Background
Patient burden - the cumulative physical, emotional, and time demand placed on participants in clinical trials - is widely recognized as a contributor to recruitment challenges, dropout, and site burden. Yet until recently, burden assessment has been largely qualitative: subjective judgments made during protocol review, often without access to validated scoring methods or comparative benchmark data.
This creates a systematic problem. Protocol designers lack objective tools to evaluate the burden implications of their design choices in real time. The result is protocols that are more burdensome than necessary - contributing to enrollment shortfalls, high dropout rates, and the staggering statistic that 80% of clinical trials fail to enroll on time.
"Patient burden is a design variable - not a post-hoc concern. Optimizing it during protocol design is one of the highest-leverage interventions available to clinical development teams."
Our Scoring Methodology
The Trials.ai Patient Burden Index scores burden at three hierarchical levels, enabling both granular analysis and aggregate comparison:
- Activity Level: Each individual procedure or assessment is scored based on its intrusiveness (invasive vs. non-invasive), duration, preparation requirements, recovery time, and frequency. Activities are drawn from our Clinical Studies Ontology, which links each procedure to standardized terminology and population-specific burden factors.
- Visit Level: Visit burden aggregates activity-level scores while accounting for visit duration, travel requirements, and the clustering of high-burden procedures. High-burden clusters - multiple invasive procedures concentrated in a single visit - are flagged as optimization targets.
- Schedule Level: Schedule burden represents the cumulative demand across the entire study - the total burden imposed on a patient who completes the full protocol. This enables meaningful cross-study comparison and benchmark-based optimization.
Burden scores are computed in real time within Smart Designer as scientists build their Schedule of Activities, enabling design decisions to be informed by burden impact at the moment they are made - not during a downstream review cycle.
The Tufts Collaboration
To validate the Burden Index as a predictor of operational outcomes, we partnered with the Tufts Center for the Study of Drug Development - the leading academic research center focused on the productivity of pharmaceutical innovation. Tufts' longitudinal database of clinical trial performance data provides ground-truth operational metrics for a broad sample of completed trials, spanning multiple therapy areas, phases, and geographic regions.
The collaboration linked our burden scores - calculated retrospectively from digitized protocols - to Tufts' operational outcomes data, enabling rigorous statistical testing of whether burden scores at the design stage predict operational performance at execution. This retrospective validation study represents one of the first systematic analyses of the relationship between quantitative protocol burden and trial operational performance.
Key Findings
Our analysis revealed several statistically significant correlations between protocol burden scores and operational outcomes:
Studies in the highest burden quartile showed dropout rates 34% higher than those in the lowest quartile, controlling for therapy area, phase, and geographic region. This finding held across oncology, metabolic disease, and central nervous system indications - suggesting that burden impact on retention is a generalizable phenomenon rather than indication-specific.
High-burden visit clustering was independently associated with increased screen failure rates. When three or more high-burden procedures were concentrated in screening visits, screen failure rates increased by an average of 22% compared to studies with distributed burden across visits. This finding suggests that burden optimization at the visit level - spreading high-burden activities across multiple visits - is a meaningful lever for improving enrollment efficiency.
Most significantly for prospective application: studies where burden was actively optimized during design - shifting procedures between visits, substituting lower-burden alternatives, or restructuring assessment schedules - showed enrollment timelines 18% shorter on average than matched controls with equivalent scientific scope. This suggests that burden optimization during protocol design translates directly to operational performance.
Implications for Protocol Design
These findings have direct implications for how clinical study design should be approached. Burden is not a post-hoc concern to be addressed during protocol review - it is a design variable that should be optimized during the protocol development process, alongside cost, scientific validity, and operational feasibility.
The Trials.ai platform integrates burden scoring directly into the study design workflow. Clinical scientists see the burden implications of their design choices in real time, model alternative scenarios, and receive Franklin-guided recommendations for burden optimization - before the protocol reaches regulatory submission. The burden score is not a report generated after design is complete. It is a live metric that evolves with every design decision.
Future Research
We are continuing to expand the validation study to broader sample sizes and additional therapy areas, with a focus on rare disease indications where enrollment challenges are particularly acute. We are also investigating the relationship between burden scores and specific regulatory outcomes, including clinical hold rates and post-submission amendment frequency.
A full peer-reviewed publication of our methodology and findings is currently in preparation. We anticipate submission to a peer-reviewed clinical trials journal in the second half of 2026. Researchers and clinical development professionals interested in collaborating on burden methodology research are encouraged to reach out.