Clinical Outcome Assessments and Digital Measures

As patient-centered clinical trials have become a regulatory expectation rather than a differentiator, sponsors have a variety of options to measure treatment benefit. Traditional clinical outcome assessment (COA) methodology now exists alongside electronic assessments (eCOA), wearable sensors, smartphone-based technologies, and other digital health technologies (DHTs). 

In fact, these different methods can appear together within the same protocol. 

How a measure is obtained and who, or what, is reporting the information (e.g. a human or a sensor) affects how it should be developed, validated, implemented, and ultimately defended during regulatory review. Most critically, the key question for sponsors is whether the measure captures the clinical concept that reflects a meaningful aspect of patients’ lives, and whether the evidence generated will support the intended use. 

At Mapi Research Trust, we work with sponsors navigating these decisions by providing access to validated COAs and DHTs and the information needed to support appropriate tool selection, licensing, use, and deployment. The principles that guide COA selection remain just as important as the field expands into digital measurement. 

This glossary explains the major terms trial teams encounter and the considerations that should guide measurement strategy.

A clinical outcome assessment is a tool that reflects how a patient feels, functions, or survives.  (BEST (Biomarkers, Endpoints and Other Tools) Resource). 

COAs are generally divided into four categories: patient-reported outcomes (PROs), observer-reported outcomes (ObsROs), clinician-reported outcomes (ClinROs), and performance outcomes (PerfOs). COAs are designed to assess outcomes meaningful to the patient’s health status, rather than simply measuring a biological signal or disease marker. 

The distinction between COA types is based on who provides the information and how that information is generated. 

A patient-reported outcome comes directly from the patient, without interpretation by a clinician, caregiver, or other observer. 

Symptoms or other unobservable concepts known only to the patient (e.g., pain severity or nausea) can only be measured by PROs (BEST (Biomarkers, Endpoints and Other Tools) Resource and FDA-PFDD Glossary). 

PROs provide something uniquely valuable in clinical research: direct insight into the patient’s own experience of disease and treatment. 

That direct perspective is also why PRO development requires careful attention. A PRO cannot simply be borrowed from another disease area or population because it appears relevant. Sponsors need evidence that the questions capture concepts that matter to the intended patient population, and that patients interpret those questions as intended. 

There is an important limitation of PROs, which is that a patient must be able to reliably report their own experience. In populations such as very young children, individuals with significant cognitive impairment, or patients with conditions affecting communication, self-reporting may not be feasible.  

An observer-reported outcome is a measurement based on a report of observable signs, events or behaviors related to a patient’s health condition by someone other than that patient or a health professional (FDA-PFDD Glossary). ObsRO shifts reporting from the patient to someone who regularly observes the patient’s condition, such as a parent, caregiver, or family member. 

ObsROs are particularly valuable when patients cannot reliably report their own symptoms or functioning. Pediatric trials often incorporate caregiver observations, as do studies involving individuals with severe cognitive limitations. 

A critical distinction between ObsROs and PROs is that observers can report what they see, not what they assume the patient feels. A caregiver may accurately report that a child is sleeping poorly, crying more often, or avoiding certain activities. That same caregiver cannot directly report the child’s internal experience of pain or emotional distress unless the patient has communicated it. 

This boundary is an important principle in ObsRO development. Instruments that unintentionally ask observers to interpret internal experiences rather than report observable behaviors risk moving beyond what an observer can validly assess. 

A clinician-reported outcome is a measurement based on a report that comes from a trained health-care professional after observation of a patient’s health condition (FDA-PFDD Glossary).  

ClinROs are often used when clinical expertise is required to interpret findings. Examples include disease severity ratings, structured clinical assessments, and physician global assessments. 

The difference between an ObsRO and a ClinRO comes down to interpretation. A caregiver describes what they observe. A clinician applies medical knowledge and defined assessment criteria to evaluate what those observations mean. 

ClinROs are particularly useful when a clinical perspective is essential to understanding treatment response. They may also complement PROs by providing two perspectives on the same disease experience: the patient’s own report and an independent clinical assessment. 

A performance outcome is a measurement based on a standardized task(s) performed by a patient that is administered and evaluated by an appropriately trained individual or is independently completed. PerfOs require patient cooperation and motivation. These include measures of gait speed (e.g., timed 25 foot walk test), memory recall (e.g., word recall test), or other cognitive testing (e.g., digit symbol substitution test) (BEST (Biomarkers, Endpoints and Other Tools) Resource and FDA-PFDD Glossary).Unlike PROs and ObsROs, PerfOs do not rely on someone reporting an experience. The patient’s performance itself generates the measurement. 

This makes PerfOs an important bridge between traditional COAs and emerging digital approaches. Many performance-based concepts, such as mobility, activity, and motor function, can now also be measured outside a clinical setting using wearable sensors or connected health devices. 

However, a digital measurement of a similar concept is not automatically equivalent to an established PerfO. A clinic-based walking assessment and a wearable-generated gait measure may both relate to mobility, but sponsors still need evidence demonstrating that the digital measure is reliable, meaningful, and appropriate for the intended purpose.

Electronic clinical outcome assessment (eCOA) is not a separate COA category, but a method of collecting COA data electronically. A patient completing a symptom diary through a smartphone application is still completing a PRO. A clinician entering an assessment through an electronic system is still completing a ClinRO. 

When migrating a paper-based COA to an electronic version, it is critical to complete a thorough review of the digital version to ensure a faithful migration, i.e., that the integrity of the original assessment has been preserved.  

The move from paper to electronic collection has become increasingly common because electronic systems can improve data quality, reduce transcription errors, support audit trails, and provide better visibility into completion patterns

Digital health technologies (DHTs) have expanded the ways researchers can collect health-related information. These technologies may include wearable sensors, smartphone applications, connected medical devices, and other systems capable of capturing data outside traditional clinical settings (U.S. Food and Drug Administration). 

The data generated by these technologies then become digital measures (Izmailova et al, 2023). 

Examples might include activity levels, gait speed, sleep patterns, or other measurements derived from sensor data. However, the existence of a digital measure does not automatically make it clinically meaningful or suitable for regulatory decision-making. 

Sponsors must establish multiple layers of evidence: 

1. The technology must accurately capture the intended signal.  
2. The algorithm must reliably convert raw data into a meaningful measurement.  
3. And the resulting measure must demonstrate relevance to the clinical concept being studied. 

These measures can result in either digital biomarkers or digitally-collected COAs, which can then become digital endpoints. 

A digital biomarker is a measure derived from a digital health technology that has been linked to a physiological or clinical state (Macias Alonso et al, 2024). For example, a continuous glucose monitor (CGM) can passively and continuously measure glucose levels in patients with Type 1 Diabetes. This validated measurement can serve as an indicator of short-term glycemic control, variability and trends. A digital endpoint is the outcome a clinical trial formally uses to test its hypothesis (Landers et al, 2021). For example, a sponsor developing a Parkinson’s disease therapy might define change in gait speed measured by a validated wearable device as an exploratory or secondary endpoint to evaluate whether treatment improves motor function. 

The same digital measure may therefore have different roles depending on the evidence available and the objectives of the study. A sponsor may initially use wearable-derived gait speed as an exploratory digital endpoint while building the evidence needed to demonstrate clinical relevance. A later trial may use that same measure as a formal digital endpoint once sufficient validation has been established. 

The use of digital health technologies in clinical trials has increased recently, giving sponsors new options for how they measure patient experience and treatment response. But there is still often a lack of clarity around when a digital measure should complement an established clinical outcome assessment, and when it might provide a stronger approach on its own (Izmailova et al, 2023). 

The decision depends on the clinical question the trial is designed to address. Digital measures can provide insight into aspects of health that may be difficult to capture through scheduled assessments alone, including day-to-day, at-home changes in mobility, symptoms, activity, or functioning. For some conditions, this additional context can be valuable.  

But more data does not automatically equate to better evidence. Digital measures introduce a new set of considerations around validation, implementation, data interpretation, and regulatory acceptance. The right measurement mix depends on what the trial needs to understand, and the right tool to use is the one that best captures the most meaningful aspects of health and generates evidence that supports the intended use. 

In deciding which measurement to use, sponsors should consider several questions: 

1. What concept(s) are being measured? A symptom experience, functional ability, disease severity, or quality-of-life impact may require different approaches. 
2. Who or what is best positioned to provide that information? Sometimes the patient is the most reliable source. In other cases, caregiver observation, clinician assessment, objective task performance or sensor-derived measuresmay be more appropriate. 
3. Does continuous measurement add meaningful scientific value? Digital approaches are especially useful when a condition changes between scheduled clinical visits. Symptoms that fluctuate throughout the day, changes in mobility, or variations in sleep patterns may be difficult to capture through periodic assessments alone. 

However, adding a digital measure simply because the technology exists can create unnecessary complexity. If a validated COA already answers the research question, introducing another measurement may increase validation requirements without improving the quality of evidence. 

The expanding measurement landscape requires expertise across multiple disciplines. Selecting an appropriate measure involves clinical science, regulatory strategy, patient engagement, technology assessment, and data expertise. 

These decisions often become most important at three points in development. 

1. During endpoint strategy discussions, sponsors must determine whether a digital measure is sufficiently mature to support a primary or secondary endpoint or whether it should remain exploratory. 
2. During tool selection, teams must weigh established COAs against newer digital approaches and determine which option provides the strongest evidence for the intended purpose. 
3. During regulatory engagement, sponsors must clearly explain why a measurement approach is appropriate and how the supporting evidence addresses potential reviewer questions. 

Bringing this expertise together early can help prevent costly challenges later, including protocol amendments, additional validation work, or uncertainty about whether collected data will support a regulatory claim. 

Digital technologies have expanded what is possible in clinical research, but they have not changed the fundamental principles of good measurement. Whether selecting an established clinical outcome assessment, evaluating a digital measure, or designing a hybrid measurement strategy, early decisions influence the quality and credibility of the evidence a trial will produce. 

Through resources such as ePROVIDE’s PROQOLID™, PROLABELS™, PROINSIGHT™ and Atlas Network databases, scientific consulting and licensing expertise, and decades of experience supporting sponsors worldwide, Mapi Research Trust helps clinical teams identify validated tools, navigate implementation considerations, and build measurement strategies that support regulatory confidence. 

Explore validated clinical outcome assessments or connect with our experts to discuss the right measurement approach for your next clinical trial. 

Thank you to Céline Desvignes-Gleizes, Caprice Sassano and Aisling Curran for their contributions to this article. 

Related reading:  

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

Selecting fit-for-purpose Clinical Outcome Assessments (COAs) using the ePROVIDE databases