Why interpreting data in clinical trials is the new sponsor challenge
For sponsors, the expansion of Digital Health Technologies (DHTs) such as wearables, remote sensors and smartphone-based tools, has opened the door to richer, more continuous insights into patient experience and disease progression. However, it has also introduced a challenge that many teams are still working through: understanding how these measurements fit together in relation to the research question and their clinical relevance for patients.
This interpretation challenge is becoming increasingly important as sponsors combine DHTs with traditional Clinical Outcome Assessments (COAs) in endpoint strategies.
COAs, including patient-reported outcomes (PROs), observer-reported outcomes (ObsROs), clinician-reported outcomes (ClinROs), and performance outcomes (PerfOs), have long provided the framework for evaluating symptoms, functioning, and treatment benefit in clinical trials. These instruments are designed around clearly defined concepts and validated using established scientific methods. They help sponsors understand whether a treatment effect is meaningful from a clinical and patient perspective.
DHTs, on the other hand, capture patient functions and physiology outside the clinic, often continuously. This includes movement, sleep, cognition, and other signals measured between site visits, at a frequency that was not feasible in traditional study designs. In many therapeutic areas, DHTs can reveal patterns that would otherwise go undetected.
Where the measurement logic breaks down
The challenge begins when sponsors attempt to combine these two measurement systems into a single endpoint strategy.
A wearable may show reduced mobility, a passive monitoring tool may detect disrupted sleep, and a smartphone assessment may identify fluctuations in cognition throughout the day. While those signals can be valuable, they do not automatically explain what patients are experiencing clinically, or whether observed changes reflect treatment benefit in a meaningful way. This is where interpretability becomes critical.
Many digital measures function as proxies for broader clinical concepts. Physical activity data may relate to fatigue, pain, physical functioning, motivation, or disease severity. Physiologic changes may correlate with symptom improvement in some patients but not others. Without a clearly defined conceptual relationship between the measures being used and their clinical meaningfulness, endpoint strategies can become difficult to interpret and harder to defend to regulatory agencies.
Sponsors are encountering this issue more often as digital endpoints mature. Data collection itself is no longer the primary hurdle, as most organizations already have access to sophisticated technologies capable of generating large volumes of longitudinal patient data. The harder question is how those measurements relate to validated clinical concepts and whether they support a coherent story about treatment effect.
This has elevated the importance of thoughtful COA strategy in digitally enabled trials.
COAs often provide the clinical context needed to interpret digital signals appropriately. For example, a PRO may help clarify whether increased activity levels correspond to improved functioning or reduced symptom burden, while a PerfO may provide additional perspective on passive mobility data collected through a wearable device. In other cases, clinician-reported assessments may help distinguish between physiologic variability and clinically meaningful improvement.
That does not mean every DHT requires direct anchoring to a COA, or that every digital measure should mirror an existing instrument. Hybrid endpoint strategies are more nuanced than that. Some digital measures may ultimately stand on their own with sufficient validation and regulatory support, as with the use of Stride velocity 95th centile (SV95C) as an alternative primary endpoint replacing the Six Minute Walk Test (6MWD) in trials of patients with Duchenne muscular dystrophy (DMD). Others may complement traditional assessments by adding sensitivity or continuous monitoring capabilities, as in the case of Atopic Dermatitis patients, for whom a wearable sensor can passively collect nocturnal movement data to complement traditional ClinROs and PROs of itch and scratch.
What matters is having a clear understanding of the clinical concepts being measured and the relationships between different types of endpoints.
What sponsors should consider before combining COAs and DHTs
For sponsors, designing endpoint strategies that integrate COAs and digital health technologies in a way that supports clinical interpretation starts long before data collection begins. Endpoint design decisions require careful evaluation of:
- measurement context and meaningful aspect of health (MAH)
- concepts of interest
- tool validity
- psychometric evidence
- interpretability across data sources
- practicability/usability of DHTs
As digital health technologies become more common in clinical trials, sponsors are generating far more continuous patient data than traditional study designs were built to handle. The central challenge is how to align these different measurements so they can be interpreted within a coherent clinical endpoint framework.
Adding wearable data or increasing the number of endpoints does not solve the underlying issue. Endpoint strategies need to start from clearly defined clinical concepts and an explicit understanding of how different assessments relate.
For many sponsors, COAs remain central to that process because they provide the clinical and patient-centered context needed to interpret emerging digital signals appropriately.
This is why structured COA information has become more important in endpoint design. Platforms like ePROVIDE from Mapi Research Trust help sponsors evaluate instruments based on their conceptual foundations, validation history, and scientific fit within broader measurement strategies.