Let's create the Book of Life!

What if we could securely exchange our health data while respecting privacy laws? As a triathlete, I rely on the Strava app, which collects all the facets of my training journey, including heart rate, and this app remains a benchmark for me in its data management, up to the implementation of the right to be forgotten. But my dream transcends the world of fitness, and I envision a singular repository for all my health data – a digital haven encompassing medical records, laboratory insights, diagnostic imagery, and the profound nuances of my genetic blueprint. In this dream, I see a new era in healthcare, empowering me to share my health data with a physician during a moment of consultation. I could extend my consent to innovative research projects, uniting my data with the collective wisdom of others. This could also revel the ability to compare myself against peers of my generation and to understand the impact of my life style on my health.

Yet my dreams don't end there; I could see pharmaceutical industries sharing more of?their non competitive data and hospitals sharing more de-identified EHR data with research laboratories. It's evident that we would better understand the progression of diseases, accelerate research, have better outcomes with new treatments, and have shorter and less expensive clinical trials, for example, by reusing the control arm of a clinical trial produced by another lab or by developing novel biomarkers. Of course, this must be done while preserving data privacy and respecting patient consent, and protecting the intellectual property of pharmaceutical companies.

??I'm sure that if we raise awareness on health data sharing and work on the necessary frameworks, we could collectively create the book of life!??

Health data in the news

Bridging the Data Sharing Divide: Unlocking the Future of Research and Healthcare"

Governments worldwide are actively working to familiarize their populations with the idea that sharing private health data can significantly boost research efforts.

Drawing inspiration from the UK's approach and the ongoing conversation surrounding data sharing, as seen in discussions with organizations like UK Biobank, highlighted in articles such as the Guardian's "Half a million of us want to donate our data to British science, but it's languishing unused due to privacy concern," a paper in the Hong Kong Medical Journal suggests that the sharing of anonymized and aggregated health data with private enterprises can play a pivotal role in advancing medical research and accelerating the development of new medications in Hong Kong.

The pharmaceutical industry finds itself at a crossroads of data sharing, with recent investigations having unveiled a startling reality: Over the past decade, more than half of the clinical trials supporting the FDA approval of 115 anticancer medicines were deemed ineligible for participant-level data sharing. Furthermore, investigations indicate that much of the participant-level data underpinning the FDA/EMA approval of COVID-19 vaccines is currently out of scope for request and will likely remain so for some time. This issue is underscored in "A 10-year update to the principles for clinical trial data sharing by pharmaceutical companies: perspectives based on a decade of literature and policies". This is a must read report on clinical datasharing making a status on 2013 data sharing commitments and emphasizing the need for continued improvement in data sharing practices in the pharmaceutical industry. This groups of scientist propose updates to data sharing principles aim to enhance the data sharing ecosystem, which is crucial for scientific progress and patient-centered care. Key recommendations include sharing participant-level data, making full Clinical Study Reports available, sharing protocols and Statistical Analysis Plans, providing summaries to participants, and maintaining clear public data sharing policies.

Sharing health data poses the challenge of establishing a clear framework that ensures data privacy while preserving data utility for potential reuse, a complexity particularly pronounced in the clinical research industry. It's as if the call from these scientists has been answered: the TransCelerate consortium and its preclinical subsidiary, BioCelerate, have introduced a novel methodology entitled "Clinical Data Sharing: A Proposed Methodology to Enable Data Privacy while Enhancing Secondary Use." The culmination of a three-year collaboration involving participation from over 20 global sponsors and substantial input from industry stakeholders, this comprehensive guidebook outlines how sponsors can enhance the sharing and repurposing of clinical data for secondary research. The best practices it outlines are designed to expedite scientific discoveries and alleviate the burdens on research participants, thereby enabling the industry to achieve these objectives. You can find a summary of these recommendations in the clinical leader article.

There are some solutions that facilitate data sharing through data trust while upholding privacy, and one notable example is Decentriq. This week, Decentriq announced its support for the new international consortium, iCARE4CVD composed of 20+ participants, which is dedicated to enhancing cardiovascular disease (CVD) outcomes through the utilization of AI and extensive datasets. Decentriq has also announced a partnership with Datavant to enable collaboration between health researchers and European hospitals.

Another example comes from Amazon Web Services who has developed and launched a new platform designed to help federal government customers integrate data across multiple systems, provide data access and facilitate the decision-making process in the cloud while building up security through the adoption of zero trust principles.

Synthetic Data Realities: Benefits, Challenges, and Opportunities in Healthcare

The Synthetic Data Summit 2023, held recently, shed light on the remarkable progress and wide-ranging applications of synthetic health data in the healthcare industry. Distinguished experts and thought leaders from various domains came together to discuss the potential benefits and challenges of using synthetic data for research, innovation, and patient care. Synthetic health data refers to artificially generated data that mimics real-world data while protecting individual privacy and confidentiality. Synthetic data can provide a cost-effective and privacy-preserving alternative to real patient data, enabling researchers and innovators to work with large, diverse, and representative datasets without compromising privacy or regulatory compliance.The event highlighted various domains where synthetic data can have a significant impact, including clinical research, drug development, predictive modeling, and precision medicine. Ensuring the quality, representativeness, and utility of synthetic data sets remains a critical task. Effective validation and evaluation frameworks are needed to establish the reliability and credibility of synthetic data for various applications.

FDA's View on Real-World Data and Real-World Evidence

FDA leaders have clarified the differences between Real-World Data (RWD) and Real-World Evidence (RWE) in a new paper. RWD refers to the data collected from a variety of sources, while RWE is the insights derived from the analysis of RWD. FDA has encountered a misconception that only non-interventional (observational) research utilizes RWD to generate RWE—in other words, a dichotomy of randomized controlled trials versus real-world evidence is said to exist. In reality, the spectrum of study design involves various combinations of data sources and design architectures; For example, externally controlled trials that utilize RWD in the comparator arm generate RWE, despite the treatment arm generating data according to a study protocol in a clinical trial environment. As another example, a randomized trial generates RWE if the primary outcome is based on an assessment of RWD (often referred to as a point-of-care trial).

Continuing to Rise Concerns in Healthcare Cybersecurity, Data Breaches and Vulnerabilities

Connecting the dots between the recent surge in data breaches and growing cybersecurity concerns in the healthcare industry, it's worth noting that the same hacker who leaked a trove of user data stolen from the genetic testing company 23andMe two weeks ago has now leaked millions of new user records. A new report predicts a significant increase in data breaches within the healthcare sector by 2023, emphasizing the rising complexity and sophistication of cyberattacks targeting healthcare organizations. These predictions are grounded in various factors, including the increasing black market value of healthcare data, the proliferation of interconnected devices and systems in healthcare, and the escalating sophistication of cyberattacks within the industry.

Furthermore, in response to the growing vulnerabilities in the healthcare sector, the Cybersecurity and Infrastructure Security Agency (CISA) has released an advisory, ICSMA-23-285-01, regarding vulnerabilities found in medical devices and industrial control systems (ICS). These vulnerabilities encompass a wide range of medical devices and ICS components from various manufacturers and models, posing threats like unauthorized access, arbitrary code execution, and potential denial-of-service scenarios. In light of these risks, CISA recommends organizations review the advisory, apply necessary security patches and mitigations provided by manufacturers, and adopt best practices like network segmentation and robust password protocols to bolster their defenses against potential cyber threats.

AI in Life Sciences and Healthcare

It does not come as a surprise that AI, machine learning come as top health CIO priorities in 2023 and that CIOs see Generative AI has the potential to revolutionize healthcare.

We need to to address patient concerns and build trust in AI technologies

There is still a huge apprehension surrounding AI's involvement in healthcare decision-making. A survey conducted by Carta, a digital health benefits platform, uncovered a startling statistic: only 39% of respondents trust AI to make decisions about their healthcare. Even more telling, a staggering 70% of those surveyed would not trust AI for personalized medical advice. Another survey revealed that approximately 54% of patients trust AI technologies in their healthcare journey. The remaining 46%, however, do not share the same confidence in AI. Interestingly, the study demonstrated that younger patients and those with chronic conditions were more likely to trust AI, while older patients and those without chronic conditions exhibited lower levels of trust. This trust will not be improved by the recent study published in Nature raising concerns about AI perpetuating racial bias in medicine. The study evaluated the responses of four commercially available large language models (LLMs) to scenarios related to race-based medicine and widespread misconceptions. The findings were concerning: all models exhibited examples of propagating race-based medicine in their responses. Moreover, the models were not consistently accurate in their responses when posed with the same questions repeatedly. These findings underscore the critical importance of addressing bias in AI algorithms, particularly in the healthcare context.

Recognizing the significance of AI in healthcare, regulatory bodies have stepped in to address concerns and set guidelines. The U.S. Food and Drug Administration (FDA) recently established the Digital Health Advisory Committee to delve into the scientific and technical aspects of digital health technologies, including artificial intelligence and machine learning. This move underscores the FDA's commitment to understanding and overseeing the role of AI in healthcare. The World Health Organization (WHO) has also taken steps to outline considerations for the regulation of AI in the health sector. Their framework emphasizes the importance of transparency, documentation, and ethical use that promotes equity, inclusiveness, accountability, privacy, and confidentiality. These principles are crucial in building patient trust and ensuring AI's responsible integration into healthcare.

According to a new study published in Nature, researchers assessed whether four commercially available large language models (LLMs) propagate harmful, inaccurate, race-based content when responding to eight different scenarios that check for race-based medicine or widespread misconceptions around race. They found out that Artificial intelligence (AI) systems used in healthcare could potentially perpetuate and intensify racial bias in medicine. The study highlights the need to address bias in AI algorithms to ensure equitable healthcare outcomes. All models had examples of perpetuating race-based medicine in their responses. Models were not always consistent in their responses when asked the same question repeatedly. LLMs are being proposed for use in the healthcare setting, with some models already connecting to electronic health record systems. However, this study shows that based on our findings, these LLMs could potentially cause harm by perpetuating debunked, racist ideas.

AI-Driven Protein Design: A Game-Changer for Healthcare and Beyond

?Researchers have used AI to design proteins that do not exist in nature but still perform valuable functions. A recent study published in the Proceedings of the National Academy of Sciences reveals a groundbreaking new AI algorithm that can accurately predict the 3D structure of proteins. This development has significant implications for drug discovery, personalized medicine, and bioengineering.

Quantum computing can also enhance protein design by providing faster, more accurate predictions of protein structures, paving the way for breakthroughs in healthcare, drug development, and our understanding of various diseases. As an example of use of quantum computing, Cleveland Clinic, in partnership with IBM Quantum and Algorithmiq, has secured two contracts through Wellcome Leap's Quantum for Bio Challenge. The first project focuses on improving protein structure prediction using quantum algorithms, potentially advancing disease understanding and therapy development. The second project aims to explore quantum computing's role in developing photon-activated drugs for cancer treatment.

AI in clinical trials

The article "Digitally Transforming Clinical Trials" explores the advantages of incorporating digital tools and platforms into the clinical trial process. It emphasizes that by leveraging technology, clinical trial sponsors can optimize processes, enhance efficiency, and accelerate the development of innovative treatments. The article further explains how digitalizing various aspects of clinical trials, such as patient recruitment, data collection, and monitoring, can lead to improved patient experiences and more reliable data. Additionally, the article highlights the potential of wearable devices and mobile applications in facilitating remote patient monitoring and data collection, thereby reducing the burden on patients and increasing the convenience of participating in clinical trials. Overall, the article demonstrates the transformative power of digital technologies in revolutionizing the field of clinical trials, ultimately benefiting patients and advancing medical research.

And if you want to know more about past clinical trials, TrialAssure Uses Artificial Intelligence to Create Plain Language Summaries for 50,000 Studies Listed on ClinicalTrials.Gov: TrialAssure, an AI-technology leader in advancing clinical trial results disclosure and data sharing in the pharmaceutical industry, announced today that it has taken the proactive effort to create and publish clinical trial results in plain language summary (PLS) format, posting directly to TrialResults.com

AI Transforming Healthcare

??Digital health solutions are enabling early detection and screening of cancer, improving patient outcomes and survival rates

??Delfi Diagnostics has introduced a blood test that employs genetic and epigenetic markers to detect lung cancer at earlier stages, uncovering cases that conventional imaging methods may overlook. This non-invasive screening test has the potential to improve survival rates and lead to better treatment outcomes.

??Technology advancements in breast cancer screening, such as 3D mammography and digital breast tomosynthesis, have greatly enhanced early detection rates and improved outcomes for patients. These innovative screening methods offer more accurate and detailed images, enabling doctors to identify breast abnormalities with greater precision.

??AI has achieved 100% accuracy at identifying melanoma. The same software detected 99.5% of all skin cancers and 92.5% of precancerous lesions.

??692 FDA-approved Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices . As of Oct. 19, 171 Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices were added to the list. FDA highlights as well radiology's continued dominance in AI-enabled device submissions

??AI detects signs of AFib in asymptomatic patients: Advanced AI models have demonstrated the capability to detect abnormal heart rhythms in patients even before they exhibit symptoms, according to research published in JAMA Cardiology.

??A new AI tool developed by researchers at the University of Pittsburgh can detect signs of distress in overburdened hospital workers, potentially offering vital support to those experiencing burnout in the healthcare setting

??While these initiatives are encouraging, implementing such models in a hospital environment remains a significant challenge. A recent study conducted by the Icahn School of Medicine and the University of Michigan highlighted the impact of implementing predictive models on the deterioration of model performance. Their findings revealed that using these models to adjust care delivery can alter the underlying assumptions on which the models were originally trained, often with negative consequences. This study, which simulated critical care scenarios at two major healthcare institutions, the Mount Sinai Health System in New York and Beth Israel Deaconess Medical Center in Boston, analyzed 130,000 critical care admissions. The results were published in the online issue of Annals of Internal Medicine on October 9.

Our findings reinforce the complexities and challenges of maintaining predictive model performance in active clinical use," says co-senior author Karandeep Singh, MD, Associate Professor of Learning Health Sciences, Internal Medicine, Urology, and Information at the University of Michigan. "Model performance can fall dramatically if patient populations change in their makeup. However, agreed-upon corrective measures may fall apart completely if we do not pay attention to what the models are doing—or more properly, what they are learning from."

If we can't fix those issues for predictive models, how can we dream about implementing LMM in a clinical setting. During an Oct. 25 National Academy of Medicine Workshop on Generative AI and Large Language Models in Health and Medicine, health system executives and other stakeholders spoke about the governance, regulation and deployment issues they are grappling with. Nigam Shah, M.B.B.S., Ph.D., professor of Medicine at Stanford University, and chief data scientist for Stanford Health Care, confirmed that there have been hundreds of predictive models developed for population health, readmissions predictions, and sepsis predictions. “Often we don't have the policies and the work capacity designs set up correctly to achieve the promised usefulness that we could have gotten,” he said. “The risk I see is that we didn't get it right for the traditional or regular AI. What are we doing as a community to ensure that our response to generative AI will be better? There are numerous challenges that LLM's based solution will have to solve such as how generative AI models are trained and how they are used making both privacy and other regulators particularly nervous, as highlighted in an article By Kirk Nahra and Ali Jessani on Privacy Concerns at the Intersection of Generative AI and Healthcare

Cutting-Edge Platforms in Clinical Data and Hospitals

Innovative Trends in Clinical Data and Drug Development: Transforming the Landscape

Driven by the need for speed and efficiency, the clinical data technology landscape is also transforming. There's a growing necessity to eliminate fragmentation and silos, fast-tracking this evolution. To future-proof clinical data infrastructure, a modern approach is required. Enter the data lakehouse architecture, capable of managing the vast scale and complexity of clinical data. It doesn't stop there; it also provides essential governance, security, and data management features to meet stringent regulatory requirements. This architecture pattern is gaining momentum, with several pharmaceutical companies embracing it today.

Specialized solutions are also making headlines. Saama, for instance, has forged a new collaboration with AstraZeneca. This partnership aims to develop AI-driven solutions for drug development, using Saama's Life Science Analytics Cloud (LSAC) platform. Leveraging advanced machine learning and analytics, the platform will scrutinize AstraZeneca's data and construct predictive models for drug development. The ultimate goal? Accelerating drug discovery and development by pinpointing crucial data patterns and insights that optimize clinical trials and enhance patient outcomes.

Meanwhile, Medidata is expanding its partnership with Catalyst Clinical Research, a full-service Contract Research Organization (CRO). The renewed collaboration seeks to elevate oncology clinical trials by harnessing advanced technology for greater study efficiency and improved data quality.

In a separate vein, Verily, an Alphabet subsidiary, is making waves in the healthcare realm. Verily's dynamic work lies at the intersection of healthcare, data, and technology. Their mission is to enhance care delivery, revolutionize research methods, and ultimately improve patient outcomes. With a focus on disease detection, population health management, and drug discovery, Verily deploys cutting-edge technologies like artificial intelligence, machine learning, and data analytics. Through this innovative approach, Verily is poised to address some of the healthcare industry's most pressing challenges.

Lastly, BC Platforms and NTT Research have joined forces to expedite the adoption of data-driven medicine in Japan. This collaboration will harness BC Platforms' genomic data management and analysis solutions, coupled with NTT's telecom and IT expertise, to bolster precision medicine initiatives in the country.

An integrated platform for hospitals

Philips has unveiled an integrated information platform for hospitals that combines clinical, operational, and financial data in real-time, enabling healthcare providers to make more informed decisions and improve patient outcome

Engineering trends

Tuning RAG in a production environment

With the rise of LLMs, the Retrieval Augmented Generation (RAG) framework also gained popularity by making it possible to build question-answering systems over data. RAG involves multiple components, including document loaders, splitters, embedding models, vector databases, prompts, and large language models (LLMs). Ahmed Besbes provides valuable insights for those working with RAG systems, emphasizing the importance of careful tuning and optimization for reliable performance in real-world applications.

Data Mesh, a socio-technical paradigm

Data Mesh is currently a trending and widely discussed concept among service companies. A notable reference book in this domain is 'Data Mesh — Delivering Data-Driven Value at Scale' by Zhamak Dehghani. Additionally, I highly recommend reading a comprehensive book review authored by Venkataraman Balasubramanian , with whom I've shared a professional relationship spanning three decades. This insightful review can provide valuable insights, and if you wish to delve deeper into this essential subject, you may consider ordering the book.

Data engineering at Meta

Data engineering is an art in itself. If you're eager to delve into the intricacies of data engineering, especially in the context of managing exabyte-scale data, we invite you to gain an insider's perspective on Meta's data engineering practices. Discover how our data engineers expertly navigate this multifaceted landscape, fine-tune their processes, and guarantee a seamless flow of data.

Upcoming Conferences

• 6th GCC eHealth Workforce Development Conference, Dubai, October 31st to November 2nd
• DigiHealthDay - 2023 Organized by Deggendorf Institute of Technology,Pfarrkirchen, November 10th
• Digital Health World Congress 2023, London, UK, Nov 29th and 30th
• Digital Health Conference, Berlin, GERMANY, Nov 30th
• AI for Life, Geneva, December 7th

Other Great Newsletter on healthcare and AI

Stay up to date with the latest insights and industry trends by subscribing to these informative newsletters from my esteemed colleagues and professionals in the field.

• Healthcare in the Metaverse?by?Philippe GERWILL
• DigitalhealthReadingList?by?Wolfgang Schleifer
• Future Tech Trends?by?Bernard Marr
• All about Blockchain Adoption?by?Kamlesh Nagware
• AI Frontier?by?Steve Nouri