approach to patient-centricity. Data warehouses are a lot of work to rebuild, the tips in this Beghou Consulting article help to provide four keys to inform that planning so that patients stay at the center of focus. #data #pharma #biotech #patient-centered

15

Thoughts on this? >> Harnessing Introspection in AI: How Large Language Models Are Learning to Understand and Predict Their Behavior for Greater Accuracy >> Comment below! >>> lqventures.com #industry40 #IoT #AI #mhealth #healthtech

2

2024 was a year of transformation, growth, and innovation! 🚀 لقدكان عامامليئا بكرم الله ولطفةوستره فالحمدلله على ماتبقي وعلى ما أتى وعلى ماهو قادم والحمدلله دائما وابدا It taught me that healthcare isn't just about algorithms it's about bridging AI innovation with real-world pharma impact. 🌟 𝐇𝐞𝐫𝐞’𝐬 𝐡𝐨𝐰 𝐢𝐭 𝐚𝐥𝐥 𝐮𝐧𝐟𝐨𝐥𝐝𝐞𝐝: 🌟 𝐊𝐞𝐲 𝐌𝐢𝐥𝐞𝐬𝐭𝐨𝐧𝐞𝐬 🌐 Reached 111,921 views sharing insights on AI and healthcare innovation! 📝 Published my first research paper on Value Innovation Strategy in pharma, introducing new frameworks for strategic decision-making. 🎓 Achieved 15 certifications in AI & leadership, including IBM AI Academy and Microsoft's Generative AI programs. 📊 Led an analytics transformation that boosted sales by 25% through smarter, data-driven decisions. 🔥 𝗧𝗼𝗽 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 𝗠𝗼𝗺𝗲𝗻𝘁𝘀 1️⃣ "First Research Publication" How Value Innovation Strategy is shaping the pharmaceutical market. 👉 "Challenges, learning, and growth are what make the journey worthwhile." Read more here "https://lnkd.in/dkS3NT-c". 2️⃣ "AI Empowerment at EPSF Congress" Sharing AI knowledge with future pharmacists. 👉 "Every idea shared is a step closer to progress." Catch the highlights!"https://lnkd.in/dDfq7puQ" 3️⃣ "ChatGPT-5 Predictions" Analyzing AI’s future capabilities and their impact. 👉 "We aren’t ready for what’s coming." Get the insights here."https://lnkd.in/dKfjYuaP" 🔭 𝐋𝐨𝐨𝐤𝐢𝐧𝐠 𝐀𝐡𝐞𝐚𝐝 𝐭𝐨 𝟐𝟎𝟐𝟓 Next year is all about scaling impact Transforming pharma and healthcare analytics through AI while continuing to share knowledge and insights. The convergence of AI and healthcare opens up endless opportunities for innovation, and I’m excited to be at that intersection! To our growing community of healthcare innovators and AI enthusiasts T♥H❤️A♥N❤️K♥❤️ ♥Y♥O♥U for engaging, challenging, and growing with me. 🙌 Together, we’re rewriting the future of healthcare, one line of code at a time. 💻💡 Let’s keep building. The best is yet to come! 🌟 #LinkedInRewind #2024Wrapped #AI #HealthcareInnovation #DataAnalytics #GrowthJourney

6

I'm thrilled to announce a new preprint describing collaborative work with Ajay Jain and Ann Cleves Jain, "Deep-Learning Based Docking Methods: Fair Comparisons to Conventional Docking Workflows". It's commonly believed that deep learning-based docking programs, like DiffDock, surpass traditional physics-based methods in performance. However, our findings challenge this notion by addressing fundamental limitations in earlier evaluations. When we removed the training set bias, fully automated evaluations using Glide and Surflex-Dock outperformed DiffDock. In addition, AutoDock Vina and Gnina demonstrated competitive results without leveraging extensive pre-trained data. Paper: https://lnkd.in/ehjXfrgn

633

45 Comments

In this post, I asked ChatGPT to compare commercial and free open-source data analytics products. ChatGPT's response reinforces that choosing the best product depends on the current infrastructure or product being used and needs, resources, skills, and budget.

3

Please read my interview: How can we build a solid analytics learning culture at Fortune 100 companies? Please read my conversation with Xingchu Liu, Ph.D. Liu, Former Chief Commercial Analytics & AI Officer at Pfizer Inc., on Published 11/20/2024

3

This is a really interesting table from a recent Nature paper - it summarizes very well the situation regarding the adoption of #AI in the research workflow: 📌 regardless which steps you look at, the industry is still experimenting (#Pilot), very few tools have been adopted in routine 📌 there is still much progress to be made (and I guess also a lot of skepticism in the scientific community). Now, one question i kept asking myself: with the pace of progress in AI, and the number of researcher papers (relevant to life science research) coming out every week, i'm wondering how organizations are managing to evaluate those new models / tools (or they don't ... and they might be missing out on something important)? I guess it's time to adopt more "Agile" research methods in Life Sciences too. This also requires adapting organizations to bring data scientists & tech closer to SMEs. Thanks Patrick Loerch for sharing! #AgilePharma

9

2 Comments

Are scientists inventing their AI replacement? https://lnkd.in/gZ6fkpd7 interesting ready - attempting to automate research itself. An end-to-end framework designed to automate the generation of research ideas, implementation, and the production of research papers.

46

A new #GenAI report from McKinsey & Company One of the first highlights being “Gen AI adoption is most common in the functions where it can create the most value” [… as opposed to where it brings the least value 🤔🤔🤔??? Not sure I needed McKinsey for this insight 🤣 … but keep reading it does get better after] Unsurprisingly #Pharma is the industry with the smallest investment in GenAI, with 61% of the companies investing less than 5% of their digital budget In term of financial impact, most organisations observe cost reduction in HR and increased revenues in supply chain management “Inaccuracy” is the most recognized and experienced risk of gen AI use, #IP security and #cybersecurity are close 2nd and 3rd risks - 25% if the respondents reported negative impact of inaccuracy on their organisation In term of governance, only 18% of the respondents reported that their organizations had “an enterprise-wide council or board with the authority to make decisions involving responsible AI governance” Here is the link to the report: https://lnkd.in/gmgSH9ZS

11

As I explained in my last post, Machine Learning models are "trained" on historical data. However, the "extent" to which these models have to be trained is not an exact science. Let me explain 2 issues that have to be taken care off, with a healthcare related analogy, for folks with a life-science background to understand - 1. Underfitting: The One-Size-Fits-None Scrubs Imagine a hospital where all the doctors and nurses have to wear the same set of scrubs, regardless of their body size or shape. Now, these scrubs are so loose that they fall off easily and offer little to no protection. This situation is like underfitting in machine learning. Just like those scrubs don't fit anyone properly, an underfit model doesn't capture the complexities of the data, resulting in poor performance. 2. Overfitting: The Super Specialized Stethoscope Picture a doctor who insists on using an incredibly specialized stethoscope for every patient, claiming it can diagnose any ailment. This stethoscope has so many attachments and features that it can pick up even the tiniest of anomalies. However, because it's so specialized, it often detects "issues" that aren't really there, leading to unnecessary treatments. This scenario mirrors overfitting in machine learning, where a model becomes overly complex and fits too closely to the training data, capturing noise as if it were signal. If there is a machine learning product that you use, and you notice errors from the product that is either missing findings that it is supposed to identify, or mistakenly picks up an artifact as a finding, it may be due to underfitting or overfitting. #machinelearning #aiforhealth #healthtech

6

Lesson Learned from a oncology DCT trial: Reference: https://lnkd.in/gyF9U-RZ Challenges in DCT Implementation Technological Integration: Lack of integration between DCT platforms and hospital systems can hinder trial conduct and increase operational complexity. Patient Referral and Management: Clear delineation of responsibilities between investigators and treating physicians, as well as effective communication with patients, is crucial for successful patient referral and management in DCTs. Recruitment Strategies: While digital recruitment holds potential, the TELEPIK trial's reliance on traditional methods limited the evaluation of its impact on enrollment rates. Lessons Learned from TELEPIK DCT Potential: DCTs offer advantages in participant representation, data collection efficiency, and reduced patient burden. Optimization Needed: Despite these benefits, the early termination of TELEPIK highlights the need for refined DCT design and execution methodologies. Knowledge Gap: Limited evidence on DCT performance compared to traditional trials necessitates further research to inform future trials. Recommendations Comprehensive Evaluation: Conduct thorough post-trial analyses to identify critical factors influencing DCT success or failure. Technological Advancements: Prioritize the development of integrated technological solutions to streamline DCT operations. Collaborative Approach: Foster collaboration between sponsors, investigators, and healthcare providers to address challenges and optimize DCT design. Patient-Centric Focus: Continuously evaluate DCTs from the patient perspective to ensure positive experiences and enhance patient engagement. Rigorous Evaluation: Conduct comparative studies to assess the effectiveness of DCTs versus traditional trials in terms of recruitment, retention, and cost. #clinicaltrial #clinicalresearch #data #healthcare #patient #dct

16

🚀 "Knowledge graph with GenAI interface at AstraZeneca": insights from Mathew Woodwark ... ... ... There’s no denying the buzz around AI in drug discovery, but where are we in terms of trust and adoption? 🔍 Early Pipeline? A Definite Yes! At the early stages, AI shows tremendous promise. It’s exciting, and scientists are asking: Can we use it for this? And that? ⚙️ Later Pipeline? Proceed with Caution While there’s enthusiasm, adoption is cautious—checks, balances, and human oversight are crucial. Why? Trust isn’t fully built yet, and challenges like fragmented data sources remain a hurdle. 📊 Knowledge Graphs & Gen AI Interfaces: A Glimpse of the Future One organization tackled this by building a Knowledge Graph for Discovery and layering it with a generative AI interface. The result? It’s incredibly popular, offering rich insights—but only from public domain data. The next step? Integrating internal data to unlock full potential. 💡 The Missing Link: Measuring Impact While these tools are delivering great results, the real question is: How are these outcomes influencing decisions? Tracking this impact will be the game-changer in validating AI’s value in the drug R&D pipeline. 🌟 Big Takeaway: We’re in the phase of cautious optimism—no internal disasters, but also no monumental breakthroughs. The journey to fully embedding AI into R&D pipelines will hinge on trust, rigorous processes, and ensuring data is ready to fuel these models. Link to the video in the comment section 👇 ________ Love ❤️ this post? Add your comments 💭, re-share ♻️ it with your network, follow me 🔔 for more posts like this, and DM me 📩 if you want to deep dive on this topic!

12

1 Comment

Did you know that there is an alliance of AI companies working in healthcare with over 40 members helping shape the industry? Check out the interview with the chairman of the The Alliance for Artificial Intelligence in Healthcare (AAIH) Mustaqhusain Kazi during the World Economic Forum event in San Francisco. 💡📈🤖

7