How To Achieve Remarkable Results in Oncology Clinical Trials

In this insightful interview, I spoke with Craig Serra, Head, Clinical Research Scientific and Technical Engagement at Flatiron Health, about the innovative ways the company is revolutionizing oncology clinical trials by integrating real-world data (RWD) and real-world evidence (RWE). Flatiron Health’s pioneering work in electronic health records (EHR) and data curation has made significant strides in enhancing oncology clinical trials, protocol optimization, site identification, and patient matching. Our discussion covered various aspects of Flatiron’s approach, including using advanced technologies and strategic partnerships to improve oncology clinical trials processes and patient outcomes.

Moe: What were the core historical businesses of Flatiron Health, and how did they impact oncology clinical trials?

Craig Serra: Flatiron Health initially focused on two core businesses: deploying an EHR system (OncoEMR^®) to community oncology practices and curating data from these practices into malignancy-level datasets for biopharma sponsors. This dual approach allowed us to serve both healthcare providers and pharmaceutical companies effectively. We helped oncology practices streamline their operations, improve patient care, and enhance their site workflow capabilities by providing an oncology-centric EHR. OncoEMR^® is an intuitive oncology EMR software platform that enables physicians to access detailed patient information, track treatment progress, and make informed clinical decisions, which is crucial for oncology clinical trials. What makes this product unique is its’ development: built by community oncologists who understand the detailed elements of care and intended for use by those community oncologists.

Moe: How has acquiring Protocol First and the Clinical Pipe EHR to EDC connector transformed your clinical research approach?

Craig Serra: The acquisition of Protocol First with subsequent expansion of our Clinical Pipe EHR to EDC connector have been pivotal for our clinical research approach. Clinical Pipe, a SMART on FHIR application, enables seamless data transfer from EHRs to EDC systems, significantly enhancing operational efficiency. This acquisition allowed us to dramatically scale Clinical Pipe to community sites using OncoEMR^® and add to the site footprint beyond Epic and Cerner sites (typically academic medical centers and health systems).

We’ve also been able to apply our deep expertise and pair our knowledge of site workflows with our industry-leading data curation infrastructure to process certain additional structured data that is often difficult to get as well as notoriously difficult-to-acquire unstructured data. We’ve remarkably achieved an ability to transfer up to 100% of EHR-sourced data to an EDC system. There are obviously a host of benefits to multiple stakeholders and users, and by ensuring transformative data flow from point A to point B, Clinical Pipe has streamlined sponsor data management and monitoring processes, making it easier for sponsors and sites to run oncology clinical trials.

Moreover, while Clinical Pipe is not currently used for patient matching, its ability to facilitate data transfer has other substantial benefits. It allows for real-time data integration, improving the accuracy and timeliness of data available for analysis. This capability is crucial for monitoring patient outcomes and treatment efficacy in oncology clinical trials. The expanded site footprint also means we can gather more diverse and comprehensive data, enhancing the robustness of clinical research. Overall, the acquisition and integration of these technologies have positioned us to conduct more efficient and effective clinical trials, ultimately accelerating the development of new oncology therapies and improving patient outcomes.

Moe: How has Flatiron leveraged oncologists’ pathways to improve patient matching and pre-screening processes within the EHR system?

Craig Serra: At Flatiron, we have leveraged our intimate knowledge of physician workflows and long-standing relationships with oncology practices to significantly enhance patient matching and pre-screening processes within the EHR system. By deeply understanding patient journeys and the complexities of cancer care, we can identify more diverse and high functioning sites for cancer studies and identify trial-eligible patients at the point of care. This enables us to unburden clinical sites from the exhaustive task of manually reviewing patient charts. Instead, we handle the pre-screening and matching process centrally, using sophisticated algorithms and curated data to surface potential trial participants directly to the sites. This streamlined approach ensures that healthcare providers can focus on delivering high-quality patient care while we surface potentially eligible patients for trial participation allowing a patient to have a clinical trial discussion with their trusted physician.

Moreover, our system’s ability to seamlessly integrate into existing EHR workflows means that oncologists receive timely notifications about eligible patients during their routine clinical activities. This proactive identification and matching process improves the efficiency of patient recruitment and enhances the accuracy of matching patients to appropriate trials. By automating and centralizing these processes, we reduce the likelihood of missed opportunities for patient enrollment and ensure that more patients have access to potentially life-saving clinical trials.

Moe: What AI-driven innovations have you implemented recently, and how are they improving patient outcomes and research accuracy?

Craig Serra: Flatiron Health uses the power of data, enhanced by AI and other technologies, to close the gap between oncology clinical trials and care.

Our recent AI-driven innovations have focused on a core part of our mission – learning from the experience of every person with cancer. This is a challenge that we boldly take on, and there are two key aspects to it related to the application of AI to data: first, extremely valuable and important patient information is often stored in unstructured documents (e.g., physician’s notes); as such, extracting all relevant data is far from trivial. Second, as we continue to assess and analyze, we find that learnings can quickly become outdated given the speed at which cancer care is evolving. New therapies are made available regularly, new biomarkers become relevant to the standard of care, new variables (like Social Determinants of Health) become integral to the types of analyses researchers want to perform. Being able to dynamically update what we learn is critically important to our customers and patients alike. AI has the potential to address these challenges. 

With AI:

• We can achieve scale, rapidly iterating through charts in parallel across our total population of over 4 million patents (a task that would take a human 800 years to complete). 
• We can harness the ability to read through hundreds of pages of documents for each patient, to learn more about the experience of every person with cancer, extracting needle-in-a-haystack information like smoking status or comorbidities, that can be buried anywhere in the chart.
• We can dynamically keep up-to-date with the standard of care and new relevant variables as they emerge. Once a model is updated or developed, it can immediately be deployed across all patients.  

The ability to leverage AI to dynamically learn from millions of patients in real time has a significant impact on patient outcomes and research accuracy. By improving the speed and precision of data curation, and with our use of AI, we ensure that clinical trial designs are based on the most accurate and relevant information and can dynamically identify eligible patients for trials in real time. This enhances the overall quality of the trials and increases the likelihood of successful outcomes. Furthermore, AI systems can identify patterns and insights within the data that might be overlooked by manual processes, leading to novel discoveries and more effective treatment strategies.

Moe: How has the partnership with the FDA influenced the development and approval of new cancer therapies?

Craig Serra: Our collaboration with the FDA has revolutionized how real-world data (RWD) and RWE are utilized in regulatory processes for cancer therapies. By maintaining open communication and sharing insights, we have contributed significantly to shaping the FDA’s policies and thinking around integrating RWD and RWE into clinical trials. This partnership has provided us with a unique platform to demonstrate the value of real-world insights in understanding treatment efficacy and patient outcomes, ultimately influencing regulatory guidelines and standards. Our joint efforts have enabled the FDA to appreciate the nuances and potential of RWD, leading to more informed and flexible regulatory decisions that can accelerate the approval of new cancer therapies.

Furthermore, this collaboration has paved the way for supporting innovative approaches to clinical trial design and execution. Initiatives like the Center for Clinical Trial Innovation (C3TI) and the Streamlined Trials Embedded in Practice (STEP) project reflect our joint efforts between industry and regulators to enhance trial efficiency and effectiveness. Through these initiatives, we continue to explore new methodologies for integrating real-world clinical practice into trial designs, reducing the burden on patients and sites, and increasing the relevance and applicability of trial results.