Issue:June 2025

CLINICAL TRIALS - Solving Clinical Trial Challenges Through Sub-Population Optimization & Modeling Solution


INTRODUCTION

Clinical trials are the foundation of medical research, paving the way for groundbreaking therapies and treatments. However, these trials often face significant challenges, including high failure rates, substantial costs, and complex patient recruitment. In recent years, a revolutionary approach known as Sub-population Opti­mization & Modeling Solution (SOMS) has emerged to transform the landscape of clinical trials and offer new hope for more effi­cient and successful studies.

WHAT IS SOMS?

SOMS represents a shift in how clinical trials are conducted and analyzed. A sophisticated tool, SOMS leverages artificial in­telligence (AI) and advanced analytics to identify biomarkers within specific patient subgroups that can predict treatment re­sponse. This capability helps researchers identify patient sub­groups with higher efficacy and/or risk for adverse events, which significantly boosts the likelihood of trial success.

A MULTITUDE OF BENEFITS

One significant advantage of SOMS is its ability to progres­sively enhance trials throughout the process. From the outset, re­searchers can track subgroups, validate initial hypotheses, and run ongoing biomarker analyses to uncover new subgroups. This dynamic approach stands in stark contrast to traditional methods, which often rely on predetermined variables and static analysis. Trials that do not start with SOMS can still benefit from its intro­duction later, especially in cases of slow patient recruitment or minimal differences in treatment responses between treatment and placebo groups.

The power of SOMS includes real-time analysis, trial simu­lation, and benchmarking. Using simulated or real-world data, researchers can predict trial outcomes based on specific patient pool characteristics. This feature allows sponsors to simulate var­ious scenarios and incorporate historical data to create more ac­curate patient pools. Moreover, SOMS can compare standard treatments or other therapies in a specific therapeutic area/indi­cation to provide important insights into the potential perform­ance of a new therapy in real-world situations.

The importance of SOMS becomes particularly apparent when considering the substantial costs and high failure rates as­sociated with clinical trials. Taking a ther­apy through all trial stages can exceed $1 billion, with failure rates of 65% in Phase 2 and 35% in Phase 3. Even more con­cerning, only 12% of therapies that com­plete Phase 3 receive FDA approval.1 In this challenging landscape, SOMS offers a lifeline, helping to optimize trials and con­trol for elements that might otherwise go unnoticed or take too long to identify.

One of the key strengths of SOMS lies in its AI integration. In contrast to tradi­tional statistical methods in which statisti­cians decide variables beforehand, SOMS uses a data-driven approach to analyze multiple variables at once. This allows the AI to evaluate exponential permutations and comprehensively analyze all possible patient subgroups that might respond bet­ter to therapy or have improved safety pro­files.

SOMS demonstrates impressive pro­cessing efficiency. Once data has been prepared, SOMS can generate analyses within 30 seconds for Phase 1 and Phase 2 data, or within a couple of hours for larger Phase 3 datasets. A comprehensive analysis, including data preparation, can be achieved within a week. This translates to a significant, near 20x reduction in analysis time compared to traditional methods, enabling continuous optimiza­tion throughout the trial lifecycle.

SOMS’s credibility stems from its use of validated, open-source algorithms with a proven track record, which lends weight to results presented to health authorities. The system also offers flexibility, allowing users to modify various criteria as needed. Its repeatability across trials within a port­folio, with consistently prepared data and adapted algorithms, has proven effective across multiple therapeutic areas.

KEY REAL-WORLD APPLICATIONS

One of the most critical applications of SOMS is to implement rescue strategies for struggling clinical trials. Common in­dicators of a trial in distress include unex­pected adverse events or lack of efficacy in the treatment group. When such issues arise, SOMS can intervene with targeted rescue strategies. For instance, in a Phase 3 trial with serious adverse events in a sub­set of patients, SOMS can quickly analyze the data to predict which subgroups are more susceptible to these events.

A real-world example of SOMS’s ef­fectiveness is from a Phase 3 trial in multi­ple myeloma that needed to identify patient subgroups at increased risk for car­diac failure. Using SOMS, researchers an­alyzed 25 various biomarkers, including demographic and disease characteristics. The analysis revealed two specific bio­markers within the patient population that indicated an increased risk of cardiac is­sues. This data empowered sponsors to focus on high-risk groups, implement pro­tective measures and introduce interven­tions to mitigate the incidence of cardiac failure.

SOMS’s ability to identify and opti­mize effective subgroups within a trial is another crucial feature. To address various analytical needs, the platform offers three versions of the core algorithm, called Sub­group Identification Based on Differential Effect Search (SIDES). These variations in­clude the basic SIDES algorithm, fixed SIDES and an adaptive SIDES. These algo­rithmic methods enable SOMS to analyze trial data with high accuracy and config­urability to effectively identify patient sub­groups with precision.

While SOMS excels at finding patterns when they exist, it is important to note that not every trial will have clear subgroup dis­tinctions. The impact of SOMS-driven safety measures on trial integrity and suc­cess rates depends on several factors, in­cluding the specific findings for each trial, the nature of the identified risks, and how researchers choose to act on the informa­tion SOMS provides.

When addressing patient recruitment challenges, SOMS takes a unique ap­proach. Rather than directly improving re­cruitment numbers, SOMS enhances the quality and relevance of the recruited pa­tients. This “fewer but right” approach can lead to better trial outcomes in the long-term, including potentially faster recruit­ment due to a more focused pool.

The therapeutic and financial out­comes of using SOMS in clinical trials are significant. In one instance, a Phase 3 trial for a new antibacterial treatment initially showed no overall treatment effect. By an­alyzing 26 biomarkers, SOMS identified a subpopulation with a strong enough re­sponse to secure FDA approval. This inter­vention not only resulted in a successful drug launch for a particular patient sub­group but also prevented a late-stage fail­ure, potentially saving pharmaceutical companies hundreds of millions of dollars.

SOMS offers valuable support to sponsors throughout clinical trials, from design to closeout. Its versatility allows ap­plication across the entire trial lifecycle, with a particularly significant impact be­tween Phase 2 and Phase 3. In Phase 1, SOMS can leverage existing data from the therapy in different indications or from similar compounds to simulate patient re­sponses. The simulated data helps inform Phase 2 design and patient selection.

The transition from Phase 2 to Phase 3 often sees the greatest impact of SOMS. By analyzing data from both Phase 1 and Phase 2, SOMS can refine inclusion/exclu­sion criteria, identify responsive sub­groups, and ultimately optimize overall trial design for Phase 3. These capabilities significantly increase the chances of trial success and regulatory approval. SOMS’s value extends beyond Phase 3. It can also be used in post-approval phases to bench­mark the therapy against standards of care.

LOOKING AHEAD

SOMS holds significant promise for the future of clinical trials. While SOMS currently operates independently, future developments aim to integrate it with data management systems. This will enable real-time analysis as data streams in, al­lowing for dynamic trial management with identification of emerging risks and oppor­tunities. In addition, incorporating SOMS into risk-based quality management tools will enhance risk prediction and mitigation to leverage automated workflows.

The evolution of algorithms is also a critical area. Currently, SOMS uses a gen­eralized approach across various indica­tions. The future lies in specialized algorithms for specific therapeutic areas. These will provide more nuanced and ac­curate insights, especially in early signal detection. Custom algorithms tailored to specific contexts are being developed to better predict efficacy and safety events, leading to more precise insights for differ­ent therapeutic areas and trial types.

As AI and advanced analytics mature, their role in clinical trial optimization will grow significantly. The trend is toward more integrated, intelligent, and responsive systems. AI is likely to impact various aspects of trial design, conduct, and analysis, including sophisticated predictive modeling, auto­mated patient matching and recruitment, real-time data analysis and AI-assisted protocol design.

Advanced analytics hold the potential for targeted ap­proaches leading to adaptive trial designs that evolve based on incoming data. This could significantly reduce the time and cost of bringing new treatments to market. Integrating diverse data sources, like real-world evidence and genetic in­formation, could lead to a more comprehensive understand­ing of treatment effects and patient responses, with AI playing a crucial role in synthesizing insights from these complex datasets.

As these technologies evolve, ethical considerations and explainable AI will be increasingly important. AI-driven deci­sions in clinical trials must remain transparent, interpretable and aligned with patient safety and regulatory requirements. The future of clinical trial optimization with tools like SOMS and other AI-driven solutions points toward more efficient, precise, and adaptive trials. This has the potential to revolu­tionize how new therapies are developed and brought to market.

SOMS is a powerful solution for many clinical trial chal­lenges. By leveraging advanced analytics and AI, it offers a data-driven approach to optimizing patient subgroups, pre­dicting outcomes and enhancing overall trial efficiency. As the field continues to evolve, SOMS and similar technologies are poised to play a crucial role in shaping the future of med­ical research and drug development, ultimately leading to better treatments and improved patient outcomes.

REFERENCE

  1. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9293739/.

Adrian Kizewski is Associate Director of Product Management at IQVIA. He brings expertise spanning R&D and clinical life sciences, business analysis, process design and improvement, and product implementation. He is currently a lead for IQVIA’s Clinical Data Analytics Solution (CDAS) as well as Sub-population Optimization and Modeling Solution (SOMS). He earned his MBA from the McDonough School of Business at Georgetown University, in addition to his MSc in Pharmacology from The Johns Hopkins University School of Medicine and a BSc in Biochemistry from Temple University.