• Patient-reported outcome (PRO): A PRO is a direct report from the patient about how they feel or function in relation to their health condition or treatment, without interpretation by anyone else. It captures subjective outcomes like symptom severity, pain levels, or quality of life, providing crucial insight into the patient’s personal experience with a treatment.
• Patient-reported outcome measure (PROM): PROM refers to the specific tools or instruments (such as questionnaires or scales) used to collect PRO data. These measures are designed to systematically quantify the patient’s perspective on their health, functioning, or symptoms.
• Observer-reported outcome (ObsRO): An ObsRO is reported by someone other than the patient, typically a caregiver or family member, who can report on observable behaviors or aspects of health. This type of outcome is often used when the patient is unable to report for themselves, such as in the case of young children or individuals with cognitive impairments. Note that observer reports are different from proxy reports. Proxy reports are also completed by someone other than the patient, such as a caregiver, but proxy reporting also includes assessment of aspects of health that cannot be comprehensively observed and are only truly known by the patient themselves, such as the degree of fatigue or pain. (We discuss the challenges with proxy reporting in pediatric trials later in this article.)

Unlike clinician-reported or performance-based assessments, PROs and ObsROs offer a unique, patient-centric view of treatment effects, especially crucial in pediatric clinical trials.

In pediatric research, COA selection is particularly complex. Study teams must consider not only the treatment effects being measured but also whether the pediatric population can understand the questions and response scales. While numerous PROMs have been validated for different pediatric age groups, factors such as cognitive ability, willingness to cooperate, and the availability of validated instruments for the specific indication must also be considered.

To ensure robust data collection in alignment with regulatory guidance, it’s vital to choose appropriate COAs for pediatric populations. Selecting the right measures is key to developing a COA strategy that supports reliable, patient-centered outcomes.

Choosing between PROs and ObsROs

ObsROs are often used in place of PROs in instances where it is impractical or impossible to gain the patient perspective directly. The observer completing an ObsRO is typically a parent, a caregiver, or someone who observes the patient in daily life. It is key that ObsROs measure signs, events, or behaviors that are observable, and do not include personal judgment or interpretation.

By contrast, a proxy-reported outcome is a report provided by someone other than the patient, such as a family member or caregiver, about aspects of health that are only truly known to the patient themselves. Proxies attempt to reflect the patient’s subjective experience, focusing on what they believe the patient would report if they could.

While both involve third-party reporting, proxy reports aim to substitute the patient’s perspective, whereas observer-reported outcomes provide an objective evaluation of observable phenomena, making them useful for patients like infants or individuals with cognitive impairments who cannot articulate their experiences. For this reason, proxy-reported measures are discouraged by the FDA because these attempt to measure concepts known only to the patient themselves.

To further illustrate the difference, an example might be an assessment of a child’s pain. A proxy-reported outcome measure might ask how severe the child’s pain was during the day, whereas an observer-reported outcome would focus on pain-related signs such as irritability, crying, or holding the stomach. ObsROs and PROs are sometimes used together in a pediatric trial; this provides a rich picture of symptoms and signs from both the child and the caregiver perspective.

Patients’ Age Considerations

Out of nearly 55,000 clinical trials performed globally in 2022, about 10% included children up to 9 years old.

Industry bodies have also provided recommendations on the age at which a child may be considered able to reliably self-report. The Professional Society for Health Economics and Outcomes Research (ISPOR) created a Task Force to establish good research practices for the assessment of children and adolescents. In their report, they acknowledge that “It is not possible to provide age cutoffs that will fit every situation. Specific age cutoffs should be determined individually for each PRO instrument and tested with cognitive interviews in each new target population.” However, whilst not generalizable to every patient population and PROM, they provide four age groups as a starting point for consideration (see Table 1).

Similarly, the Patient-Reported Outcome Measurement Group at the University of Oxford agrees that it is not possible to provide specific age cutoffs applicable across all instances but suggests that “Children aged eight years and above are widely believed to be competent to complete such questionnaires; children aged five to eight years may be able to self-report but some are likely to require assistance to complete a questionnaire.”

To assist stakeholders on the collection and submission of PRO data, FDA released guidance in 2009. More recently, the agency supplemented this document by releasing a series of guidance and draft guidance documents on patient-focused drug development (PFDD) with the third installment on Selecting, Developing, or Modifying Fit-for-Purpose Clinical Outcome Assessments. According to these documents, the FDA encourages that children self-report whenever appropriate. That said, the FDA concedes that not all PROMs are appropriate, as certain response scales (such as 0-10 numeric rating scales, more/less comparisons) and choice of recall periods may be hard to understand in certain age groups. Additionally, research studies conducted in pediatric populations have shown that children under seven years old may have difficulty in understanding and using a visual analog scale (VAS), and that the use of pictorial scales improves the understanding of response options by mapping words to pictures.

In addition, FDA recommends using one version of a pediatric PROM, if possible, to avoid unwanted measurement variability, but also recognizes that there are instances where different age-appropriate versions may be needed to ensure full understanding. Depending on the duration of the clinical trial, a study may extend across various age groups through multiple years, which may create the challenge of patients aging up to the next age group.

To address this, study teams should decide a priori whether the PROM version should change as the patient ages or should stay the same to minimize variability. Precedent, input from key experts, and regulatory consultation may all play a part in such decisions. When using multiple age-related versions of a PROM within a clinical trial, researchers should consider how to interpret the data.

Where it is possible to demonstrate the comparability of scores derived from different age-appropriate versions, and that items are assessing the same constructs, it may be possible to pool the data collected across the different versions to make inferences. More likely, however, the data may need to be formally analyzed separately and assessed for commonality in findings. The approach taken should be determined a priori and documented in the statistical analysis plan, and in the case of pooling supported by evidence of comparability across versions (e.g., published literature or provided by the PROM license holder).

Measure Selection

When selecting a PROM or ObsRO to be used to derive an endpoint for a study, FDA encourages that these COAs should “reflect an aspect of the patient’s health that is meaningful; and be capable of supporting an inference of treatment effect within the context of the planned clinical trial.” Additional efforts need to be made in pediatric trials to ensure the assessment is suitable for completion by the appropriate age group.

The availability of a patient-completed or caregiver-report option for PROM completion depends on whether the assessment was validated for these use cases. Certain assessments do not have validated pediatric or observer versions, so it is important to work with psychometricians and instrument owners to fully understand the requirements associated with each measure. In these cases, the development of supporting content validation evidence may be required, which could include prospective qualitative studies or accomplished using embedded interviews within existing trials.

Moreover, additional content validity and validation testing may need to be done if applying an existing PROM to a new patient population. Researchers should consider whether original validation work might be generalizable to these new patient populations. Again, exit or embedded interviews might provide a useful means of collection of additional content validity data alongside an existing trial.

These evidentiary requirements also apply to custom-developed diaries and measures that are designed specifically for the study. They still require development of evidence to demonstrate their content validity, and that they are measuring aspects of health that are meaningful to the patient.

Response Scale Selection, Wording, and Use of Pictographs

The FDA guidance states that pediatric PROMs require additional reviews during content validity research to ensure that patients fully understand the concepts associated with each question and the response scale. Many assessments use a verbal response scale (VRS), a set of discrete ordered verbal descriptors to describe a variable of interest. Studies have shown that children can understand a 5-point VRS scale adequately, but the wording of these options needs to be crafted carefully to ensure proper understanding. For younger children under the age of 8, evidence suggests that a simplified 3-point Likert scale can be supplemented with cartoons (such as a smiling face, neutral face, and frowning face) to facilitate comprehension, keep children engaged, and decrease completion time. An interview-administered approach in which the questions are read to children can also be used for younger children and children with learning disabilities.

Item Skipping

A well-designed electronic implementation of PROMs and ObsROs offers the ability to clearly distinguish between questions missed inadvertently and those skipped deliberately. While it is attractive to eliminate missing data where possible, skipping ensures that patients do not simply make up an answer to enable them to continue PROM completion, for example, in instances in which they encounter a question to which they are unable or unwilling to respond. Enabling question skipping in pediatric research may be important if there is a perceived risk that some questions may not be fully comprehended. When implementing skipping, an electronic COA administration solution can ensure that the collected data reflects that the item was skipped intentionally; for example, by implementing a question or a pop up to confirm that it is the user’s intention to skip the item.

Placebo Response

In clinical trials, all health changes resulting from administering an inactive treatment, including regression towards the mean and natural course of the disease, also known as placebo response, should be mitigated as much as possible to prove that the study medication is effective. While placebo response is investigated heavily in indications such as ADHD, depression, and migraine, there is little knowledge in pediatric trials, especially in patients under 6 years old. Bridge et al. (2009) found that placebo response for major depressive disorder was higher in children under 12 years old (54.3%) than children older than 12 years old (44.9%).

Studies have shown that parents’ beliefs can influence physical and psychological functioning in children and adolescents. Additional research has shown that placebos can produce changes in how caregivers behave toward children which can then produce behavioral changes in the child, creating the concept of “placebo by proxy.” Study teams should mitigate placebo response as much as possible to prove the effectiveness of the study drug and reduce the risk of failed trials.

In trials requiring a placebo control, the treatment should be limited to the shortest duration needed to establish scientific validity. With the aim of minimizing the effects of placebo response, sites should provide information about the trial to the patient and the caregivers about placebo response and actions they can take to reduce its effects, and site personnel should themselves receive placebo response mitigation training (Table 2). Additionally, from a design perspective, studies have shown an increase in placebo response as the probability of receiving active agent increases. Thus, studies with 1:1 placebo to drug randomization have lower placebo response rates than studies with, for example, a 1:3 placebo to drug randomization scheme. Design elements, as well as training, can help reduce placebo response.

In the event that the investigational drug has already been proven in the adult population, and safety and response similarity can be modeled and extrapolated to the satisfaction of regulators, some placebo-controlled studies in children may not need to occur. The EU guidance Ethical Considerations for Clinical Trials on Medicinal Products Conducted with Minors states: “The choice of active or placebo control for randomized pediatric trials requires a situation of equipoise” and that “as the level of evidence in favor of an effective treatment increases, the ethical justification for placebo use weakens.” Recent examples of abbreviated pediatric exposures in placebo-controlled trials include trials of some adult antipsychotic medications for adolescent schizophrenia.

Recommendations

Because there are no black-and-white rules, it is important to develop an evidence-based rationale for the measurement strategy proposed, and, particularly where this is likely to be questioned, to engage with the regulators for early feedback and comment. Based on extensive experience and feedback from sponsors, regulatory bodies, sites, parents/caregivers, and patients, the authors offer the following recommendations for effectively collecting eCOA data in pediatric clinical trials.

Utilize PROMs Whenever Possible: For pediatric studies, as with adult studies, we recommend utilizing PROMs whenever possible. In addition, pediatric studies often benefit from the input of parents/caregivers via Observer-Reported Outcomes. The goal of each of these measures is to gain the unique perspective of the patient and family, separate from the viewpoint of the treating clinician. The use of Proxy-Reported Outcomes is discouraged. However, depending on the indication and the ability of the patient to answer, Proxy-Reported Outcomes might be the only option to obtain patient insight.

Thorough Review of Pediatric Condition and Measures: Thorough review of the pediatric condition under study, along with the age group, patients’ cognitive ability, validation of available measurements, and assessment of what has been done previously should be undertaken to inform the selection of a PRO or ObsRO. When selecting a PRO or ObsRO that will be used in a clinical trial, FDA encourages the assessments that are selected be used to accurately reflect a meaningful aspect of the patient’s health and assist in endpoint collection.

Age-Appropriate Assessments: It is imperative to examine the assessment’s supporting information to identify whether they are validated for the appropriate age groups and indications. If an assessment hasn’t been validated for the age group in question, content validity and cognitive debriefing assessment should be performed to verify that the version is understood and appropriate for use. Age-appropriate assessments should contain response scales that are suitable and well understood by the age group, such as pictorial scales in younger children.

Neutral Administration of Measures: It is essential that study staff be educated to administer measures in a manner that is in accord with how the scale was validated and in a manner that is neutral and does not influence responses or inflate placebo response. In pediatric studies, staff may need to be reminded of the importance of refraining from assisting/guiding their young patients as they naturally might. Staff should understand that their natural impulse to help a child determine a response to a measure could easily alter the response and undermine the integrity of the data collected.

Consistency in Observer Reporting: Study staff should remind caregivers that, where possible, a single observer should complete the assessments to ensure consistency and reduce variability in the data collected. If this is impractical, it will be important to be able to identify the caregiver associated with each completed assessment so that this can be accounted for or examined in the statistical analysis.

Minimize Placebo Response: Study teams should aim to minimize placebo response as much as possible through careful study design. Placebo response mitigation training should also be considered for site staff, caregivers, and patients.

Conclusion

The selection and implementation of patient-reported outcomes (PROs) and observer-reported outcomes (ObsROs) in pediatric clinical trials demand careful consideration and adherence to regulatory guidelines. PROs and ObsROs offer valuable insights into the patient’s perspective, contributing to a more comprehensive understanding of treatment effects and clinical meaningfulness. Challenges in measurement selection for pediatric trials include age, disease, and cognitive developmental status. Recommendations from clinical outcomes experts and regulatory agencies provide valuable guidance, emphasizing the importance of age-appropriate instruments and standardized administration protocols. Study teams should prioritize education and training of patient-facing staff to ensure neutral administration and minimize potential biases. Ultimately, a thorough understanding of the pediatric condition under study, combined with thoughtful consideration of available measures and, when available, previous research, is essential for developing a robust COA strategy that captures clinically meaningful information, aligns with regulatory standards, and enhances understanding of treatment effects in pediatric clinical trials.