Why Powerful Real-World Evidence Requires Machine Learning

Real-world evidence (RWE) is a major component in helping life sciences companies predict patient outcomes and optimize treatment plans. So, how is it generated, and how do powerful RWE and machine learning work together?

It begins with having the right artificial intelligence (AI) technology and expertise to curate large amounts of real-world data (RWD), such as electronic health records (EHRs), claims, imaging, and genomics. One of the AI techniques that’s helping unlock the full potential of RWD is machine learning. 

What Is Machine Learning?

The Role of Machine Learning in RWE

Through the power of AI and ML algorithms, our Verana Health team has developed Qdata^® modules, which are fit-for-purpose datasets ready for research use to generate RWE

Combined with the expertise of physicians and data scientists, ML is redefining the scope and depth of RWE. 

Verana Health specializes in applying AI-powered techniques, such as ML and natural language processing (NLP), to analyze patterns of language in unstructured clinical notes of EHRs that signal key milestones and clinical insights throughout the patient journey. These clues can help researchers determine if patients are receiving the right treatments, based on disease progression. 

Our team of physicians and data scientists, with deep expertise in data-driven research, ensure that AI models are continually being trained and validated to catalog and categorize unstructured data to make it more easily analyzable. This approach also includes robust testing and validation before moving any model into production via comprehensive quality metrics.

Through the power of AI and ML algorithms, our Verana Health team has developed Qdata^® modules, which are fit-for-purpose datasets ready for research use to generate RWE. Qdata is powered by our VeraQ^® population health data engine which contains exclusive medical specialty data on more than 90 million de-identified patients from 20,000 contributing clinicians in the fields of ophthalmology, urology and neurology.

By leveraging this expansive RWD network, we’ve derived meaningful insights on a variety of conditions, including geographic atrophy, prostate cancer, Parkinson’s Disease and more!

The Impact of AI on Medical Research 

The integration of RWE and AI with medical research holds the potential to revolutionize healthcare and bring about significant advancements in medicine.

Here are some examples of how we’ve applied ML towards geographic atrophy RWE:

• Automated Segmentation of Geographic Atrophy Using Real-World Fundus Autofluorescence Ophthalmic Images
  • Verana Health developed an ML pipeline to automatically segment and assess the size of atrophic lesions in patients with geographic atrophy using real-world fundus autofluorescence images. 
  • Conclusion: The ML pipeline could be leveraged to assess real-world geographic atrophy disease progression at scale, helping to inform monitoring and treatment decisions. 
• Automated Identification of Geographic Atrophy Eyes with and without Subfoveal Involvement Using Machine Learning and Real-World Ophthalmic Images in the IRIS Registry
  • Verana Health developed an ML pipeline to identify eyes with GA that have subfoveal involvement and eyes without subfoveal involvement using real-world optical coherence tomography images. 
  • Conclusion: The proposed pipeline demonstrated satisfactory performance in identifying eyes with and without subfoveal involvement using images collected in real-world practice. This could potentially be useful in screening patients for geographic atrophy clinical trials and identifying patients for future treatments.

How ML and RWD Are Addressing Challenges

Real-world data and machine learning are driving innovation in many aspects of healthcare, including drug development, personalized medicine, and healthcare management. 

Here are some of the challenges that ML can help address by analyzing RWD:

• Cleansing and Preprocessing of Data: ML models excel in identifying and correcting data inconsistencies. Automated data quality processes ensure that the data used for analysis is robust, thereby enhancing the reliability of the results.
• Uncovering Hidden Patterns: ML can uncover hidden patterns and relationships within the data. For example, deep learning models can predict patient outcomes, identify risk factors, and suggest personalized treatment options based on historical data.
• Personalizing Medicine: By analyzing genetic, phenotypic, and environmental data, ML algorithms can determine which treatments are most effective for specific patient subgroups. This leads to personalized medicine, where treatments are tailored to individual patient profiles, improving efficacy and reducing the likelihood of adverse effects.
• Aiding in Biomarker Discovery: ML facilitates the discovery of new biomarkers that can predict disease progression and treatment response.
• Monitoring Post-market Surveillance: ML assists in monitoring the safety and efficacy of drugs post-approval by continuously analyzing RWE.
• Reducing Costs: ML can assess the cost-effectiveness of treatments and interventions by analyzing outcomes and costs in real-world settings. 
• Improving Clinical Trials: ML enhances the design and execution of clinical trials. When leveraging ML models, RWE can be applied to inform and streamline protocol development, enhance trial design, facilitate the selection of qualified clinical trial sites, and improve the recruitment and engagement of diverse patient populations.
• Supporting Decision-making: ML can support decision-making by providing evidence-based recommendations. For example, ML algorithms can assist clinicians in diagnosing conditions, suggesting treatment plans, and predicting patient outcomes, thereby augmenting human expertise with data-driven insights.

ML and RWE Are Driving Healthcare Innovation

Incorporating machine learning in real-world evidence analysis is not just a technological advancement; it’s a necessary evolution for anyone looking to stay at the forefront of data-driven decision-making and healthcare innovation. By utilizing ML, life sciences companies can unlock the full potential of RWD, leading to more accurate, timely, and actionable insights.