Selecting between an actively- or passively-collected COA when designing your endpoint strategy

As drug development and clinical trial operations become increasingly more complex, a robust and streamlined endpoint strategy is a critical component to an asset's success. Digital health technologies (DHTs) are contributing to the rapidly evolving landscape of measures as they become more widespread and validated, presenting sponsors with a new challenge: choosing between a Clinical Outcome Assessment (COA) or a DHT when developing their endpoint strategy. The increasing availability of DHTs such as sensors, wearables, apps, and connected devices offers the opportunity to capture objective and continuous free-living data. This additional granularity can provide a real-world picture of a patient’s lived experience and serve as a way to passively collect clinical outcomes. On the other hand, the more traditional COAs are often the gold standard for many outcomes, offering interpretable scoring and helping to standardize results within and between studies in the same populations. These measures are familiar to researchers and regulators, and have been implemented in clinical trials and acknowledged in drug labels for decades.

As DHTs continue to demonstrate analytical and clinical validity, there are now multiple measurement options for the same concept of interest. Researchers must choose whether an actively-collected clinical outcome, like a PRO, or a passively-collected  outcome collected with a DHT is most appropriate for their protocol. These choices need to be grounded in what is most meaningful to patients, and must balance strategic considerations like cost, patient burden, trial feasibility and evolving Clinical Outcome Assessment and digital health regulatory requirements. 

Most often when developing an endpoint strategy, a trial will aim to collect one or more COAs, formally recognized and defined by the FDA as a measure of how a patient feels, functions or experiences their disease (Glossary - BEST (Biomarkers, EndpointS, and other Tools) Resource - NCBI Bookshelf). Traditionally, COAs have been collected through active means, whether that is through direct reporting from a patient (PRO), a clinician (ClinRO) or other observer (ObsRO), or through the performance of the patient during a standardized task (PerfO). These may be administered via paper or through an electronic clinical outcome assessment (eCOA) platform in eCOA clinical trials.

A digital health technology is “a system that uses computing platforms, connectivity, software, and/or sensors, for healthcare and related uses” (FDA). This is a broad category that can cover many types of technologies, such as medical devices, digital therapeutics, apps that deliver electronic versions of questionnaires (known as electronic clinical outcome assessments, or eCOAs), wearables, software and more. . In recent years, the FDA has published guidance on successfully incorporating DHTs into digital clinical trials to collect outcomes and has recognized DHTs' ability to collect both digital biomarkers, as well as measures that fall within the definition of a COA. Increasingly, digital biomarkers in clinical trials are being evaluated alongside traditional COAs to support endpoint strategies. Here our discussion focuses on those that can capture unique health metrics through hardware and/or software (sensors and wearables).

As shown in the table below, a COA assesses what a patient feels and how they function and a DHT assesses a biomarker or how a patient functions:

Actively- and passively-collected outcome assessments have significant overlap, but very often are each providing unique elements of a single concept of interest. This is where opportunities lie to leverage both types of assessment to create a more robust endpoint strategy and enhance the digital patient experience within trials.

There are a few key considerations when developing an endpoint strategy:

• Align the type of assessment with the concept of interest (and meaningful aspect of health): The first step researchers must take is to ensure the assessment type used to capture an endpoint is actually measuring what it is intended to. Reviewing the concepts and domains from actively collected PROs alongside passively collected digital measures using a unified framework can streamline the process, allowing for evaluation of the alignment among concepts, and ultimately, whether these concepts are measuring the most meaningful aspects of health according to patients. The Table 1 of the 3rd guidance in the FDA’s PFDD series illustrates a clear format to build compelling justifications for endpoint selection, applicable to both COAs and DHTs. This structured approach helps ensure both traditional COAs and digital endpoints support regulatory expectations.

• Evidence maturity and regulatory expectations: DHTs or COAs must have sufficient evidentiary support that it accurately reflects the concepts it aims to capture. For traditional COAs like PROs, this includes psychometric properties and content, linguistic and other forms of validation. For DHTs, this means that, when applicable, both the hardware (e.g. the sensors and materials) and the software (e.g. the algorithms and processing) have demonstrated the ability to obtain consistent, reliable, and valid measures. In addition to this verification and analytical validation, a digital measure must also be clinically validated for the context of use. Researchers often rely on the developers of these device to ensure the first two, but the responsibility of ensuring clinical validity can fall to researchers themselves, leveraging industry standards such as the V3+ framework for clear evaluation of DHTs. Regardless of assessment type, regulatory bodies, such as FDA and EMA, require clear justification and evidentiary support to be submitted with endpoints and are looking for these benchmarks to show proof of validity within a specified context of use.
• Participant burden and feasibility: Finally, it is critical to understand how an endpoint will be deployed at scale within a clinical trial. The schedule of assessments ultimately reflects what the patient will be asked to do within a trial, and how often. Understanding the trade-offs between actively collecting point-in-time data or more continuous passive collection helps highlight where and how the patient will participate, as well as the cost of the trial itself. Leveraging DHTs for free-living data can help reduce the burden of travel and study visits for patients, but increase the cost of analysis and securing the sensors and ensuring proper use and data transfer. Additionally, digital literacy, data privacy and security and user acceptance also factor into the success of DHT deployment. Actively-collected COAs often require participants to come to sites, which can be difficult for patients with mobility, transportation or other responsibilities that limit their ability to travel for study visits. However, these assessment types can be more familiar to patients and clinicians and widely available to license and administer.

Very often the conversation of selecting assessment types can seem dichotomous. Do we select a PRO or deploy a DHT? And while this may be the case in some instances, more often than not, there is value in recognizing the complimentary or additive nature of combining these two assessment types.

Below are a few real-world examples of when researchers have successfully leveraged COAs and DHTs in varying capacities.

• Interchangeable: In 2023, a breakthrough for digital endpoints came when EMA qualified the 95 Centile Stride Velocity (SV95) as a primary endpoint in ambulatory patients with Duchenne Muscular Dystrophy. The new digital endpoint serves as a replacement for the traditional PerfO, the 6 Minute Walk Test (6MWT). Researchers provided enough evidentiary support that the sensors worn were valid and reliable compared to the gold standard previously used, and now allowed for the capture of the primary outcome outside of clinic visits, offering a clearer view into a patient’s daily functioning.
• Additive: Researchers in the PROActive COPD initiative understood that physical activity, and, more specifically, being physically able to perform activities of daily living, is among the most meaningful aspects of health for patient’s living with COPD. However, they also recognized that no current assessment types measured the comprehensive experience of physical activity. By developing a hybrid measure, the PROactive Physical Activity in COPD (D-PPAC) captures both the objective aspects of activity as well as the subjective experience around being active. This novel measure demonstrated validity and superiority to using an existing PRO and DHT in parallel, and was qualified by EMA in 2018.
• Complementary: Perhaps the most ubiquitous use of COAs and DHTs is in their ability to complement each other by capturing different aspects of the same aspect of health, and when endpoint selection is grounded in the most meaningful aspects of health, the value of mixing assessment types becomes clear. In a qualitative study done in patients with atopic dermatitis, researchers built a conceptual model of several aspects of the disease that were most important to patients, and the available tools that can capture them. Scratching during the night, the feeling of itch, and sleep quality are all related concepts, none of which can be fully measured by a DHT or a COA on their own. Teasing out these concepts into outcomes to be measured shows where DHTs can lend the most value in capturing objective sleep measurements like number of awakenings, how long patients were asleep, as well as capturing scratch movements to quantify time scratching and intensity of scratching during the sleeping hours. Alternatively, PROs will be best suited to capture patient experience on how their itch feels, their perception of sleep quality and how these signs and symptoms impact their overall quality of life.

While the landscape of clinical trials continues to evolve, the foundational approach of the assessment type selection remains the same. Endpoint strategy should always be driven by the concept of interest, which should reflect something meaningful to the patient. The modality of data collection comes after the selection of the concepts of interest, but both digital health technologies and Clinical Outcome Assessments should be evaluated carefully for robust evidence in the context of use. By reviewing COAs and DHTs in an aligned and systematic way, researchers can easily identify where the two can be successfully combined or exchanged for one another, and how their complementary nature can help to support more meaningful trial outcomes. Ultimately, regulators require clear justification and evidence that the endpoints selected are valid, reliable, and patient-centered.

Mapi Research Trust and ICON have developed comprehensive evidence-based platforms designed to aid sponsors in navigating the complexity of outcome assessment and fit-for-purpose tool selection. Through ePROVIDE, a centralized COA database, sponsors can identify, review, and implement appropriate clinical outcome assessment tools, supported by structured COA licensing and linguistic validation processes. The ePROVIDE database, combined with the Atlas database for digital endpoints and digital measures, enable the identification, selection and deployment of COAs and DHTs, and our subject matter experts work closely with study teams to analyse the vast landscape of potential measures, build patient-centered conceptual models and allow for confident endpoint selection in both traditional and digital clinical trials.

Contact us for support with your clinical trials strategy today.

Thank you to Céline Desvignes-Gleizes, Scientific Principal, and Caprice Sassano, Sr. Outcomes Researcher, for their contributions to this article.