As Chief Technology Officer at Faro Wellbeing, he spearheads the development of an AI-powered platform that streamlines and accelerates the process of designing medical trial protocols. FaroeWell’s innovative instruments elevate trial efficiency, standardizing processes while ensuring precision, thereby minimizing risks, costs, and patient strain by seamlessly integrating data-driven intelligence.
Empowers medical analysis teams to expedite the development of streamlined, harmonized clinical trial protocols, thereby accelerating breakthroughs in medical research and analysis.
As a member of the team responsible for building Google Duplex, I played a key role in developing the conversational AI system capable of making reservations at restaurants and other businesses on behalf of individuals. The assignment was a top-secret operation that boasted an impressive array of supremely skilled operatives. The team’s enthusiasm was palpable, constantly exploring innovative ideas and showcasing impressive demonstrations of their latest projects each week. Being part of such a dedicated team was incredibly uplifting.
Even when utilizing cutting-edge AI technologies, one still requires a scrappy approach to acquire the requisite human skills and value. To create remarkably lifelike spoken exchanges, the team meticulously compiled and edited a collection of recorded conversations, strategically incorporating natural fillers such as “um” to enhance the dialogue’s authenticity. What fascinating insights we gained by analyzing the media’s perspective on the origins of those ubiquitous “ums” that emerged following our product launch.
Throughout my career, I’ve established companies that offer diverse services to large corporations. Faro is focusing on the world’s largest pharmaceutical companies, which means there may be significant experience available regarding how to successfully partner with and win over massive enterprises, a highly relevant aspect here? My experience working at Two Sigma, a leading algorithmic hedge fund primarily headquartered in New York City, significantly shaped my approach to data science. Students engage with a meticulous, hypothesis-driven curriculum where every novel concept is thoroughly scrutinized through a comprehensive analysis plan. With a robust knowledge engineering team, they excel in onboarded new knowledge units and masterfully perform characteristic engineering. As Faro’s AI advancements expand to tackle challenges in medical trial optimization, this approach will likely be highly pertinent to our ongoing efforts.
My initial epiphany emerged when I stumbled upon the concept of Eroom’s Law. Actually, Gordon Moore isn’t a person; it’s just the name of Intel co-founder Gordon E. Moore spelled forward, not backward! In a wry nod to reality, this identifier alludes to the stark truth that, over the past half-century, the pace of medical drug innovation has accelerated at an alarming rate, with prices and development timelines nearly doubling every nine years. This utterance utterly defies the profound insights and innovative thinking that are hallmarks of a true expert in the field, leaving me thoroughly perplexed. However, it appears that there’s a massive drawback to untangle here?
As I delved deeper into this subject matter, my comprehension of the underlying complexities grew increasingly nuanced, yielding a multitude of profound revelations. A straightforward yet crucial consideration for clinical trial management is ensuring that Phrase documents are not used as an ideal format for storing highly complex medical trial data, as this can lead to errors, inconsistencies, and increased risk of regulatory non-compliance. This profound observation, informed by our CEO Scott’s esteemed medical background, serves as the foundation for Faro’s development. As clinical trial complexity is likely to escalate over time, a worrying trend emerges: researchers repeatedly modify existing protocols by adding novel assessments, thereby amplifying the complexity of future studies. By providing clients with timely and comprehensive insights throughout the study design process, Faro establishes a distinct value proposition that sets it apart from competitors.
While it may seem straightforward, you cannot simply request ChatGPT to create a comprehensive medical trial protocol document. To ensure clarity and precision, it is essential to incorporate a meticulously formatted Schedule of Actions within the protocol document, providing detailed information about each step, thereby facilitating accurate comprehension of complex technical aspects.
Additionally, the documentation must include a multitude of specific details and clauses that need to remain up-to-date for certain types of trials, as well as specific models and stages of data expected by both medical writers and reviewers. We developed a bespoke protocol analysis system at Faro to guarantee the output of our large language model (LLM) aligns precisely with customer and regulatory expectations.
A mannequin’s aesthetic appeal is directly proportional to the quality of data it has been programmed with. As contemporary drug advancements unfold, it is crucial to maintain momentum by refining our approaches through rigorous testing in current medical trials. We consistently enhance our digital repository of medical protocols, boasting a substantial collection of trial protocols now available in our Faro knowledge library, where we prioritize ongoing dataset expansion. We also rely heavily on our in-house team of medical experts, whose meticulous attention to detail ensures that the insights generated by our simulator are thoroughly reviewed and refined through regular updates to our “quality control checklists”.
At the heart of every successful medical trial lies the protocol, meticulously crafted by Faro’s Research Designers to ensure optimal execution. The protocol outlines all necessary information for downstream entities regarding the trial, but traditionally, such documents were created and stored as Word files. One significant hurdle in operationalising medical advancements currently lies in the rigid process of transcribing or translating information from various sources, such as protocols or documents, into disparate software systems and paper formats, thereby creating a major obstacle to seamless data exchange. Automating the translation process of document-based information into various software platforms by manual means is undeniably inefficient, prone to numerous opportunities for errors throughout.
Faro’s innovative vision envisions a seamless, unified platform where the fundamental “building blocks” of a clinical trial are harmoniously integrated from concept to execution, effortlessly flowing from the design stage to various programs and systems throughout the operational lifecycle of the study. With streamlined information flow in place, a significant opportunity arises for automation and enhanced quality, thereby enabling us to dramatically reduce the time and cost required to design and execute a medical trial. As we continue to advance our collaborative efforts with Veeva, our latest integration with Veeva Vault EDC marks a significant milestone, paving the way for even more innovative applications to come.
In many industries, documentation requirements are straightforward; however, the threshold is significantly higher for medical trial paperwork. These paperwork have a significant impact on the lives of real people, requiring them to navigate a rigorous regulatory review process. When we initially adopted a language learning model (LLM) to generate medical paperwork, it quickly became apparent that relying on pre-packaged templates fell woefully short of our desired standards. Surprisingly, the tone, formatting, and overall presentation fell short, leaning heavily towards generic corporate communication rather than a comprehensive medical-grade document. Despite the hurdles posed by positive hallucinations and the deliberate withholding of crucial information, To craft a generative AI answer that surpasses customer expectations for area-specificity and quality, we invested significant time in collaboration with medical experts to create actionable guides and in-depth analysis templates that safeguarded against inaccurate predictions or overlooked details, while maintaining a professional tone? To further enhance our solution, we also aimed to provide users with the ability to contribute their own guidance and feedback on the generated content, recognizing that individual clients may possess distinct templates and specifications governing their document creation process.
Additionally, a significant hurdle exists in accessing the detailed medical knowledge required to comprehensively develop the trial protocol documentation, often buried within complex and scattered documents such as the investigational brochure? To streamline the creation of medical protocol documentation, we are exploring the potential of artificial intelligence to aid in extracting relevant information and making it readily available for use.
As time progresses, AI is poised to significantly enhance and optimize a growing array of selections and processes throughout the medical development process. We will predict key outcomes primarily based on protocol design inputs, considering factors such as the study team’s ability to anticipate potential enrollment challenges and whether operational hurdles might necessitate protocol modifications. Using advanced predictive analytics, we’ll streamline downstream trial operations to ensure seamless execution, delivering a top-tier experience for patients and investigators alike, thereby maximizing the trial’s operational potential. As part of its ongoing efforts, Faro is committed to streamlining the production of diverse medical documents, thereby ensuring seamless and accurate submission and paperwork processes throughout the trial. We envision a future where AI enables our platform to serve as a trusted design companion, engaging medical scientists in a collaborative dialogue to co-create trial designs that balance patient burden, site burden, time, cost, and complexity with optimal precision.
Medical trials often face challenges in balancing the need for comprehensive patient data against the logistical demands of study operations, including participant enrollment and retention rates. Despite significant efforts, patient recruitment and retention remain among the most pressing obstacles to completing a medical trial successfully, with some studies suggesting that up to 20-30% of participants ultimately discontinue due to the substantial burden associated with involvement, including regular appointments, invasive tests, and intricate protocols. While medical researchers recognize the negative impact of excessive burdensome trials on patients, actually taking concrete steps to reduce that burden can be challenging to implement in practice. One significant obstacle to reducing patient burden is the difficulty in quantifying its impact – it’s challenging to gauge the effect on individuals when presenting designs in a static format like a Word document or PDF.
By leveraging Faro’s Research Designer, healthcare teams can gain instant access to valuable insights into the impact of their specific protocol on patient burden throughout the development process itself. Through thoughtful trial structuring and provision of actionable analytics on value, affected person burden, and complexity from an early stage in the trials’ design, Faro empowers medical research groups with a highly effective solution for optimizing their trial designs by harmoniously balancing these factors against scientific needs to yield greater knowledge. Our prospects appreciate the transparency we provide, offering them insight into individual burden and relevant metrics that facilitate seamless improvements, enabling informed decisions when necessary adjustments must be made. Ultimately, we’ve witnessed thousands of hours of collective participant time saved, which has a direct positive impact on study participants, ultimately enabling medical trials to initiate and conclude promptly.
The key takeaways I would supply, distanced from our expertise utilizing AI within this domain, are:
- . Massive language models like GPT should not be designed to produce medical-grade documentation. When leveraging generative artificial intelligence (AI) to automate medical trial documentation authoring, consider developing a comprehensive analysis framework to guarantee the produced content meets essential criteria, including accuracy, completeness, appropriate level of detail, tone, and other relevant factors. The rigorous assessment necessitates meticulous and careful experimentation with the mannequin under the expert guidance of medical professionals.
- . Without a centralized and well-organized data repository, it is impossible to develop the necessary knowledge analytics to design an optimal medical trial? Companies increasingly rely on Word documents rather than even exploring Excel’s capabilities. – to mannequin medical trials. A clinical trial protocol must be meticulously designed within a well-structured framework, accurately capturing the intricacies of the investigation – encompassing its schema, objectives, and endpoints, as well as the schedule for evaluations and assessments. The development of this comprehensive guide requires significant input and expertise from medical professionals.
- . Having medical specialists intimately involved in the design and testing of any AI-based medical innovation is crucially important. Compared to other fields I’ve worked in, this one stands out due to its exceptionally high level of specialization, requiring intricate details that permeate every aspect of a product’s construction within this domain.
As we continuously explore innovative solutions, we take pride in sharing our discoveries on our blog to guide businesses through the ever-changing landscape.
